diff --git "a/checkpoint-2322/trainer_state.json" "b/checkpoint-2322/trainer_state.json" new file mode 100644--- /dev/null +++ "b/checkpoint-2322/trainer_state.json" @@ -0,0 +1,38675 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 1.8013964313421256, + "eval_steps": 20, + "global_step": 2322, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.0007757951900698216, + "grad_norm": Infinity, + "learning_rate": 0.0, + "loss": 6.6798, + "step": 1 + }, + { + "epoch": 0.0015515903801396431, + "grad_norm": 19.460369110107422, + "learning_rate": 4.112808460634547e-08, + "loss": 6.1418, + "step": 2 + }, + { + "epoch": 0.0023273855702094647, + "grad_norm": 38.97050857543945, + "learning_rate": 8.225616921269094e-08, + "loss": 6.8984, + "step": 3 + }, + { + "epoch": 0.0031031807602792862, + "grad_norm": 25.763484954833984, + "learning_rate": 1.2338425381903642e-07, + "loss": 7.1881, + "step": 4 + }, + { + "epoch": 0.003878975950349108, + "grad_norm": 11.17786979675293, + "learning_rate": 1.6451233842538187e-07, + "loss": 5.5114, + "step": 5 + }, + { + "epoch": 0.004654771140418929, + "grad_norm": 15.640860557556152, + "learning_rate": 2.0564042303172737e-07, + "loss": 6.2131, + "step": 6 + }, + { + "epoch": 0.005430566330488751, + "grad_norm": 26.026487350463867, + "learning_rate": 2.4676850763807285e-07, + "loss": 7.43, + "step": 7 + }, + { + "epoch": 0.0062063615205585725, + "grad_norm": Infinity, + "learning_rate": 2.4676850763807285e-07, + "loss": 9.8237, + "step": 8 + }, + { + "epoch": 0.0069821567106283944, + "grad_norm": 12.019311904907227, + "learning_rate": 2.878965922444183e-07, + "loss": 5.4554, + "step": 9 + }, + { + "epoch": 0.007757951900698216, + "grad_norm": 19.58979606628418, + "learning_rate": 3.2902467685076374e-07, + "loss": 6.6825, + "step": 10 + }, + { + "epoch": 0.008533747090768037, + "grad_norm": 11.048348426818848, + "learning_rate": 3.701527614571093e-07, + "loss": 5.6752, + "step": 11 + }, + { + "epoch": 0.009309542280837859, + "grad_norm": 10.552579879760742, + "learning_rate": 4.1128084606345474e-07, + "loss": 4.8126, + "step": 12 + }, + { + "epoch": 0.01008533747090768, + "grad_norm": 10.90353775024414, + "learning_rate": 4.524089306698002e-07, + "loss": 4.7993, + "step": 13 + }, + { + "epoch": 0.010861132660977503, + "grad_norm": 14.389288902282715, + "learning_rate": 4.935370152761457e-07, + "loss": 6.2433, + "step": 14 + }, + { + "epoch": 0.011636927851047323, + "grad_norm": 32.469276428222656, + "learning_rate": 5.346650998824911e-07, + "loss": 7.8983, + "step": 15 + }, + { + "epoch": 0.012412723041117145, + "grad_norm": 17.988374710083008, + "learning_rate": 5.757931844888366e-07, + "loss": 5.9294, + "step": 16 + }, + { + "epoch": 0.013188518231186967, + "grad_norm": 53.49104309082031, + "learning_rate": 6.169212690951821e-07, + "loss": 10.0069, + "step": 17 + }, + { + "epoch": 0.013964313421256789, + "grad_norm": 51.57087707519531, + "learning_rate": 6.580493537015275e-07, + "loss": 9.761, + "step": 18 + }, + { + "epoch": 0.01474010861132661, + "grad_norm": 10.312777519226074, + "learning_rate": 6.99177438307873e-07, + "loss": 4.7492, + "step": 19 + }, + { + "epoch": 0.015515903801396431, + "grad_norm": 24.298805236816406, + "learning_rate": 7.403055229142186e-07, + "loss": 7.0924, + "step": 20 + }, + { + "epoch": 0.015515903801396431, + "eval_Qnli-dev_cosine_accuracy": 0.58984375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.9331285953521729, + "eval_Qnli-dev_cosine_ap": 0.5485925209297758, + "eval_Qnli-dev_cosine_f1": 0.6291834002677376, + "eval_Qnli-dev_cosine_f1_threshold": 0.6565657258033752, + "eval_Qnli-dev_cosine_precision": 0.4598825831702544, + "eval_Qnli-dev_cosine_recall": 0.9957627118644068, + "eval_Qnli-dev_dot_accuracy": 0.541015625, + "eval_Qnli-dev_dot_accuracy_threshold": 1048.2359619140625, + "eval_Qnli-dev_dot_ap": 0.4748490344531905, + "eval_Qnli-dev_dot_f1": 0.6300268096514745, + "eval_Qnli-dev_dot_f1_threshold": 380.322998046875, + "eval_Qnli-dev_dot_precision": 0.46078431372549017, + "eval_Qnli-dev_dot_recall": 0.9957627118644068, + "eval_Qnli-dev_euclidean_accuracy": 0.587890625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 9.073010444641113, + "eval_Qnli-dev_euclidean_ap": 0.5650640204703478, + "eval_Qnli-dev_euclidean_f1": 0.6332882273342355, + "eval_Qnli-dev_euclidean_f1_threshold": 18.225353240966797, + "eval_Qnli-dev_euclidean_precision": 0.46520874751491054, + "eval_Qnli-dev_euclidean_recall": 0.9915254237288136, + "eval_Qnli-dev_manhattan_accuracy": 0.6171875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 155.41647338867188, + "eval_Qnli-dev_manhattan_ap": 0.6018518855824669, + "eval_Qnli-dev_manhattan_f1": 0.6291834002677376, + "eval_Qnli-dev_manhattan_f1_threshold": 286.5093994140625, + "eval_Qnli-dev_manhattan_precision": 0.4598825831702544, + "eval_Qnli-dev_manhattan_recall": 0.9957627118644068, + "eval_Qnli-dev_max_accuracy": 0.6171875, + "eval_Qnli-dev_max_accuracy_threshold": 1048.2359619140625, + "eval_Qnli-dev_max_ap": 0.6018518855824669, + "eval_Qnli-dev_max_f1": 0.6332882273342355, + "eval_Qnli-dev_max_f1_threshold": 380.322998046875, + "eval_Qnli-dev_max_precision": 0.46520874751491054, + "eval_Qnli-dev_max_recall": 0.9957627118644068, + "eval_allNLI-dev_cosine_accuracy": 0.666015625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9902692437171936, + "eval_allNLI-dev_cosine_ap": 0.36630271296437167, + "eval_allNLI-dev_cosine_f1": 0.5088235294117648, + "eval_allNLI-dev_cosine_f1_threshold": 0.7946319580078125, + "eval_allNLI-dev_cosine_precision": 0.34122287968441817, + "eval_allNLI-dev_cosine_recall": 1.0, + "eval_allNLI-dev_dot_accuracy": 0.662109375, + "eval_allNLI-dev_dot_accuracy_threshold": 940.7398681640625, + "eval_allNLI-dev_dot_ap": 0.33409368825504626, + "eval_allNLI-dev_dot_f1": 0.5081240768094535, + "eval_allNLI-dev_dot_f1_threshold": 585.4312744140625, + "eval_allNLI-dev_dot_precision": 0.3412698412698413, + "eval_allNLI-dev_dot_recall": 0.9942196531791907, + "eval_allNLI-dev_euclidean_accuracy": 0.666015625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 5.293120384216309, + "eval_allNLI-dev_euclidean_ap": 0.36165210170894113, + "eval_allNLI-dev_euclidean_f1": 0.5111441307578009, + "eval_allNLI-dev_euclidean_f1_threshold": 16.931983947753906, + "eval_allNLI-dev_euclidean_precision": 0.344, + "eval_allNLI-dev_euclidean_recall": 0.9942196531791907, + "eval_allNLI-dev_manhattan_accuracy": 0.666015625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 74.05484008789062, + "eval_allNLI-dev_manhattan_ap": 0.37228608061825896, + "eval_allNLI-dev_manhattan_f1": 0.5081240768094535, + "eval_allNLI-dev_manhattan_f1_threshold": 240.91050720214844, + "eval_allNLI-dev_manhattan_precision": 0.3412698412698413, + "eval_allNLI-dev_manhattan_recall": 0.9942196531791907, + "eval_allNLI-dev_max_accuracy": 0.666015625, + "eval_allNLI-dev_max_accuracy_threshold": 940.7398681640625, + "eval_allNLI-dev_max_ap": 0.37228608061825896, + "eval_allNLI-dev_max_f1": 0.5111441307578009, + "eval_allNLI-dev_max_f1_threshold": 585.4312744140625, + "eval_allNLI-dev_max_precision": 0.344, + "eval_allNLI-dev_max_recall": 1.0, + "eval_sequential_score": 0.6018518855824669, + "eval_sts-test_pearson_cosine": 0.14468011325378688, + "eval_sts-test_pearson_dot": 0.12373166012012136, + "eval_sts-test_pearson_euclidean": 0.15040853567589774, + "eval_sts-test_pearson_manhattan": 0.14680735741048356, + "eval_sts-test_pearson_max": 0.15040853567589774, + "eval_sts-test_spearman_cosine": 0.1980088477205014, + "eval_sts-test_spearman_dot": 0.12132071438334546, + "eval_sts-test_spearman_euclidean": 0.17802306863688658, + "eval_sts-test_spearman_manhattan": 0.17413140944376768, + "eval_sts-test_spearman_max": 0.1980088477205014, + "eval_vitaminc-pairs_loss": 4.239284515380859, + "eval_vitaminc-pairs_runtime": 5.4781, + "eval_vitaminc-pairs_samples_per_second": 23.366, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 20 + }, + { + "epoch": 0.015515903801396431, + "eval_negation-triplets_loss": 5.499993801116943, + "eval_negation-triplets_runtime": 1.0772, + "eval_negation-triplets_samples_per_second": 118.828, + "eval_negation-triplets_steps_per_second": 1.857, + "step": 20 + }, + { + "epoch": 0.015515903801396431, + "eval_scitail-pairs-pos_loss": 2.806995391845703, + "eval_scitail-pairs-pos_runtime": 1.221, + "eval_scitail-pairs-pos_samples_per_second": 104.828, + "eval_scitail-pairs-pos_steps_per_second": 1.638, + "step": 20 + }, + { + "epoch": 0.015515903801396431, + "eval_scitail-pairs-qa_loss": 4.214225769042969, + "eval_scitail-pairs-qa_runtime": 0.7726, + "eval_scitail-pairs-qa_samples_per_second": 165.67, + "eval_scitail-pairs-qa_steps_per_second": 2.589, + "step": 20 + }, + { + "epoch": 0.015515903801396431, + "eval_xsum-pairs_loss": 5.982180118560791, + "eval_xsum-pairs_runtime": 3.9493, + "eval_xsum-pairs_samples_per_second": 32.411, + "eval_xsum-pairs_steps_per_second": 0.506, + "step": 20 + }, + { + "epoch": 0.015515903801396431, + "eval_sciq_pairs_loss": 0.733872652053833, + "eval_sciq_pairs_runtime": 5.8101, + "eval_sciq_pairs_samples_per_second": 22.031, + "eval_sciq_pairs_steps_per_second": 0.344, + "step": 20 + }, + { + "epoch": 0.015515903801396431, + "eval_qasc_pairs_loss": 8.005105018615723, + "eval_qasc_pairs_runtime": 0.89, + "eval_qasc_pairs_samples_per_second": 143.822, + "eval_qasc_pairs_steps_per_second": 2.247, + "step": 20 + }, + { + "epoch": 0.015515903801396431, + "eval_openbookqa_pairs_loss": 5.631566047668457, + "eval_openbookqa_pairs_runtime": 0.861, + "eval_openbookqa_pairs_samples_per_second": 148.656, + "eval_openbookqa_pairs_steps_per_second": 2.323, + "step": 20 + }, + { + "epoch": 0.015515903801396431, + "eval_msmarco_pairs_loss": 10.92141342163086, + "eval_msmarco_pairs_runtime": 1.9495, + "eval_msmarco_pairs_samples_per_second": 65.659, + "eval_msmarco_pairs_steps_per_second": 1.026, + "step": 20 + }, + { + "epoch": 0.015515903801396431, + "eval_nq_pairs_loss": 9.838966369628906, + "eval_nq_pairs_runtime": 3.3145, + "eval_nq_pairs_samples_per_second": 38.618, + "eval_nq_pairs_steps_per_second": 0.603, + "step": 20 + }, + { + "epoch": 0.015515903801396431, + "eval_trivia_pairs_loss": 7.321401596069336, + "eval_trivia_pairs_runtime": 4.4673, + "eval_trivia_pairs_samples_per_second": 28.653, + "eval_trivia_pairs_steps_per_second": 0.448, + "step": 20 + }, + { + "epoch": 0.015515903801396431, + "eval_gooaq_pairs_loss": 7.844480037689209, + "eval_gooaq_pairs_runtime": 1.4628, + "eval_gooaq_pairs_samples_per_second": 87.505, + "eval_gooaq_pairs_steps_per_second": 1.367, + "step": 20 + }, + { + "epoch": 0.015515903801396431, + "eval_paws-pos_loss": 2.474581718444824, + "eval_paws-pos_runtime": 1.0227, + "eval_paws-pos_samples_per_second": 125.159, + "eval_paws-pos_steps_per_second": 1.956, + "step": 20 + }, + { + "epoch": 0.015515903801396431, + "eval_global_dataset_loss": 5.820813179016113, + "eval_global_dataset_runtime": 19.3774, + "eval_global_dataset_samples_per_second": 21.468, + "eval_global_dataset_steps_per_second": 0.361, + "step": 20 + }, + { + "epoch": 0.016291698991466253, + "grad_norm": 16.32306671142578, + "learning_rate": 7.81433607520564e-07, + "loss": 6.3011, + "step": 21 + }, + { + "epoch": 0.017067494181536073, + "grad_norm": 10.361891746520996, + "learning_rate": 8.225616921269095e-07, + "loss": 5.6518, + "step": 22 + }, + { + "epoch": 0.017843289371605897, + "grad_norm": 18.645973205566406, + "learning_rate": 8.636897767332549e-07, + "loss": 6.5445, + "step": 23 + }, + { + "epoch": 0.018619084561675717, + "grad_norm": 14.296663284301758, + "learning_rate": 9.048178613396004e-07, + "loss": 6.0762, + "step": 24 + }, + { + "epoch": 0.019394879751745538, + "grad_norm": 12.288186073303223, + "learning_rate": 9.459459459459459e-07, + "loss": 4.9043, + "step": 25 + }, + { + "epoch": 0.02017067494181536, + "grad_norm": 48.58110046386719, + "learning_rate": 9.870740305522914e-07, + "loss": 10.3442, + "step": 26 + }, + { + "epoch": 0.02094647013188518, + "grad_norm": 17.945547103881836, + "learning_rate": 1.0282021151586369e-06, + "loss": 6.3305, + "step": 27 + }, + { + "epoch": 0.021722265321955005, + "grad_norm": 16.731460571289062, + "learning_rate": 1.0693301997649822e-06, + "loss": 6.3366, + "step": 28 + }, + { + "epoch": 0.022498060512024826, + "grad_norm": 10.825992584228516, + "learning_rate": 1.1104582843713277e-06, + "loss": 5.6012, + "step": 29 + }, + { + "epoch": 0.023273855702094646, + "grad_norm": 13.152975082397461, + "learning_rate": 1.1515863689776732e-06, + "loss": 5.9001, + "step": 30 + }, + { + "epoch": 0.02404965089216447, + "grad_norm": 17.892440795898438, + "learning_rate": 1.1927144535840187e-06, + "loss": 6.3815, + "step": 31 + }, + { + "epoch": 0.02482544608223429, + "grad_norm": 10.021342277526855, + "learning_rate": 1.2338425381903642e-06, + "loss": 4.8941, + "step": 32 + }, + { + "epoch": 0.02560124127230411, + "grad_norm": 17.672340393066406, + "learning_rate": 1.2749706227967097e-06, + "loss": 6.2971, + "step": 33 + }, + { + "epoch": 0.026377036462373934, + "grad_norm": 13.532915115356445, + "learning_rate": 1.316098707403055e-06, + "loss": 5.6313, + "step": 34 + }, + { + "epoch": 0.027152831652443754, + "grad_norm": 13.363771438598633, + "learning_rate": 1.3572267920094007e-06, + "loss": 5.6635, + "step": 35 + }, + { + "epoch": 0.027928626842513578, + "grad_norm": 11.951986312866211, + "learning_rate": 1.398354876615746e-06, + "loss": 6.0632, + "step": 36 + }, + { + "epoch": 0.028704422032583398, + "grad_norm": 12.715692520141602, + "learning_rate": 1.4394829612220915e-06, + "loss": 6.0125, + "step": 37 + }, + { + "epoch": 0.02948021722265322, + "grad_norm": 44.77717208862305, + "learning_rate": 1.4806110458284372e-06, + "loss": 7.9108, + "step": 38 + }, + { + "epoch": 0.030256012412723042, + "grad_norm": 16.372045516967773, + "learning_rate": 1.5217391304347825e-06, + "loss": 6.3741, + "step": 39 + }, + { + "epoch": 0.031031807602792862, + "grad_norm": 10.656290054321289, + "learning_rate": 1.562867215041128e-06, + "loss": 5.3981, + "step": 40 + }, + { + "epoch": 0.031031807602792862, + "eval_Qnli-dev_cosine_accuracy": 0.6015625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.9430716633796692, + "eval_Qnli-dev_cosine_ap": 0.5565875119399253, + "eval_Qnli-dev_cosine_f1": 0.6291834002677376, + "eval_Qnli-dev_cosine_f1_threshold": 0.6778229475021362, + "eval_Qnli-dev_cosine_precision": 0.4598825831702544, + "eval_Qnli-dev_cosine_recall": 0.9957627118644068, + "eval_Qnli-dev_dot_accuracy": 0.548828125, + "eval_Qnli-dev_dot_accuracy_threshold": 709.1221923828125, + "eval_Qnli-dev_dot_ap": 0.47478000038083695, + "eval_Qnli-dev_dot_f1": 0.6291834002677376, + "eval_Qnli-dev_dot_f1_threshold": 357.37890625, + "eval_Qnli-dev_dot_precision": 0.4598825831702544, + "eval_Qnli-dev_dot_recall": 0.9957627118644068, + "eval_Qnli-dev_euclidean_accuracy": 0.599609375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 8.403730392456055, + "eval_Qnli-dev_euclidean_ap": 0.5735439175412295, + "eval_Qnli-dev_euclidean_f1": 0.6324324324324324, + "eval_Qnli-dev_euclidean_f1_threshold": 17.377689361572266, + "eval_Qnli-dev_euclidean_precision": 0.4642857142857143, + "eval_Qnli-dev_euclidean_recall": 0.9915254237288136, + "eval_Qnli-dev_manhattan_accuracy": 0.625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 152.123046875, + "eval_Qnli-dev_manhattan_ap": 0.6110278070440811, + "eval_Qnli-dev_manhattan_f1": 0.6291834002677376, + "eval_Qnli-dev_manhattan_f1_threshold": 275.72344970703125, + "eval_Qnli-dev_manhattan_precision": 0.4598825831702544, + "eval_Qnli-dev_manhattan_recall": 0.9957627118644068, + "eval_Qnli-dev_max_accuracy": 0.625, + "eval_Qnli-dev_max_accuracy_threshold": 709.1221923828125, + "eval_Qnli-dev_max_ap": 0.6110278070440811, + "eval_Qnli-dev_max_f1": 0.6324324324324324, + "eval_Qnli-dev_max_f1_threshold": 357.37890625, + "eval_Qnli-dev_max_precision": 0.4642857142857143, + "eval_Qnli-dev_max_recall": 0.9957627118644068, + "eval_allNLI-dev_cosine_accuracy": 0.666015625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9908591508865356, + "eval_allNLI-dev_cosine_ap": 0.3677090943014154, + "eval_allNLI-dev_cosine_f1": 0.5088235294117648, + "eval_allNLI-dev_cosine_f1_threshold": 0.8096739053726196, + "eval_allNLI-dev_cosine_precision": 0.34122287968441817, + "eval_allNLI-dev_cosine_recall": 1.0, + "eval_allNLI-dev_dot_accuracy": 0.66015625, + "eval_allNLI-dev_dot_accuracy_threshold": 1065.9342041015625, + "eval_allNLI-dev_dot_ap": 0.33310258182405583, + "eval_allNLI-dev_dot_f1": 0.5081240768094535, + "eval_allNLI-dev_dot_f1_threshold": 600.9107666015625, + "eval_allNLI-dev_dot_precision": 0.3412698412698413, + "eval_allNLI-dev_dot_recall": 0.9942196531791907, + "eval_allNLI-dev_euclidean_accuracy": 0.666015625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 5.249004364013672, + "eval_allNLI-dev_euclidean_ap": 0.36405301502121135, + "eval_allNLI-dev_euclidean_f1": 0.5103857566765578, + "eval_allNLI-dev_euclidean_f1_threshold": 16.79437828063965, + "eval_allNLI-dev_euclidean_precision": 0.34331337325349304, + "eval_allNLI-dev_euclidean_recall": 0.9942196531791907, + "eval_allNLI-dev_manhattan_accuracy": 0.666015625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 71.94633483886719, + "eval_allNLI-dev_manhattan_ap": 0.3767657897123271, + "eval_allNLI-dev_manhattan_f1": 0.5073746312684366, + "eval_allNLI-dev_manhattan_f1_threshold": 236.0598602294922, + "eval_allNLI-dev_manhattan_precision": 0.3405940594059406, + "eval_allNLI-dev_manhattan_recall": 0.9942196531791907, + "eval_allNLI-dev_max_accuracy": 0.666015625, + "eval_allNLI-dev_max_accuracy_threshold": 1065.9342041015625, + "eval_allNLI-dev_max_ap": 0.3767657897123271, + "eval_allNLI-dev_max_f1": 0.5103857566765578, + "eval_allNLI-dev_max_f1_threshold": 600.9107666015625, + "eval_allNLI-dev_max_precision": 0.34331337325349304, + "eval_allNLI-dev_max_recall": 1.0, + "eval_sequential_score": 0.6110278070440811, + "eval_sts-test_pearson_cosine": 0.14752508082997276, + "eval_sts-test_pearson_dot": 0.13488717804142986, + "eval_sts-test_pearson_euclidean": 0.15185506014664446, + "eval_sts-test_pearson_manhattan": 0.14687917919770682, + "eval_sts-test_pearson_max": 0.15185506014664446, + "eval_sts-test_spearman_cosine": 0.20049724161017018, + "eval_sts-test_spearman_dot": 0.1353592298330889, + "eval_sts-test_spearman_euclidean": 0.17951910330520107, + "eval_sts-test_spearman_manhattan": 0.17376785019622396, + "eval_sts-test_spearman_max": 0.20049724161017018, + "eval_vitaminc-pairs_loss": 4.123702049255371, + "eval_vitaminc-pairs_runtime": 5.4736, + "eval_vitaminc-pairs_samples_per_second": 23.385, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 40 + }, + { + "epoch": 0.031031807602792862, + "eval_negation-triplets_loss": 5.45994758605957, + "eval_negation-triplets_runtime": 1.0747, + "eval_negation-triplets_samples_per_second": 119.102, + "eval_negation-triplets_steps_per_second": 1.861, + "step": 40 + }, + { + "epoch": 0.031031807602792862, + "eval_scitail-pairs-pos_loss": 2.758348226547241, + "eval_scitail-pairs-pos_runtime": 1.2143, + "eval_scitail-pairs-pos_samples_per_second": 105.41, + "eval_scitail-pairs-pos_steps_per_second": 1.647, + "step": 40 + }, + { + "epoch": 0.031031807602792862, + "eval_scitail-pairs-qa_loss": 4.180817127227783, + "eval_scitail-pairs-qa_runtime": 0.766, + "eval_scitail-pairs-qa_samples_per_second": 167.105, + "eval_scitail-pairs-qa_steps_per_second": 2.611, + "step": 40 + }, + { + "epoch": 0.031031807602792862, + "eval_xsum-pairs_loss": 5.874823093414307, + "eval_xsum-pairs_runtime": 3.924, + "eval_xsum-pairs_samples_per_second": 32.62, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 40 + }, + { + "epoch": 0.031031807602792862, + "eval_sciq_pairs_loss": 0.6976423859596252, + "eval_sciq_pairs_runtime": 5.8211, + "eval_sciq_pairs_samples_per_second": 21.989, + "eval_sciq_pairs_steps_per_second": 0.344, + "step": 40 + }, + { + "epoch": 0.031031807602792862, + "eval_qasc_pairs_loss": 7.423002243041992, + "eval_qasc_pairs_runtime": 0.9034, + "eval_qasc_pairs_samples_per_second": 141.693, + "eval_qasc_pairs_steps_per_second": 2.214, + "step": 40 + }, + { + "epoch": 0.031031807602792862, + "eval_openbookqa_pairs_loss": 5.564730644226074, + "eval_openbookqa_pairs_runtime": 0.8569, + "eval_openbookqa_pairs_samples_per_second": 149.376, + "eval_openbookqa_pairs_steps_per_second": 2.334, + "step": 40 + }, + { + "epoch": 0.031031807602792862, + "eval_msmarco_pairs_loss": 10.023717880249023, + "eval_msmarco_pairs_runtime": 1.9602, + "eval_msmarco_pairs_samples_per_second": 65.298, + "eval_msmarco_pairs_steps_per_second": 1.02, + "step": 40 + }, + { + "epoch": 0.031031807602792862, + "eval_nq_pairs_loss": 8.906123161315918, + "eval_nq_pairs_runtime": 3.3287, + "eval_nq_pairs_samples_per_second": 38.453, + "eval_nq_pairs_steps_per_second": 0.601, + "step": 40 + }, + { + "epoch": 0.031031807602792862, + "eval_trivia_pairs_loss": 6.981179714202881, + "eval_trivia_pairs_runtime": 4.4701, + "eval_trivia_pairs_samples_per_second": 28.635, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 40 + }, + { + "epoch": 0.031031807602792862, + "eval_gooaq_pairs_loss": 7.387185096740723, + "eval_gooaq_pairs_runtime": 1.453, + "eval_gooaq_pairs_samples_per_second": 88.091, + "eval_gooaq_pairs_steps_per_second": 1.376, + "step": 40 + }, + { + "epoch": 0.031031807602792862, + "eval_paws-pos_loss": 2.5774118900299072, + "eval_paws-pos_runtime": 1.0228, + "eval_paws-pos_samples_per_second": 125.149, + "eval_paws-pos_steps_per_second": 1.955, + "step": 40 + }, + { + "epoch": 0.031031807602792862, + "eval_global_dataset_loss": 5.574436187744141, + "eval_global_dataset_runtime": 19.3888, + "eval_global_dataset_samples_per_second": 21.456, + "eval_global_dataset_steps_per_second": 0.361, + "step": 40 + }, + { + "epoch": 0.03180760279286268, + "grad_norm": 10.962615966796875, + "learning_rate": 1.6039952996474733e-06, + "loss": 5.3344, + "step": 41 + }, + { + "epoch": 0.032583397982932506, + "grad_norm": 14.594836235046387, + "learning_rate": 1.645123384253819e-06, + "loss": 6.1365, + "step": 42 + }, + { + "epoch": 0.03335919317300233, + "grad_norm": 11.795042037963867, + "learning_rate": 1.6862514688601645e-06, + "loss": 5.818, + "step": 43 + }, + { + "epoch": 0.03413498836307215, + "grad_norm": 12.332319259643555, + "learning_rate": 1.7273795534665098e-06, + "loss": 5.8133, + "step": 44 + }, + { + "epoch": 0.03491078355314197, + "grad_norm": 14.156538963317871, + "learning_rate": 1.7685076380728553e-06, + "loss": 5.9534, + "step": 45 + }, + { + "epoch": 0.035686578743211794, + "grad_norm": 11.36040210723877, + "learning_rate": 1.8096357226792008e-06, + "loss": 5.6302, + "step": 46 + }, + { + "epoch": 0.03646237393328161, + "grad_norm": 12.837271690368652, + "learning_rate": 1.8507638072855463e-06, + "loss": 5.9389, + "step": 47 + }, + { + "epoch": 0.037238169123351435, + "grad_norm": 13.556758880615234, + "learning_rate": 1.8918918918918918e-06, + "loss": 5.7893, + "step": 48 + }, + { + "epoch": 0.03801396431342126, + "grad_norm": 9.904508590698242, + "learning_rate": 1.933019976498237e-06, + "loss": 4.725, + "step": 49 + }, + { + "epoch": 0.038789759503491075, + "grad_norm": 13.982979774475098, + "learning_rate": 1.9741480611045828e-06, + "loss": 6.1877, + "step": 50 + }, + { + "epoch": 0.0395655546935609, + "grad_norm": 13.776800155639648, + "learning_rate": 2.015276145710928e-06, + "loss": 6.001, + "step": 51 + }, + { + "epoch": 0.04034134988363072, + "grad_norm": 21.959096908569336, + "learning_rate": 2.0564042303172738e-06, + "loss": 7.0504, + "step": 52 + }, + { + "epoch": 0.041117145073700546, + "grad_norm": 14.812617301940918, + "learning_rate": 2.097532314923619e-06, + "loss": 5.7706, + "step": 53 + }, + { + "epoch": 0.04189294026377036, + "grad_norm": 13.54603099822998, + "learning_rate": 2.1386603995299644e-06, + "loss": 6.3886, + "step": 54 + }, + { + "epoch": 0.04266873545384019, + "grad_norm": 11.500264167785645, + "learning_rate": 2.17978848413631e-06, + "loss": 6.0503, + "step": 55 + }, + { + "epoch": 0.04344453064391001, + "grad_norm": 7.893326759338379, + "learning_rate": 2.2209165687426554e-06, + "loss": 5.9136, + "step": 56 + }, + { + "epoch": 0.04422032583397983, + "grad_norm": 9.733207702636719, + "learning_rate": 2.2620446533490006e-06, + "loss": 5.6769, + "step": 57 + }, + { + "epoch": 0.04499612102404965, + "grad_norm": 12.644614219665527, + "learning_rate": 2.3031727379553464e-06, + "loss": 5.6418, + "step": 58 + }, + { + "epoch": 0.045771916214119475, + "grad_norm": 8.392154693603516, + "learning_rate": 2.344300822561692e-06, + "loss": 5.3225, + "step": 59 + }, + { + "epoch": 0.04654771140418929, + "grad_norm": 6.115021228790283, + "learning_rate": 2.3854289071680374e-06, + "loss": 5.2977, + "step": 60 + }, + { + "epoch": 0.04654771140418929, + "eval_Qnli-dev_cosine_accuracy": 0.59765625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.9555542469024658, + "eval_Qnli-dev_cosine_ap": 0.5645675307780116, + "eval_Qnli-dev_cosine_f1": 0.6291834002677376, + "eval_Qnli-dev_cosine_f1_threshold": 0.7179017066955566, + "eval_Qnli-dev_cosine_precision": 0.4598825831702544, + "eval_Qnli-dev_cosine_recall": 0.9957627118644068, + "eval_Qnli-dev_dot_accuracy": 0.55078125, + "eval_Qnli-dev_dot_accuracy_threshold": 727.389404296875, + "eval_Qnli-dev_dot_ap": 0.47243395169886154, + "eval_Qnli-dev_dot_f1": 0.629878869448183, + "eval_Qnli-dev_dot_f1_threshold": 461.4835205078125, + "eval_Qnli-dev_dot_precision": 0.46153846153846156, + "eval_Qnli-dev_dot_recall": 0.9915254237288136, + "eval_Qnli-dev_euclidean_accuracy": 0.599609375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 8.283002853393555, + "eval_Qnli-dev_euclidean_ap": 0.5826604533188524, + "eval_Qnli-dev_euclidean_f1": 0.6314363143631436, + "eval_Qnli-dev_euclidean_f1_threshold": 16.081790924072266, + "eval_Qnli-dev_euclidean_precision": 0.4641434262948207, + "eval_Qnli-dev_euclidean_recall": 0.9872881355932204, + "eval_Qnli-dev_manhattan_accuracy": 0.630859375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 146.17575073242188, + "eval_Qnli-dev_manhattan_ap": 0.6168306227629134, + "eval_Qnli-dev_manhattan_f1": 0.6296296296296297, + "eval_Qnli-dev_manhattan_f1_threshold": 204.5635528564453, + "eval_Qnli-dev_manhattan_precision": 0.4742489270386266, + "eval_Qnli-dev_manhattan_recall": 0.9364406779661016, + "eval_Qnli-dev_max_accuracy": 0.630859375, + "eval_Qnli-dev_max_accuracy_threshold": 727.389404296875, + "eval_Qnli-dev_max_ap": 0.6168306227629134, + "eval_Qnli-dev_max_f1": 0.6314363143631436, + "eval_Qnli-dev_max_f1_threshold": 461.4835205078125, + "eval_Qnli-dev_max_precision": 0.4742489270386266, + "eval_Qnli-dev_max_recall": 0.9957627118644068, + "eval_allNLI-dev_cosine_accuracy": 0.666015625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9849413633346558, + "eval_allNLI-dev_cosine_ap": 0.37307202329042877, + "eval_allNLI-dev_cosine_f1": 0.5104477611940298, + "eval_allNLI-dev_cosine_f1_threshold": 0.8623183965682983, + "eval_allNLI-dev_cosine_precision": 0.3440643863179074, + "eval_allNLI-dev_cosine_recall": 0.9884393063583815, + "eval_allNLI-dev_dot_accuracy": 0.66015625, + "eval_allNLI-dev_dot_accuracy_threshold": 1062.3370361328125, + "eval_allNLI-dev_dot_ap": 0.3297699984610426, + "eval_allNLI-dev_dot_f1": 0.5088757396449705, + "eval_allNLI-dev_dot_f1_threshold": 622.9761962890625, + "eval_allNLI-dev_dot_precision": 0.341948310139165, + "eval_allNLI-dev_dot_recall": 0.9942196531791907, + "eval_allNLI-dev_euclidean_accuracy": 0.666015625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 5.127579689025879, + "eval_allNLI-dev_euclidean_ap": 0.36909730374104466, + "eval_allNLI-dev_euclidean_f1": 0.5113464447806354, + "eval_allNLI-dev_euclidean_f1_threshold": 14.735725402832031, + "eval_allNLI-dev_euclidean_precision": 0.3463114754098361, + "eval_allNLI-dev_euclidean_recall": 0.976878612716763, + "eval_allNLI-dev_manhattan_accuracy": 0.6640625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 54.84022521972656, + "eval_allNLI-dev_manhattan_ap": 0.3817636902039587, + "eval_allNLI-dev_manhattan_f1": 0.5103857566765578, + "eval_allNLI-dev_manhattan_f1_threshold": 225.62132263183594, + "eval_allNLI-dev_manhattan_precision": 0.34331337325349304, + "eval_allNLI-dev_manhattan_recall": 0.9942196531791907, + "eval_allNLI-dev_max_accuracy": 0.666015625, + "eval_allNLI-dev_max_accuracy_threshold": 1062.3370361328125, + "eval_allNLI-dev_max_ap": 0.3817636902039587, + "eval_allNLI-dev_max_f1": 0.5113464447806354, + "eval_allNLI-dev_max_f1_threshold": 622.9761962890625, + "eval_allNLI-dev_max_precision": 0.3463114754098361, + "eval_allNLI-dev_max_recall": 0.9942196531791907, + "eval_sequential_score": 0.6168306227629134, + "eval_sts-test_pearson_cosine": 0.15615571514608637, + "eval_sts-test_pearson_dot": 0.1565591340193878, + "eval_sts-test_pearson_euclidean": 0.15789005051166094, + "eval_sts-test_pearson_manhattan": 0.15049180567530787, + "eval_sts-test_pearson_max": 0.15789005051166094, + "eval_sts-test_spearman_cosine": 0.21185036811759986, + "eval_sts-test_spearman_dot": 0.15568507375201698, + "eval_sts-test_spearman_euclidean": 0.1835967487406626, + "eval_sts-test_spearman_manhattan": 0.17528196437414056, + "eval_sts-test_spearman_max": 0.21185036811759986, + "eval_vitaminc-pairs_loss": 3.992605209350586, + "eval_vitaminc-pairs_runtime": 5.4384, + "eval_vitaminc-pairs_samples_per_second": 23.536, + "eval_vitaminc-pairs_steps_per_second": 0.368, + "step": 60 + }, + { + "epoch": 0.04654771140418929, + "eval_negation-triplets_loss": 5.40188455581665, + "eval_negation-triplets_runtime": 1.0636, + "eval_negation-triplets_samples_per_second": 120.348, + "eval_negation-triplets_steps_per_second": 1.88, + "step": 60 + }, + { + "epoch": 0.04654771140418929, + "eval_scitail-pairs-pos_loss": 2.6962366104125977, + "eval_scitail-pairs-pos_runtime": 1.1899, + "eval_scitail-pairs-pos_samples_per_second": 107.569, + "eval_scitail-pairs-pos_steps_per_second": 1.681, + "step": 60 + }, + { + "epoch": 0.04654771140418929, + "eval_scitail-pairs-qa_loss": 4.189697742462158, + "eval_scitail-pairs-qa_runtime": 0.7567, + "eval_scitail-pairs-qa_samples_per_second": 169.153, + "eval_scitail-pairs-qa_steps_per_second": 2.643, + "step": 60 + }, + { + "epoch": 0.04654771140418929, + "eval_xsum-pairs_loss": 5.7225446701049805, + "eval_xsum-pairs_runtime": 3.9013, + "eval_xsum-pairs_samples_per_second": 32.809, + "eval_xsum-pairs_steps_per_second": 0.513, + "step": 60 + }, + { + "epoch": 0.04654771140418929, + "eval_sciq_pairs_loss": 0.6581735610961914, + "eval_sciq_pairs_runtime": 5.7872, + "eval_sciq_pairs_samples_per_second": 22.118, + "eval_sciq_pairs_steps_per_second": 0.346, + "step": 60 + }, + { + "epoch": 0.04654771140418929, + "eval_qasc_pairs_loss": 6.56049919128418, + "eval_qasc_pairs_runtime": 0.8954, + "eval_qasc_pairs_samples_per_second": 142.954, + "eval_qasc_pairs_steps_per_second": 2.234, + "step": 60 + }, + { + "epoch": 0.04654771140418929, + "eval_openbookqa_pairs_loss": 5.458502292633057, + "eval_openbookqa_pairs_runtime": 0.8618, + "eval_openbookqa_pairs_samples_per_second": 148.532, + "eval_openbookqa_pairs_steps_per_second": 2.321, + "step": 60 + }, + { + "epoch": 0.04654771140418929, + "eval_msmarco_pairs_loss": 8.660462379455566, + "eval_msmarco_pairs_runtime": 1.9838, + "eval_msmarco_pairs_samples_per_second": 64.521, + "eval_msmarco_pairs_steps_per_second": 1.008, + "step": 60 + }, + { + "epoch": 0.04654771140418929, + "eval_nq_pairs_loss": 7.642305850982666, + "eval_nq_pairs_runtime": 3.3253, + "eval_nq_pairs_samples_per_second": 38.492, + "eval_nq_pairs_steps_per_second": 0.601, + "step": 60 + }, + { + "epoch": 0.04654771140418929, + "eval_trivia_pairs_loss": 6.512197494506836, + "eval_trivia_pairs_runtime": 4.4671, + "eval_trivia_pairs_samples_per_second": 28.654, + "eval_trivia_pairs_steps_per_second": 0.448, + "step": 60 + }, + { + "epoch": 0.04654771140418929, + "eval_gooaq_pairs_loss": 6.7339911460876465, + "eval_gooaq_pairs_runtime": 1.4577, + "eval_gooaq_pairs_samples_per_second": 87.81, + "eval_gooaq_pairs_steps_per_second": 1.372, + "step": 60 + }, + { + "epoch": 0.04654771140418929, + "eval_paws-pos_loss": 2.735227584838867, + "eval_paws-pos_runtime": 1.0334, + "eval_paws-pos_samples_per_second": 123.864, + "eval_paws-pos_steps_per_second": 1.935, + "step": 60 + }, + { + "epoch": 0.04654771140418929, + "eval_global_dataset_loss": 5.233480453491211, + "eval_global_dataset_runtime": 19.3785, + "eval_global_dataset_samples_per_second": 21.467, + "eval_global_dataset_steps_per_second": 0.361, + "step": 60 + }, + { + "epoch": 0.047323506594259115, + "grad_norm": 17.255535125732422, + "learning_rate": 2.426556991774383e-06, + "loss": 3.5938, + "step": 61 + }, + { + "epoch": 0.04809930178432894, + "grad_norm": 8.553607940673828, + "learning_rate": 2.4676850763807284e-06, + "loss": 6.1306, + "step": 62 + }, + { + "epoch": 0.048875096974398756, + "grad_norm": 29.715600967407227, + "learning_rate": 2.5088131609870737e-06, + "loss": 8.328, + "step": 63 + }, + { + "epoch": 0.04965089216446858, + "grad_norm": 11.584660530090332, + "learning_rate": 2.5499412455934194e-06, + "loss": 6.0765, + "step": 64 + }, + { + "epoch": 0.0504266873545384, + "grad_norm": 9.648879051208496, + "learning_rate": 2.5910693301997647e-06, + "loss": 4.808, + "step": 65 + }, + { + "epoch": 0.05120248254460822, + "grad_norm": 10.802507400512695, + "learning_rate": 2.63219741480611e-06, + "loss": 5.9628, + "step": 66 + }, + { + "epoch": 0.051978277734678044, + "grad_norm": 8.42625904083252, + "learning_rate": 2.6733254994124557e-06, + "loss": 5.197, + "step": 67 + }, + { + "epoch": 0.05275407292474787, + "grad_norm": 18.369491577148438, + "learning_rate": 2.7144535840188014e-06, + "loss": 6.7851, + "step": 68 + }, + { + "epoch": 0.05352986811481769, + "grad_norm": 16.186237335205078, + "learning_rate": 2.7555816686251467e-06, + "loss": 3.5248, + "step": 69 + }, + { + "epoch": 0.05430566330488751, + "grad_norm": 25.963136672973633, + "learning_rate": 2.796709753231492e-06, + "loss": 7.614, + "step": 70 + }, + { + "epoch": 0.05508145849495733, + "grad_norm": 10.458888053894043, + "learning_rate": 2.8378378378378377e-06, + "loss": 5.9919, + "step": 71 + }, + { + "epoch": 0.055857253685027156, + "grad_norm": 22.169750213623047, + "learning_rate": 2.878965922444183e-06, + "loss": 6.5741, + "step": 72 + }, + { + "epoch": 0.05663304887509697, + "grad_norm": 8.526782989501953, + "learning_rate": 2.9200940070505282e-06, + "loss": 5.5377, + "step": 73 + }, + { + "epoch": 0.057408844065166796, + "grad_norm": 17.269418716430664, + "learning_rate": 2.9612220916568744e-06, + "loss": 6.6046, + "step": 74 + }, + { + "epoch": 0.05818463925523662, + "grad_norm": 25.390901565551758, + "learning_rate": 3.0023501762632197e-06, + "loss": 6.6433, + "step": 75 + }, + { + "epoch": 0.05896043444530644, + "grad_norm": 9.188263893127441, + "learning_rate": 3.043478260869565e-06, + "loss": 5.347, + "step": 76 + }, + { + "epoch": 0.05973622963537626, + "grad_norm": 19.360456466674805, + "learning_rate": 3.0846063454759102e-06, + "loss": 6.3361, + "step": 77 + }, + { + "epoch": 0.060512024825446084, + "grad_norm": 16.59157371520996, + "learning_rate": 3.125734430082256e-06, + "loss": 6.672, + "step": 78 + }, + { + "epoch": 0.0612878200155159, + "grad_norm": 21.274341583251953, + "learning_rate": 3.1668625146886012e-06, + "loss": 7.2266, + "step": 79 + }, + { + "epoch": 0.062063615205585725, + "grad_norm": 7.30432653427124, + "learning_rate": 3.2079905992949465e-06, + "loss": 5.2962, + "step": 80 + }, + { + "epoch": 0.062063615205585725, + "eval_Qnli-dev_cosine_accuracy": 0.609375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.9602965116500854, + "eval_Qnli-dev_cosine_ap": 0.5666608532248261, + "eval_Qnli-dev_cosine_f1": 0.6300268096514745, + "eval_Qnli-dev_cosine_f1_threshold": 0.8070105314254761, + "eval_Qnli-dev_cosine_precision": 0.46078431372549017, + "eval_Qnli-dev_cosine_recall": 0.9957627118644068, + "eval_Qnli-dev_dot_accuracy": 0.55078125, + "eval_Qnli-dev_dot_accuracy_threshold": 752.1351318359375, + "eval_Qnli-dev_dot_ap": 0.47532727039762024, + "eval_Qnli-dev_dot_f1": 0.6307277628032345, + "eval_Qnli-dev_dot_f1_threshold": 492.98980712890625, + "eval_Qnli-dev_dot_precision": 0.4624505928853755, + "eval_Qnli-dev_dot_recall": 0.9915254237288136, + "eval_Qnli-dev_euclidean_accuracy": 0.609375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 7.72648811340332, + "eval_Qnli-dev_euclidean_ap": 0.5815079822504383, + "eval_Qnli-dev_euclidean_f1": 0.6300268096514745, + "eval_Qnli-dev_euclidean_f1_threshold": 16.727632522583008, + "eval_Qnli-dev_euclidean_precision": 0.46078431372549017, + "eval_Qnli-dev_euclidean_recall": 0.9957627118644068, + "eval_Qnli-dev_manhattan_accuracy": 0.625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 138.83407592773438, + "eval_Qnli-dev_manhattan_ap": 0.6149910998453122, + "eval_Qnli-dev_manhattan_f1": 0.6338028169014084, + "eval_Qnli-dev_manhattan_f1_threshold": 200.78419494628906, + "eval_Qnli-dev_manhattan_precision": 0.47468354430379744, + "eval_Qnli-dev_manhattan_recall": 0.9533898305084746, + "eval_Qnli-dev_max_accuracy": 0.625, + "eval_Qnli-dev_max_accuracy_threshold": 752.1351318359375, + "eval_Qnli-dev_max_ap": 0.6149910998453122, + "eval_Qnli-dev_max_f1": 0.6338028169014084, + "eval_Qnli-dev_max_f1_threshold": 492.98980712890625, + "eval_Qnli-dev_max_precision": 0.47468354430379744, + "eval_Qnli-dev_max_recall": 0.9957627118644068, + "eval_allNLI-dev_cosine_accuracy": 0.66796875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9806065559387207, + "eval_allNLI-dev_cosine_ap": 0.3777500660247708, + "eval_allNLI-dev_cosine_f1": 0.5105105105105106, + "eval_allNLI-dev_cosine_f1_threshold": 0.8768577575683594, + "eval_allNLI-dev_cosine_precision": 0.3448275862068966, + "eval_allNLI-dev_cosine_recall": 0.9826589595375722, + "eval_allNLI-dev_dot_accuracy": 0.66015625, + "eval_allNLI-dev_dot_accuracy_threshold": 1061.25537109375, + "eval_allNLI-dev_dot_ap": 0.32808659027780296, + "eval_allNLI-dev_dot_f1": 0.5103244837758112, + "eval_allNLI-dev_dot_f1_threshold": 628.6370239257812, + "eval_allNLI-dev_dot_precision": 0.3425742574257426, + "eval_allNLI-dev_dot_recall": 1.0, + "eval_allNLI-dev_euclidean_accuracy": 0.6640625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 2.8695411682128906, + "eval_allNLI-dev_euclidean_ap": 0.3719354048335051, + "eval_allNLI-dev_euclidean_f1": 0.5098634294385432, + "eval_allNLI-dev_euclidean_f1_threshold": 14.270792007446289, + "eval_allNLI-dev_euclidean_precision": 0.345679012345679, + "eval_allNLI-dev_euclidean_recall": 0.9710982658959537, + "eval_allNLI-dev_manhattan_accuracy": 0.6640625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 52.462921142578125, + "eval_allNLI-dev_manhattan_ap": 0.385436980166003, + "eval_allNLI-dev_manhattan_f1": 0.5121212121212121, + "eval_allNLI-dev_manhattan_f1_threshold": 210.9059295654297, + "eval_allNLI-dev_manhattan_precision": 0.3470225872689938, + "eval_allNLI-dev_manhattan_recall": 0.976878612716763, + "eval_allNLI-dev_max_accuracy": 0.66796875, + "eval_allNLI-dev_max_accuracy_threshold": 1061.25537109375, + "eval_allNLI-dev_max_ap": 0.385436980166003, + "eval_allNLI-dev_max_f1": 0.5121212121212121, + "eval_allNLI-dev_max_f1_threshold": 628.6370239257812, + "eval_allNLI-dev_max_precision": 0.3470225872689938, + "eval_allNLI-dev_max_recall": 1.0, + "eval_sequential_score": 0.6149910998453122, + "eval_sts-test_pearson_cosine": 0.16481809744211354, + "eval_sts-test_pearson_dot": 0.1687336929448065, + "eval_sts-test_pearson_euclidean": 0.16621273435380343, + "eval_sts-test_pearson_manhattan": 0.15708095069304162, + "eval_sts-test_pearson_max": 0.1687336929448065, + "eval_sts-test_spearman_cosine": 0.22067965773633885, + "eval_sts-test_spearman_dot": 0.16797399388640222, + "eval_sts-test_spearman_euclidean": 0.19304745437508256, + "eval_sts-test_spearman_manhattan": 0.18393475501795184, + "eval_sts-test_spearman_max": 0.22067965773633885, + "eval_vitaminc-pairs_loss": 3.9390244483947754, + "eval_vitaminc-pairs_runtime": 5.4765, + "eval_vitaminc-pairs_samples_per_second": 23.373, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 80 + }, + { + "epoch": 0.062063615205585725, + "eval_negation-triplets_loss": 5.378707408905029, + "eval_negation-triplets_runtime": 1.0807, + "eval_negation-triplets_samples_per_second": 118.445, + "eval_negation-triplets_steps_per_second": 1.851, + "step": 80 + }, + { + "epoch": 0.062063615205585725, + "eval_scitail-pairs-pos_loss": 2.6731348037719727, + "eval_scitail-pairs-pos_runtime": 1.2466, + "eval_scitail-pairs-pos_samples_per_second": 102.681, + "eval_scitail-pairs-pos_steps_per_second": 1.604, + "step": 80 + }, + { + "epoch": 0.062063615205585725, + "eval_scitail-pairs-qa_loss": 4.19725227355957, + "eval_scitail-pairs-qa_runtime": 0.7667, + "eval_scitail-pairs-qa_samples_per_second": 166.945, + "eval_scitail-pairs-qa_steps_per_second": 2.609, + "step": 80 + }, + { + "epoch": 0.062063615205585725, + "eval_xsum-pairs_loss": 5.497353553771973, + "eval_xsum-pairs_runtime": 3.9143, + "eval_xsum-pairs_samples_per_second": 32.701, + "eval_xsum-pairs_steps_per_second": 0.511, + "step": 80 + }, + { + "epoch": 0.062063615205585725, + "eval_sciq_pairs_loss": 0.6366308927536011, + "eval_sciq_pairs_runtime": 5.8407, + "eval_sciq_pairs_samples_per_second": 21.915, + "eval_sciq_pairs_steps_per_second": 0.342, + "step": 80 + }, + { + "epoch": 0.062063615205585725, + "eval_qasc_pairs_loss": 5.937916278839111, + "eval_qasc_pairs_runtime": 0.9043, + "eval_qasc_pairs_samples_per_second": 141.552, + "eval_qasc_pairs_steps_per_second": 2.212, + "step": 80 + }, + { + "epoch": 0.062063615205585725, + "eval_openbookqa_pairs_loss": 5.383679389953613, + "eval_openbookqa_pairs_runtime": 0.8569, + "eval_openbookqa_pairs_samples_per_second": 149.38, + "eval_openbookqa_pairs_steps_per_second": 2.334, + "step": 80 + }, + { + "epoch": 0.062063615205585725, + "eval_msmarco_pairs_loss": 7.691795825958252, + "eval_msmarco_pairs_runtime": 1.9631, + "eval_msmarco_pairs_samples_per_second": 65.203, + "eval_msmarco_pairs_steps_per_second": 1.019, + "step": 80 + }, + { + "epoch": 0.062063615205585725, + "eval_nq_pairs_loss": 6.876422882080078, + "eval_nq_pairs_runtime": 3.3205, + "eval_nq_pairs_samples_per_second": 38.549, + "eval_nq_pairs_steps_per_second": 0.602, + "step": 80 + }, + { + "epoch": 0.062063615205585725, + "eval_trivia_pairs_loss": 6.1294965744018555, + "eval_trivia_pairs_runtime": 4.4744, + "eval_trivia_pairs_samples_per_second": 28.607, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 80 + }, + { + "epoch": 0.062063615205585725, + "eval_gooaq_pairs_loss": 6.205443859100342, + "eval_gooaq_pairs_runtime": 1.4558, + "eval_gooaq_pairs_samples_per_second": 87.923, + "eval_gooaq_pairs_steps_per_second": 1.374, + "step": 80 + }, + { + "epoch": 0.062063615205585725, + "eval_paws-pos_loss": 2.7808141708374023, + "eval_paws-pos_runtime": 1.0141, + "eval_paws-pos_samples_per_second": 126.225, + "eval_paws-pos_steps_per_second": 1.972, + "step": 80 + }, + { + "epoch": 0.062063615205585725, + "eval_global_dataset_loss": 5.0193328857421875, + "eval_global_dataset_runtime": 19.3707, + "eval_global_dataset_samples_per_second": 21.476, + "eval_global_dataset_steps_per_second": 0.361, + "step": 80 + }, + { + "epoch": 0.06283941039565555, + "grad_norm": 17.17608070373535, + "learning_rate": 3.2491186839012927e-06, + "loss": 6.1576, + "step": 81 + }, + { + "epoch": 0.06361520558572537, + "grad_norm": 22.93490982055664, + "learning_rate": 3.290246768507638e-06, + "loss": 7.3243, + "step": 82 + }, + { + "epoch": 0.0643910007757952, + "grad_norm": 7.930976867675781, + "learning_rate": 3.3313748531139832e-06, + "loss": 5.4146, + "step": 83 + }, + { + "epoch": 0.06516679596586501, + "grad_norm": 9.221315383911133, + "learning_rate": 3.372502937720329e-06, + "loss": 5.8422, + "step": 84 + }, + { + "epoch": 0.06594259115593483, + "grad_norm": 16.351512908935547, + "learning_rate": 3.4136310223266742e-06, + "loss": 3.7286, + "step": 85 + }, + { + "epoch": 0.06671838634600466, + "grad_norm": 6.853544235229492, + "learning_rate": 3.4547591069330195e-06, + "loss": 5.478, + "step": 86 + }, + { + "epoch": 0.06749418153607448, + "grad_norm": 7.5860700607299805, + "learning_rate": 3.495887191539365e-06, + "loss": 5.7321, + "step": 87 + }, + { + "epoch": 0.0682699767261443, + "grad_norm": 16.70624351501465, + "learning_rate": 3.5370152761457105e-06, + "loss": 3.6133, + "step": 88 + }, + { + "epoch": 0.06904577191621412, + "grad_norm": 20.827497482299805, + "learning_rate": 3.5781433607520563e-06, + "loss": 7.1329, + "step": 89 + }, + { + "epoch": 0.06982156710628394, + "grad_norm": 22.313615798950195, + "learning_rate": 3.6192714453584015e-06, + "loss": 7.1766, + "step": 90 + }, + { + "epoch": 0.07059736229635376, + "grad_norm": 12.5264892578125, + "learning_rate": 3.6603995299647473e-06, + "loss": 5.9805, + "step": 91 + }, + { + "epoch": 0.07137315748642359, + "grad_norm": 5.0581955909729, + "learning_rate": 3.7015276145710925e-06, + "loss": 5.797, + "step": 92 + }, + { + "epoch": 0.0721489526764934, + "grad_norm": 11.093539237976074, + "learning_rate": 3.742655699177438e-06, + "loss": 6.2104, + "step": 93 + }, + { + "epoch": 0.07292474786656322, + "grad_norm": 18.35002326965332, + "learning_rate": 3.7837837837837835e-06, + "loss": 6.7904, + "step": 94 + }, + { + "epoch": 0.07370054305663305, + "grad_norm": 4.62692928314209, + "learning_rate": 3.824911868390129e-06, + "loss": 5.6134, + "step": 95 + }, + { + "epoch": 0.07447633824670287, + "grad_norm": 5.769094944000244, + "learning_rate": 3.866039952996474e-06, + "loss": 5.5922, + "step": 96 + }, + { + "epoch": 0.07525213343677269, + "grad_norm": 6.5628461837768555, + "learning_rate": 3.907168037602819e-06, + "loss": 5.0351, + "step": 97 + }, + { + "epoch": 0.07602792862684252, + "grad_norm": 18.365070343017578, + "learning_rate": 3.9482961222091655e-06, + "loss": 7.0271, + "step": 98 + }, + { + "epoch": 0.07680372381691233, + "grad_norm": 12.308566093444824, + "learning_rate": 3.989424206815511e-06, + "loss": 6.2418, + "step": 99 + }, + { + "epoch": 0.07757951900698215, + "grad_norm": 15.452849388122559, + "learning_rate": 4.030552291421856e-06, + "loss": 6.6114, + "step": 100 + }, + { + "epoch": 0.07757951900698215, + "eval_Qnli-dev_cosine_accuracy": 0.6015625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.9667960405349731, + "eval_Qnli-dev_cosine_ap": 0.5703877819293376, + "eval_Qnli-dev_cosine_f1": 0.6317204301075269, + "eval_Qnli-dev_cosine_f1_threshold": 0.8393322229385376, + "eval_Qnli-dev_cosine_precision": 0.4625984251968504, + "eval_Qnli-dev_cosine_recall": 0.9957627118644068, + "eval_Qnli-dev_dot_accuracy": 0.546875, + "eval_Qnli-dev_dot_accuracy_threshold": 798.5418701171875, + "eval_Qnli-dev_dot_ap": 0.47597628614259624, + "eval_Qnli-dev_dot_f1": 0.6307277628032345, + "eval_Qnli-dev_dot_f1_threshold": 552.041015625, + "eval_Qnli-dev_dot_precision": 0.4624505928853755, + "eval_Qnli-dev_dot_recall": 0.9915254237288136, + "eval_Qnli-dev_euclidean_accuracy": 0.603515625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 7.820992469787598, + "eval_Qnli-dev_euclidean_ap": 0.5870442553034873, + "eval_Qnli-dev_euclidean_f1": 0.6300268096514745, + "eval_Qnli-dev_euclidean_f1_threshold": 16.982101440429688, + "eval_Qnli-dev_euclidean_precision": 0.46078431372549017, + "eval_Qnli-dev_euclidean_recall": 0.9957627118644068, + "eval_Qnli-dev_manhattan_accuracy": 0.62109375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 138.17556762695312, + "eval_Qnli-dev_manhattan_ap": 0.613470320794373, + "eval_Qnli-dev_manhattan_f1": 0.6304347826086957, + "eval_Qnli-dev_manhattan_f1_threshold": 210.9376220703125, + "eval_Qnli-dev_manhattan_precision": 0.464, + "eval_Qnli-dev_manhattan_recall": 0.9830508474576272, + "eval_Qnli-dev_max_accuracy": 0.62109375, + "eval_Qnli-dev_max_accuracy_threshold": 798.5418701171875, + "eval_Qnli-dev_max_ap": 0.613470320794373, + "eval_Qnli-dev_max_f1": 0.6317204301075269, + "eval_Qnli-dev_max_f1_threshold": 552.041015625, + "eval_Qnli-dev_max_precision": 0.464, + "eval_Qnli-dev_max_recall": 0.9957627118644068, + "eval_allNLI-dev_cosine_accuracy": 0.669921875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9819352626800537, + "eval_allNLI-dev_cosine_ap": 0.3832808530001383, + "eval_allNLI-dev_cosine_f1": 0.5082706766917294, + "eval_allNLI-dev_cosine_f1_threshold": 0.8935015201568604, + "eval_allNLI-dev_cosine_precision": 0.3434959349593496, + "eval_allNLI-dev_cosine_recall": 0.976878612716763, + "eval_allNLI-dev_dot_accuracy": 0.66015625, + "eval_allNLI-dev_dot_accuracy_threshold": 1097.2392578125, + "eval_allNLI-dev_dot_ap": 0.32480014044358735, + "eval_allNLI-dev_dot_f1": 0.5095729013254787, + "eval_allNLI-dev_dot_f1_threshold": 669.5403442382812, + "eval_allNLI-dev_dot_precision": 0.34189723320158105, + "eval_allNLI-dev_dot_recall": 1.0, + "eval_allNLI-dev_euclidean_accuracy": 0.666015625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 5.188655376434326, + "eval_allNLI-dev_euclidean_ap": 0.3752436345647485, + "eval_allNLI-dev_euclidean_f1": 0.5112781954887218, + "eval_allNLI-dev_euclidean_f1_threshold": 14.686256408691406, + "eval_allNLI-dev_euclidean_precision": 0.34552845528455284, + "eval_allNLI-dev_euclidean_recall": 0.9826589595375722, + "eval_allNLI-dev_manhattan_accuracy": 0.66796875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 81.46263122558594, + "eval_allNLI-dev_manhattan_ap": 0.39261471479691196, + "eval_allNLI-dev_manhattan_f1": 0.513677811550152, + "eval_allNLI-dev_manhattan_f1_threshold": 204.65167236328125, + "eval_allNLI-dev_manhattan_precision": 0.34845360824742266, + "eval_allNLI-dev_manhattan_recall": 0.976878612716763, + "eval_allNLI-dev_max_accuracy": 0.669921875, + "eval_allNLI-dev_max_accuracy_threshold": 1097.2392578125, + "eval_allNLI-dev_max_ap": 0.39261471479691196, + "eval_allNLI-dev_max_f1": 0.513677811550152, + "eval_allNLI-dev_max_f1_threshold": 669.5403442382812, + "eval_allNLI-dev_max_precision": 0.34845360824742266, + "eval_allNLI-dev_max_recall": 1.0, + "eval_sequential_score": 0.613470320794373, + "eval_sts-test_pearson_cosine": 0.1720525202149525, + "eval_sts-test_pearson_dot": 0.17715667152519826, + "eval_sts-test_pearson_euclidean": 0.17331039588378508, + "eval_sts-test_pearson_manhattan": 0.16267378650643669, + "eval_sts-test_pearson_max": 0.17715667152519826, + "eval_sts-test_spearman_cosine": 0.2278064303667034, + "eval_sts-test_spearman_dot": 0.17562364025486388, + "eval_sts-test_spearman_euclidean": 0.19856242082391304, + "eval_sts-test_spearman_manhattan": 0.1882869494420496, + "eval_sts-test_spearman_max": 0.2278064303667034, + "eval_vitaminc-pairs_loss": 3.920011281967163, + "eval_vitaminc-pairs_runtime": 5.4953, + "eval_vitaminc-pairs_samples_per_second": 23.293, + "eval_vitaminc-pairs_steps_per_second": 0.364, + "step": 100 + }, + { + "epoch": 0.07757951900698215, + "eval_negation-triplets_loss": 5.354557991027832, + "eval_negation-triplets_runtime": 1.0836, + "eval_negation-triplets_samples_per_second": 118.12, + "eval_negation-triplets_steps_per_second": 1.846, + "step": 100 + }, + { + "epoch": 0.07757951900698215, + "eval_scitail-pairs-pos_loss": 2.650925636291504, + "eval_scitail-pairs-pos_runtime": 1.2361, + "eval_scitail-pairs-pos_samples_per_second": 103.552, + "eval_scitail-pairs-pos_steps_per_second": 1.618, + "step": 100 + }, + { + "epoch": 0.07757951900698215, + "eval_scitail-pairs-qa_loss": 4.260052680969238, + "eval_scitail-pairs-qa_runtime": 0.7647, + "eval_scitail-pairs-qa_samples_per_second": 167.375, + "eval_scitail-pairs-qa_steps_per_second": 2.615, + "step": 100 + }, + { + "epoch": 0.07757951900698215, + "eval_xsum-pairs_loss": 5.345217227935791, + "eval_xsum-pairs_runtime": 3.917, + "eval_xsum-pairs_samples_per_second": 32.678, + "eval_xsum-pairs_steps_per_second": 0.511, + "step": 100 + }, + { + "epoch": 0.07757951900698215, + "eval_sciq_pairs_loss": 0.6231272220611572, + "eval_sciq_pairs_runtime": 5.8736, + "eval_sciq_pairs_samples_per_second": 21.793, + "eval_sciq_pairs_steps_per_second": 0.341, + "step": 100 + }, + { + "epoch": 0.07757951900698215, + "eval_qasc_pairs_loss": 5.399855136871338, + "eval_qasc_pairs_runtime": 0.9118, + "eval_qasc_pairs_samples_per_second": 140.388, + "eval_qasc_pairs_steps_per_second": 2.194, + "step": 100 + }, + { + "epoch": 0.07757951900698215, + "eval_openbookqa_pairs_loss": 5.2847161293029785, + "eval_openbookqa_pairs_runtime": 0.8603, + "eval_openbookqa_pairs_samples_per_second": 148.787, + "eval_openbookqa_pairs_steps_per_second": 2.325, + "step": 100 + }, + { + "epoch": 0.07757951900698215, + "eval_msmarco_pairs_loss": 6.665152072906494, + "eval_msmarco_pairs_runtime": 1.9686, + "eval_msmarco_pairs_samples_per_second": 65.021, + "eval_msmarco_pairs_steps_per_second": 1.016, + "step": 100 + }, + { + "epoch": 0.07757951900698215, + "eval_nq_pairs_loss": 6.199686050415039, + "eval_nq_pairs_runtime": 3.3173, + "eval_nq_pairs_samples_per_second": 38.585, + "eval_nq_pairs_steps_per_second": 0.603, + "step": 100 + }, + { + "epoch": 0.07757951900698215, + "eval_trivia_pairs_loss": 5.7401909828186035, + "eval_trivia_pairs_runtime": 4.4762, + "eval_trivia_pairs_samples_per_second": 28.595, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 100 + }, + { + "epoch": 0.07757951900698215, + "eval_gooaq_pairs_loss": 5.792109489440918, + "eval_gooaq_pairs_runtime": 1.4527, + "eval_gooaq_pairs_samples_per_second": 88.113, + "eval_gooaq_pairs_steps_per_second": 1.377, + "step": 100 + }, + { + "epoch": 0.07757951900698215, + "eval_paws-pos_loss": 2.8563427925109863, + "eval_paws-pos_runtime": 1.0232, + "eval_paws-pos_samples_per_second": 125.099, + "eval_paws-pos_steps_per_second": 1.955, + "step": 100 + }, + { + "epoch": 0.07757951900698215, + "eval_global_dataset_loss": 4.84829568862915, + "eval_global_dataset_runtime": 19.3977, + "eval_global_dataset_samples_per_second": 21.446, + "eval_global_dataset_steps_per_second": 0.361, + "step": 100 + }, + { + "epoch": 0.07835531419705198, + "grad_norm": 7.275325298309326, + "learning_rate": 4.071680376028202e-06, + "loss": 5.3294, + "step": 101 + }, + { + "epoch": 0.0791311093871218, + "grad_norm": 6.854203224182129, + "learning_rate": 4.1128084606345476e-06, + "loss": 5.0933, + "step": 102 + }, + { + "epoch": 0.07990690457719161, + "grad_norm": 7.766842365264893, + "learning_rate": 4.153936545240893e-06, + "loss": 4.9333, + "step": 103 + }, + { + "epoch": 0.08068269976726145, + "grad_norm": 7.3934478759765625, + "learning_rate": 4.195064629847238e-06, + "loss": 5.1327, + "step": 104 + }, + { + "epoch": 0.08145849495733126, + "grad_norm": 4.127883434295654, + "learning_rate": 4.236192714453583e-06, + "loss": 5.5859, + "step": 105 + }, + { + "epoch": 0.08223429014740109, + "grad_norm": 8.2522554397583, + "learning_rate": 4.277320799059929e-06, + "loss": 4.9192, + "step": 106 + }, + { + "epoch": 0.08301008533747091, + "grad_norm": 16.312501907348633, + "learning_rate": 4.318448883666274e-06, + "loss": 3.8092, + "step": 107 + }, + { + "epoch": 0.08378588052754073, + "grad_norm": 9.926016807556152, + "learning_rate": 4.35957696827262e-06, + "loss": 6.0247, + "step": 108 + }, + { + "epoch": 0.08456167571761056, + "grad_norm": 7.373856544494629, + "learning_rate": 4.400705052878966e-06, + "loss": 5.6878, + "step": 109 + }, + { + "epoch": 0.08533747090768037, + "grad_norm": 7.337434768676758, + "learning_rate": 4.441833137485311e-06, + "loss": 5.0953, + "step": 110 + }, + { + "epoch": 0.08611326609775019, + "grad_norm": 12.865263938903809, + "learning_rate": 4.482961222091657e-06, + "loss": 6.3449, + "step": 111 + }, + { + "epoch": 0.08688906128782002, + "grad_norm": 6.9256591796875, + "learning_rate": 4.524089306698001e-06, + "loss": 4.9109, + "step": 112 + }, + { + "epoch": 0.08766485647788984, + "grad_norm": 4.341299057006836, + "learning_rate": 4.5652173913043474e-06, + "loss": 5.6, + "step": 113 + }, + { + "epoch": 0.08844065166795965, + "grad_norm": 8.670815467834473, + "learning_rate": 4.606345475910693e-06, + "loss": 4.8765, + "step": 114 + }, + { + "epoch": 0.08921644685802949, + "grad_norm": 10.17563247680664, + "learning_rate": 4.647473560517038e-06, + "loss": 5.7062, + "step": 115 + }, + { + "epoch": 0.0899922420480993, + "grad_norm": 7.781512260437012, + "learning_rate": 4.688601645123384e-06, + "loss": 4.7221, + "step": 116 + }, + { + "epoch": 0.09076803723816912, + "grad_norm": 8.361579895019531, + "learning_rate": 4.7297297297297294e-06, + "loss": 4.8871, + "step": 117 + }, + { + "epoch": 0.09154383242823895, + "grad_norm": 9.360175132751465, + "learning_rate": 4.770857814336075e-06, + "loss": 4.7828, + "step": 118 + }, + { + "epoch": 0.09231962761830877, + "grad_norm": 4.893143177032471, + "learning_rate": 4.81198589894242e-06, + "loss": 5.5502, + "step": 119 + }, + { + "epoch": 0.09309542280837858, + "grad_norm": 9.300333976745605, + "learning_rate": 4.853113983548766e-06, + "loss": 4.6165, + "step": 120 + }, + { + "epoch": 0.09309542280837858, + "eval_Qnli-dev_cosine_accuracy": 0.6015625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.9670753479003906, + "eval_Qnli-dev_cosine_ap": 0.5722899415341274, + "eval_Qnli-dev_cosine_f1": 0.6317204301075269, + "eval_Qnli-dev_cosine_f1_threshold": 0.79914391040802, + "eval_Qnli-dev_cosine_precision": 0.4625984251968504, + "eval_Qnli-dev_cosine_recall": 0.9957627118644068, + "eval_Qnli-dev_dot_accuracy": 0.548828125, + "eval_Qnli-dev_dot_accuracy_threshold": 770.8974609375, + "eval_Qnli-dev_dot_ap": 0.47908167649664435, + "eval_Qnli-dev_dot_f1": 0.6317204301075269, + "eval_Qnli-dev_dot_f1_threshold": 503.69384765625, + "eval_Qnli-dev_dot_precision": 0.4625984251968504, + "eval_Qnli-dev_dot_recall": 0.9957627118644068, + "eval_Qnli-dev_euclidean_accuracy": 0.603515625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 8.770221710205078, + "eval_Qnli-dev_euclidean_ap": 0.5921268027422615, + "eval_Qnli-dev_euclidean_f1": 0.6308724832214765, + "eval_Qnli-dev_euclidean_f1_threshold": 16.821897506713867, + "eval_Qnli-dev_euclidean_precision": 0.46168958742632615, + "eval_Qnli-dev_euclidean_recall": 0.9957627118644068, + "eval_Qnli-dev_manhattan_accuracy": 0.625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 144.48236083984375, + "eval_Qnli-dev_manhattan_ap": 0.6088252763279582, + "eval_Qnli-dev_manhattan_f1": 0.6329113924050632, + "eval_Qnli-dev_manhattan_f1_threshold": 209.13174438476562, + "eval_Qnli-dev_manhattan_precision": 0.47368421052631576, + "eval_Qnli-dev_manhattan_recall": 0.9533898305084746, + "eval_Qnli-dev_max_accuracy": 0.625, + "eval_Qnli-dev_max_accuracy_threshold": 770.8974609375, + "eval_Qnli-dev_max_ap": 0.6088252763279582, + "eval_Qnli-dev_max_f1": 0.6329113924050632, + "eval_Qnli-dev_max_f1_threshold": 503.69384765625, + "eval_Qnli-dev_max_precision": 0.47368421052631576, + "eval_Qnli-dev_max_recall": 0.9957627118644068, + "eval_allNLI-dev_cosine_accuracy": 0.6796875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9780304431915283, + "eval_allNLI-dev_cosine_ap": 0.3992786539720722, + "eval_allNLI-dev_cosine_f1": 0.5098634294385432, + "eval_allNLI-dev_cosine_f1_threshold": 0.8867166042327881, + "eval_allNLI-dev_cosine_precision": 0.345679012345679, + "eval_allNLI-dev_cosine_recall": 0.9710982658959537, + "eval_allNLI-dev_dot_accuracy": 0.66015625, + "eval_allNLI-dev_dot_accuracy_threshold": 993.18896484375, + "eval_allNLI-dev_dot_ap": 0.32359114658021904, + "eval_allNLI-dev_dot_f1": 0.5080763582966226, + "eval_allNLI-dev_dot_f1_threshold": 613.2178344726562, + "eval_allNLI-dev_dot_precision": 0.3405511811023622, + "eval_allNLI-dev_dot_recall": 1.0, + "eval_allNLI-dev_euclidean_accuracy": 0.671875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 6.144635200500488, + "eval_allNLI-dev_euclidean_ap": 0.3919696045318126, + "eval_allNLI-dev_euclidean_f1": 0.5120481927710844, + "eval_allNLI-dev_euclidean_f1_threshold": 14.77474594116211, + "eval_allNLI-dev_euclidean_precision": 0.34623217922606925, + "eval_allNLI-dev_euclidean_recall": 0.9826589595375722, + "eval_allNLI-dev_manhattan_accuracy": 0.66796875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 91.93115234375, + "eval_allNLI-dev_manhattan_ap": 0.404528316560946, + "eval_allNLI-dev_manhattan_f1": 0.5082212257100149, + "eval_allNLI-dev_manhattan_f1_threshold": 230.32806396484375, + "eval_allNLI-dev_manhattan_precision": 0.34274193548387094, + "eval_allNLI-dev_manhattan_recall": 0.9826589595375722, + "eval_allNLI-dev_max_accuracy": 0.6796875, + "eval_allNLI-dev_max_accuracy_threshold": 993.18896484375, + "eval_allNLI-dev_max_ap": 0.404528316560946, + "eval_allNLI-dev_max_f1": 0.5120481927710844, + "eval_allNLI-dev_max_f1_threshold": 613.2178344726562, + "eval_allNLI-dev_max_precision": 0.34623217922606925, + "eval_allNLI-dev_max_recall": 1.0, + "eval_sequential_score": 0.6088252763279582, + "eval_sts-test_pearson_cosine": 0.18225596814280462, + "eval_sts-test_pearson_dot": 0.19389246184304787, + "eval_sts-test_pearson_euclidean": 0.18398275691369742, + "eval_sts-test_pearson_manhattan": 0.17222751044724327, + "eval_sts-test_pearson_max": 0.19389246184304787, + "eval_sts-test_spearman_cosine": 0.23913221385342298, + "eval_sts-test_spearman_dot": 0.19319844513101708, + "eval_sts-test_spearman_euclidean": 0.2106999921922726, + "eval_sts-test_spearman_manhattan": 0.1993430170763632, + "eval_sts-test_spearman_max": 0.23913221385342298, + "eval_vitaminc-pairs_loss": 3.8911848068237305, + "eval_vitaminc-pairs_runtime": 5.47, + "eval_vitaminc-pairs_samples_per_second": 23.4, + "eval_vitaminc-pairs_steps_per_second": 0.366, + "step": 120 + }, + { + "epoch": 0.09309542280837858, + "eval_negation-triplets_loss": 5.301531791687012, + "eval_negation-triplets_runtime": 1.0751, + "eval_negation-triplets_samples_per_second": 119.055, + "eval_negation-triplets_steps_per_second": 1.86, + "step": 120 + }, + { + "epoch": 0.09309542280837858, + "eval_scitail-pairs-pos_loss": 2.535888910293579, + "eval_scitail-pairs-pos_runtime": 1.2447, + "eval_scitail-pairs-pos_samples_per_second": 102.834, + "eval_scitail-pairs-pos_steps_per_second": 1.607, + "step": 120 + }, + { + "epoch": 0.09309542280837858, + "eval_scitail-pairs-qa_loss": 4.098081588745117, + "eval_scitail-pairs-qa_runtime": 0.7828, + "eval_scitail-pairs-qa_samples_per_second": 163.513, + "eval_scitail-pairs-qa_steps_per_second": 2.555, + "step": 120 + }, + { + "epoch": 0.09309542280837858, + "eval_xsum-pairs_loss": 5.268539905548096, + "eval_xsum-pairs_runtime": 3.9213, + "eval_xsum-pairs_samples_per_second": 32.642, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 120 + }, + { + "epoch": 0.09309542280837858, + "eval_sciq_pairs_loss": 0.607820987701416, + "eval_sciq_pairs_runtime": 5.8857, + "eval_sciq_pairs_samples_per_second": 21.748, + "eval_sciq_pairs_steps_per_second": 0.34, + "step": 120 + }, + { + "epoch": 0.09309542280837858, + "eval_qasc_pairs_loss": 5.1934356689453125, + "eval_qasc_pairs_runtime": 0.9116, + "eval_qasc_pairs_samples_per_second": 140.413, + "eval_qasc_pairs_steps_per_second": 2.194, + "step": 120 + }, + { + "epoch": 0.09309542280837858, + "eval_openbookqa_pairs_loss": 5.243656158447266, + "eval_openbookqa_pairs_runtime": 0.8604, + "eval_openbookqa_pairs_samples_per_second": 148.766, + "eval_openbookqa_pairs_steps_per_second": 2.324, + "step": 120 + }, + { + "epoch": 0.09309542280837858, + "eval_msmarco_pairs_loss": 6.208409309387207, + "eval_msmarco_pairs_runtime": 1.9688, + "eval_msmarco_pairs_samples_per_second": 65.013, + "eval_msmarco_pairs_steps_per_second": 1.016, + "step": 120 + }, + { + "epoch": 0.09309542280837858, + "eval_nq_pairs_loss": 5.938248157501221, + "eval_nq_pairs_runtime": 3.3259, + "eval_nq_pairs_samples_per_second": 38.486, + "eval_nq_pairs_steps_per_second": 0.601, + "step": 120 + }, + { + "epoch": 0.09309542280837858, + "eval_trivia_pairs_loss": 5.63157844543457, + "eval_trivia_pairs_runtime": 4.4797, + "eval_trivia_pairs_samples_per_second": 28.574, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 120 + }, + { + "epoch": 0.09309542280837858, + "eval_gooaq_pairs_loss": 5.568259239196777, + "eval_gooaq_pairs_runtime": 1.4576, + "eval_gooaq_pairs_samples_per_second": 87.818, + "eval_gooaq_pairs_steps_per_second": 1.372, + "step": 120 + }, + { + "epoch": 0.09309542280837858, + "eval_paws-pos_loss": 2.3947365283966064, + "eval_paws-pos_runtime": 1.0203, + "eval_paws-pos_samples_per_second": 125.448, + "eval_paws-pos_steps_per_second": 1.96, + "step": 120 + }, + { + "epoch": 0.09309542280837858, + "eval_global_dataset_loss": 4.51961088180542, + "eval_global_dataset_runtime": 19.3948, + "eval_global_dataset_samples_per_second": 21.449, + "eval_global_dataset_steps_per_second": 0.361, + "step": 120 + }, + { + "epoch": 0.09387121799844841, + "grad_norm": 8.517334938049316, + "learning_rate": 4.894242068155111e-06, + "loss": 4.892, + "step": 121 + }, + { + "epoch": 0.09464701318851823, + "grad_norm": 11.708131790161133, + "learning_rate": 4.935370152761457e-06, + "loss": 4.6732, + "step": 122 + }, + { + "epoch": 0.09542280837858805, + "grad_norm": 8.222532272338867, + "learning_rate": 4.976498237367803e-06, + "loss": 5.4697, + "step": 123 + }, + { + "epoch": 0.09619860356865788, + "grad_norm": 9.485191345214844, + "learning_rate": 5.017626321974147e-06, + "loss": 4.777, + "step": 124 + }, + { + "epoch": 0.0969743987587277, + "grad_norm": 19.364856719970703, + "learning_rate": 5.0587544065804934e-06, + "loss": 6.3552, + "step": 125 + }, + { + "epoch": 0.09775019394879751, + "grad_norm": 9.71495246887207, + "learning_rate": 5.099882491186839e-06, + "loss": 4.8758, + "step": 126 + }, + { + "epoch": 0.09852598913886734, + "grad_norm": 23.59428596496582, + "learning_rate": 5.141010575793184e-06, + "loss": 3.029, + "step": 127 + }, + { + "epoch": 0.09930178432893716, + "grad_norm": 15.757760047912598, + "learning_rate": 5.182138660399529e-06, + "loss": 5.7908, + "step": 128 + }, + { + "epoch": 0.10007757951900698, + "grad_norm": 20.3161678314209, + "learning_rate": 5.2232667450058754e-06, + "loss": 6.3179, + "step": 129 + }, + { + "epoch": 0.1008533747090768, + "grad_norm": 13.285233497619629, + "learning_rate": 5.26439482961222e-06, + "loss": 5.7515, + "step": 130 + }, + { + "epoch": 0.10162916989914662, + "grad_norm": 11.527393341064453, + "learning_rate": 5.305522914218566e-06, + "loss": 4.4534, + "step": 131 + }, + { + "epoch": 0.10240496508921644, + "grad_norm": 10.770364761352539, + "learning_rate": 5.346650998824911e-06, + "loss": 4.5598, + "step": 132 + }, + { + "epoch": 0.10318076027928627, + "grad_norm": 13.983633995056152, + "learning_rate": 5.387779083431257e-06, + "loss": 4.3048, + "step": 133 + }, + { + "epoch": 0.10395655546935609, + "grad_norm": 24.10936164855957, + "learning_rate": 5.428907168037603e-06, + "loss": 2.7566, + "step": 134 + }, + { + "epoch": 0.10473235065942592, + "grad_norm": 10.636360168457031, + "learning_rate": 5.470035252643947e-06, + "loss": 4.3254, + "step": 135 + }, + { + "epoch": 0.10550814584949574, + "grad_norm": 10.882925987243652, + "learning_rate": 5.511163337250293e-06, + "loss": 4.4835, + "step": 136 + }, + { + "epoch": 0.10628394103956555, + "grad_norm": 23.658288955688477, + "learning_rate": 5.5522914218566394e-06, + "loss": 4.4302, + "step": 137 + }, + { + "epoch": 0.10705973622963538, + "grad_norm": 11.732893943786621, + "learning_rate": 5.593419506462984e-06, + "loss": 4.2167, + "step": 138 + }, + { + "epoch": 0.1078355314197052, + "grad_norm": 15.095183372497559, + "learning_rate": 5.63454759106933e-06, + "loss": 4.3158, + "step": 139 + }, + { + "epoch": 0.10861132660977502, + "grad_norm": 19.106853485107422, + "learning_rate": 5.675675675675675e-06, + "loss": 5.9198, + "step": 140 + }, + { + "epoch": 0.10861132660977502, + "eval_Qnli-dev_cosine_accuracy": 0.6015625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.932540237903595, + "eval_Qnli-dev_cosine_ap": 0.590162641547892, + "eval_Qnli-dev_cosine_f1": 0.6358381502890174, + "eval_Qnli-dev_cosine_f1_threshold": 0.8430850505828857, + "eval_Qnli-dev_cosine_precision": 0.4824561403508772, + "eval_Qnli-dev_cosine_recall": 0.9322033898305084, + "eval_Qnli-dev_dot_accuracy": 0.568359375, + "eval_Qnli-dev_dot_accuracy_threshold": 651.802734375, + "eval_Qnli-dev_dot_ap": 0.4990308476722506, + "eval_Qnli-dev_dot_f1": 0.6318607764390897, + "eval_Qnli-dev_dot_f1_threshold": 359.81072998046875, + "eval_Qnli-dev_dot_precision": 0.461839530332681, + "eval_Qnli-dev_dot_recall": 1.0, + "eval_Qnli-dev_euclidean_accuracy": 0.607421875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 10.347358703613281, + "eval_Qnli-dev_euclidean_ap": 0.5989545722402487, + "eval_Qnli-dev_euclidean_f1": 0.6300268096514745, + "eval_Qnli-dev_euclidean_f1_threshold": 19.583602905273438, + "eval_Qnli-dev_euclidean_precision": 0.46078431372549017, + "eval_Qnli-dev_euclidean_recall": 0.9957627118644068, + "eval_Qnli-dev_manhattan_accuracy": 0.62109375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 161.458251953125, + "eval_Qnli-dev_manhattan_ap": 0.6134087495053078, + "eval_Qnli-dev_manhattan_f1": 0.6406685236768802, + "eval_Qnli-dev_manhattan_f1_threshold": 260.0387268066406, + "eval_Qnli-dev_manhattan_precision": 0.47717842323651455, + "eval_Qnli-dev_manhattan_recall": 0.9745762711864406, + "eval_Qnli-dev_max_accuracy": 0.62109375, + "eval_Qnli-dev_max_accuracy_threshold": 651.802734375, + "eval_Qnli-dev_max_ap": 0.6134087495053078, + "eval_Qnli-dev_max_f1": 0.6406685236768802, + "eval_Qnli-dev_max_f1_threshold": 359.81072998046875, + "eval_Qnli-dev_max_precision": 0.4824561403508772, + "eval_Qnli-dev_max_recall": 1.0, + "eval_allNLI-dev_cosine_accuracy": 0.68359375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.971808135509491, + "eval_allNLI-dev_cosine_ap": 0.41916615465917384, + "eval_allNLI-dev_cosine_f1": 0.5088, + "eval_allNLI-dev_cosine_f1_threshold": 0.8694682121276855, + "eval_allNLI-dev_cosine_precision": 0.35176991150442477, + "eval_allNLI-dev_cosine_recall": 0.9190751445086706, + "eval_allNLI-dev_dot_accuracy": 0.6640625, + "eval_allNLI-dev_dot_accuracy_threshold": 875.8065185546875, + "eval_allNLI-dev_dot_ap": 0.33345963244863186, + "eval_allNLI-dev_dot_f1": 0.5081723625557207, + "eval_allNLI-dev_dot_f1_threshold": 514.4766845703125, + "eval_allNLI-dev_dot_precision": 0.342, + "eval_allNLI-dev_dot_recall": 0.9884393063583815, + "eval_allNLI-dev_euclidean_accuracy": 0.673828125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 6.979422092437744, + "eval_allNLI-dev_euclidean_ap": 0.41272090015066604, + "eval_allNLI-dev_euclidean_f1": 0.5102362204724409, + "eval_allNLI-dev_euclidean_f1_threshold": 14.31092643737793, + "eval_allNLI-dev_euclidean_precision": 0.35064935064935066, + "eval_allNLI-dev_euclidean_recall": 0.9364161849710982, + "eval_allNLI-dev_manhattan_accuracy": 0.673828125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 113.5325698852539, + "eval_allNLI-dev_manhattan_ap": 0.4170134796589544, + "eval_allNLI-dev_manhattan_f1": 0.5065885797950219, + "eval_allNLI-dev_manhattan_f1_threshold": 311.357177734375, + "eval_allNLI-dev_manhattan_precision": 0.3392156862745098, + "eval_allNLI-dev_manhattan_recall": 1.0, + "eval_allNLI-dev_max_accuracy": 0.68359375, + "eval_allNLI-dev_max_accuracy_threshold": 875.8065185546875, + "eval_allNLI-dev_max_ap": 0.41916615465917384, + "eval_allNLI-dev_max_f1": 0.5102362204724409, + "eval_allNLI-dev_max_f1_threshold": 514.4766845703125, + "eval_allNLI-dev_max_precision": 0.35176991150442477, + "eval_allNLI-dev_max_recall": 1.0, + "eval_sequential_score": 0.6134087495053078, + "eval_sts-test_pearson_cosine": 0.19606202486057506, + "eval_sts-test_pearson_dot": 0.20910750938672273, + "eval_sts-test_pearson_euclidean": 0.20204170742350583, + "eval_sts-test_pearson_manhattan": 0.19980066844086258, + "eval_sts-test_pearson_max": 0.20910750938672273, + "eval_sts-test_spearman_cosine": 0.2570204565643615, + "eval_sts-test_spearman_dot": 0.2142791085758339, + "eval_sts-test_spearman_euclidean": 0.22658230436260302, + "eval_sts-test_spearman_manhattan": 0.2257667126237448, + "eval_sts-test_spearman_max": 0.2570204565643615, + "eval_vitaminc-pairs_loss": 3.860930919647217, + "eval_vitaminc-pairs_runtime": 5.4616, + "eval_vitaminc-pairs_samples_per_second": 23.437, + "eval_vitaminc-pairs_steps_per_second": 0.366, + "step": 140 + }, + { + "epoch": 0.10861132660977502, + "eval_negation-triplets_loss": 5.22545051574707, + "eval_negation-triplets_runtime": 1.061, + "eval_negation-triplets_samples_per_second": 120.643, + "eval_negation-triplets_steps_per_second": 1.885, + "step": 140 + }, + { + "epoch": 0.10861132660977502, + "eval_scitail-pairs-pos_loss": 2.2156143188476562, + "eval_scitail-pairs-pos_runtime": 1.2311, + "eval_scitail-pairs-pos_samples_per_second": 103.974, + "eval_scitail-pairs-pos_steps_per_second": 1.625, + "step": 140 + }, + { + "epoch": 0.10861132660977502, + "eval_scitail-pairs-qa_loss": 3.2973973751068115, + "eval_scitail-pairs-qa_runtime": 0.7709, + "eval_scitail-pairs-qa_samples_per_second": 166.038, + "eval_scitail-pairs-qa_steps_per_second": 2.594, + "step": 140 + }, + { + "epoch": 0.10861132660977502, + "eval_xsum-pairs_loss": 5.160762310028076, + "eval_xsum-pairs_runtime": 3.9073, + "eval_xsum-pairs_samples_per_second": 32.759, + "eval_xsum-pairs_steps_per_second": 0.512, + "step": 140 + }, + { + "epoch": 0.10861132660977502, + "eval_sciq_pairs_loss": 0.6155482530593872, + "eval_sciq_pairs_runtime": 5.8124, + "eval_sciq_pairs_samples_per_second": 22.022, + "eval_sciq_pairs_steps_per_second": 0.344, + "step": 140 + }, + { + "epoch": 0.10861132660977502, + "eval_qasc_pairs_loss": 5.0954060554504395, + "eval_qasc_pairs_runtime": 0.9074, + "eval_qasc_pairs_samples_per_second": 141.069, + "eval_qasc_pairs_steps_per_second": 2.204, + "step": 140 + }, + { + "epoch": 0.10861132660977502, + "eval_openbookqa_pairs_loss": 5.480427265167236, + "eval_openbookqa_pairs_runtime": 0.8517, + "eval_openbookqa_pairs_samples_per_second": 150.279, + "eval_openbookqa_pairs_steps_per_second": 2.348, + "step": 140 + }, + { + "epoch": 0.10861132660977502, + "eval_msmarco_pairs_loss": 6.218542575836182, + "eval_msmarco_pairs_runtime": 1.9635, + "eval_msmarco_pairs_samples_per_second": 65.189, + "eval_msmarco_pairs_steps_per_second": 1.019, + "step": 140 + }, + { + "epoch": 0.10861132660977502, + "eval_nq_pairs_loss": 5.9674296379089355, + "eval_nq_pairs_runtime": 3.3526, + "eval_nq_pairs_samples_per_second": 38.179, + "eval_nq_pairs_steps_per_second": 0.597, + "step": 140 + }, + { + "epoch": 0.10861132660977502, + "eval_trivia_pairs_loss": 6.06894063949585, + "eval_trivia_pairs_runtime": 4.4755, + "eval_trivia_pairs_samples_per_second": 28.6, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 140 + }, + { + "epoch": 0.10861132660977502, + "eval_gooaq_pairs_loss": 5.648138523101807, + "eval_gooaq_pairs_runtime": 1.4628, + "eval_gooaq_pairs_samples_per_second": 87.504, + "eval_gooaq_pairs_steps_per_second": 1.367, + "step": 140 + }, + { + "epoch": 0.10861132660977502, + "eval_paws-pos_loss": 1.3320106267929077, + "eval_paws-pos_runtime": 1.0268, + "eval_paws-pos_samples_per_second": 124.656, + "eval_paws-pos_steps_per_second": 1.948, + "step": 140 + }, + { + "epoch": 0.10861132660977502, + "eval_global_dataset_loss": 4.069947242736816, + "eval_global_dataset_runtime": 19.4022, + "eval_global_dataset_samples_per_second": 21.441, + "eval_global_dataset_steps_per_second": 0.361, + "step": 140 + }, + { + "epoch": 0.10938712179984485, + "grad_norm": 17.907608032226562, + "learning_rate": 5.716803760282021e-06, + "loss": 5.8619, + "step": 141 + }, + { + "epoch": 0.11016291698991466, + "grad_norm": 12.681406021118164, + "learning_rate": 5.757931844888366e-06, + "loss": 4.42, + "step": 142 + }, + { + "epoch": 0.11093871217998448, + "grad_norm": 29.748260498046875, + "learning_rate": 5.799059929494712e-06, + "loss": 6.5458, + "step": 143 + }, + { + "epoch": 0.11171450737005431, + "grad_norm": 17.36277198791504, + "learning_rate": 5.8401880141010565e-06, + "loss": 3.7463, + "step": 144 + }, + { + "epoch": 0.11249030256012413, + "grad_norm": 15.745108604431152, + "learning_rate": 5.881316098707403e-06, + "loss": 4.8236, + "step": 145 + }, + { + "epoch": 0.11326609775019394, + "grad_norm": 17.49993896484375, + "learning_rate": 5.922444183313749e-06, + "loss": 5.9082, + "step": 146 + }, + { + "epoch": 0.11404189294026378, + "grad_norm": 11.602448463439941, + "learning_rate": 5.963572267920093e-06, + "loss": 3.9276, + "step": 147 + }, + { + "epoch": 0.11481768813033359, + "grad_norm": 14.728039741516113, + "learning_rate": 6.004700352526439e-06, + "loss": 4.3073, + "step": 148 + }, + { + "epoch": 0.11559348332040341, + "grad_norm": 12.118491172790527, + "learning_rate": 6.045828437132785e-06, + "loss": 4.06, + "step": 149 + }, + { + "epoch": 0.11636927851047324, + "grad_norm": 14.151768684387207, + "learning_rate": 6.08695652173913e-06, + "loss": 4.7036, + "step": 150 + }, + { + "epoch": 0.11714507370054306, + "grad_norm": 12.606175422668457, + "learning_rate": 6.128084606345475e-06, + "loss": 3.9509, + "step": 151 + }, + { + "epoch": 0.11792086889061287, + "grad_norm": 38.992183685302734, + "learning_rate": 6.1692126909518205e-06, + "loss": 6.3894, + "step": 152 + }, + { + "epoch": 0.1186966640806827, + "grad_norm": 13.61286449432373, + "learning_rate": 6.210340775558167e-06, + "loss": 4.4063, + "step": 153 + }, + { + "epoch": 0.11947245927075252, + "grad_norm": 17.750974655151367, + "learning_rate": 6.251468860164512e-06, + "loss": 3.6723, + "step": 154 + }, + { + "epoch": 0.12024825446082234, + "grad_norm": 17.65484046936035, + "learning_rate": 6.292596944770857e-06, + "loss": 5.462, + "step": 155 + }, + { + "epoch": 0.12102404965089217, + "grad_norm": 12.64035415649414, + "learning_rate": 6.3337250293772025e-06, + "loss": 4.579, + "step": 156 + }, + { + "epoch": 0.12179984484096198, + "grad_norm": 11.979784965515137, + "learning_rate": 6.374853113983549e-06, + "loss": 3.9228, + "step": 157 + }, + { + "epoch": 0.1225756400310318, + "grad_norm": 12.76739501953125, + "learning_rate": 6.415981198589893e-06, + "loss": 3.6178, + "step": 158 + }, + { + "epoch": 0.12335143522110163, + "grad_norm": 27.8046817779541, + "learning_rate": 6.457109283196239e-06, + "loss": 5.686, + "step": 159 + }, + { + "epoch": 0.12412723041117145, + "grad_norm": 12.14214038848877, + "learning_rate": 6.498237367802585e-06, + "loss": 3.5491, + "step": 160 + }, + { + "epoch": 0.12412723041117145, + "eval_Qnli-dev_cosine_accuracy": 0.611328125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.9235547780990601, + "eval_Qnli-dev_cosine_ap": 0.5831965528167927, + "eval_Qnli-dev_cosine_f1": 0.6318758815232722, + "eval_Qnli-dev_cosine_f1_threshold": 0.8174055218696594, + "eval_Qnli-dev_cosine_precision": 0.47357293868921774, + "eval_Qnli-dev_cosine_recall": 0.9491525423728814, + "eval_Qnli-dev_dot_accuracy": 0.564453125, + "eval_Qnli-dev_dot_accuracy_threshold": 650.8389892578125, + "eval_Qnli-dev_dot_ap": 0.5067183783295084, + "eval_Qnli-dev_dot_f1": 0.6327077747989276, + "eval_Qnli-dev_dot_f1_threshold": 399.63836669921875, + "eval_Qnli-dev_dot_precision": 0.4627450980392157, + "eval_Qnli-dev_dot_recall": 1.0, + "eval_Qnli-dev_euclidean_accuracy": 0.6171875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 10.054267883300781, + "eval_Qnli-dev_euclidean_ap": 0.5948450213638221, + "eval_Qnli-dev_euclidean_f1": 0.6318607764390897, + "eval_Qnli-dev_euclidean_f1_threshold": 20.041950225830078, + "eval_Qnli-dev_euclidean_precision": 0.461839530332681, + "eval_Qnli-dev_euclidean_recall": 1.0, + "eval_Qnli-dev_manhattan_accuracy": 0.625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 173.52857971191406, + "eval_Qnli-dev_manhattan_ap": 0.6122239571226092, + "eval_Qnli-dev_manhattan_f1": 0.6353591160220995, + "eval_Qnli-dev_manhattan_f1_threshold": 274.06744384765625, + "eval_Qnli-dev_manhattan_precision": 0.4713114754098361, + "eval_Qnli-dev_manhattan_recall": 0.9745762711864406, + "eval_Qnli-dev_max_accuracy": 0.625, + "eval_Qnli-dev_max_accuracy_threshold": 650.8389892578125, + "eval_Qnli-dev_max_ap": 0.6122239571226092, + "eval_Qnli-dev_max_f1": 0.6353591160220995, + "eval_Qnli-dev_max_f1_threshold": 399.63836669921875, + "eval_Qnli-dev_max_precision": 0.47357293868921774, + "eval_Qnli-dev_max_recall": 1.0, + "eval_allNLI-dev_cosine_accuracy": 0.6796875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9759693145751953, + "eval_allNLI-dev_cosine_ap": 0.42359590129311003, + "eval_allNLI-dev_cosine_f1": 0.5105740181268882, + "eval_allNLI-dev_cosine_f1_threshold": 0.8282982110977173, + "eval_allNLI-dev_cosine_precision": 0.3456032719836401, + "eval_allNLI-dev_cosine_recall": 0.976878612716763, + "eval_allNLI-dev_dot_accuracy": 0.66796875, + "eval_allNLI-dev_dot_accuracy_threshold": 799.5001831054688, + "eval_allNLI-dev_dot_ap": 0.3510979636740506, + "eval_allNLI-dev_dot_f1": 0.5089820359281437, + "eval_allNLI-dev_dot_f1_threshold": 532.1497802734375, + "eval_allNLI-dev_dot_precision": 0.3434343434343434, + "eval_allNLI-dev_dot_recall": 0.9826589595375722, + "eval_allNLI-dev_euclidean_accuracy": 0.673828125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 6.110721588134766, + "eval_allNLI-dev_euclidean_ap": 0.4178547437007212, + "eval_allNLI-dev_euclidean_f1": 0.519134775374376, + "eval_allNLI-dev_euclidean_f1_threshold": 13.405746459960938, + "eval_allNLI-dev_euclidean_precision": 0.3644859813084112, + "eval_allNLI-dev_euclidean_recall": 0.9017341040462428, + "eval_allNLI-dev_manhattan_accuracy": 0.67578125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 103.17061614990234, + "eval_allNLI-dev_manhattan_ap": 0.4272300717524159, + "eval_allNLI-dev_manhattan_f1": 0.5107033639143731, + "eval_allNLI-dev_manhattan_f1_threshold": 245.95425415039062, + "eval_allNLI-dev_manhattan_precision": 0.3471933471933472, + "eval_allNLI-dev_manhattan_recall": 0.9653179190751445, + "eval_allNLI-dev_max_accuracy": 0.6796875, + "eval_allNLI-dev_max_accuracy_threshold": 799.5001831054688, + "eval_allNLI-dev_max_ap": 0.4272300717524159, + "eval_allNLI-dev_max_f1": 0.519134775374376, + "eval_allNLI-dev_max_f1_threshold": 532.1497802734375, + "eval_allNLI-dev_max_precision": 0.3644859813084112, + "eval_allNLI-dev_max_recall": 0.9826589595375722, + "eval_sequential_score": 0.6122239571226092, + "eval_sts-test_pearson_cosine": 0.1983814843936258, + "eval_sts-test_pearson_dot": 0.21567697770934743, + "eval_sts-test_pearson_euclidean": 0.2076928261154226, + "eval_sts-test_pearson_manhattan": 0.2134685079189917, + "eval_sts-test_pearson_max": 0.21567697770934743, + "eval_sts-test_spearman_cosine": 0.2667653122399071, + "eval_sts-test_spearman_dot": 0.21757047434864857, + "eval_sts-test_spearman_euclidean": 0.2401902679264998, + "eval_sts-test_spearman_manhattan": 0.24235085882647658, + "eval_sts-test_spearman_max": 0.2667653122399071, + "eval_vitaminc-pairs_loss": 3.85381817817688, + "eval_vitaminc-pairs_runtime": 5.4462, + "eval_vitaminc-pairs_samples_per_second": 23.503, + "eval_vitaminc-pairs_steps_per_second": 0.367, + "step": 160 + }, + { + "epoch": 0.12412723041117145, + "eval_negation-triplets_loss": 5.059137344360352, + "eval_negation-triplets_runtime": 1.0529, + "eval_negation-triplets_samples_per_second": 121.569, + "eval_negation-triplets_steps_per_second": 1.9, + "step": 160 + }, + { + "epoch": 0.12412723041117145, + "eval_scitail-pairs-pos_loss": 1.9513375759124756, + "eval_scitail-pairs-pos_runtime": 1.2192, + "eval_scitail-pairs-pos_samples_per_second": 104.99, + "eval_scitail-pairs-pos_steps_per_second": 1.64, + "step": 160 + }, + { + "epoch": 0.12412723041117145, + "eval_scitail-pairs-qa_loss": 2.7738723754882812, + "eval_scitail-pairs-qa_runtime": 0.7658, + "eval_scitail-pairs-qa_samples_per_second": 167.149, + "eval_scitail-pairs-qa_steps_per_second": 2.612, + "step": 160 + }, + { + "epoch": 0.12412723041117145, + "eval_xsum-pairs_loss": 4.9543375968933105, + "eval_xsum-pairs_runtime": 3.9138, + "eval_xsum-pairs_samples_per_second": 32.705, + "eval_xsum-pairs_steps_per_second": 0.511, + "step": 160 + }, + { + "epoch": 0.12412723041117145, + "eval_sciq_pairs_loss": 0.5856431722640991, + "eval_sciq_pairs_runtime": 5.846, + "eval_sciq_pairs_samples_per_second": 21.895, + "eval_sciq_pairs_steps_per_second": 0.342, + "step": 160 + }, + { + "epoch": 0.12412723041117145, + "eval_qasc_pairs_loss": 4.644654750823975, + "eval_qasc_pairs_runtime": 0.9037, + "eval_qasc_pairs_samples_per_second": 141.642, + "eval_qasc_pairs_steps_per_second": 2.213, + "step": 160 + }, + { + "epoch": 0.12412723041117145, + "eval_openbookqa_pairs_loss": 5.493581771850586, + "eval_openbookqa_pairs_runtime": 0.8617, + "eval_openbookqa_pairs_samples_per_second": 148.55, + "eval_openbookqa_pairs_steps_per_second": 2.321, + "step": 160 + }, + { + "epoch": 0.12412723041117145, + "eval_msmarco_pairs_loss": 5.456345081329346, + "eval_msmarco_pairs_runtime": 1.9747, + "eval_msmarco_pairs_samples_per_second": 64.82, + "eval_msmarco_pairs_steps_per_second": 1.013, + "step": 160 + }, + { + "epoch": 0.12412723041117145, + "eval_nq_pairs_loss": 5.471042633056641, + "eval_nq_pairs_runtime": 3.332, + "eval_nq_pairs_samples_per_second": 38.416, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 160 + }, + { + "epoch": 0.12412723041117145, + "eval_trivia_pairs_loss": 5.90379524230957, + "eval_trivia_pairs_runtime": 4.4746, + "eval_trivia_pairs_samples_per_second": 28.606, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 160 + }, + { + "epoch": 0.12412723041117145, + "eval_gooaq_pairs_loss": 5.094766139984131, + "eval_gooaq_pairs_runtime": 1.4626, + "eval_gooaq_pairs_samples_per_second": 87.517, + "eval_gooaq_pairs_steps_per_second": 1.367, + "step": 160 + }, + { + "epoch": 0.12412723041117145, + "eval_paws-pos_loss": 1.0914796590805054, + "eval_paws-pos_runtime": 1.0277, + "eval_paws-pos_samples_per_second": 124.555, + "eval_paws-pos_steps_per_second": 1.946, + "step": 160 + }, + { + "epoch": 0.12412723041117145, + "eval_global_dataset_loss": 3.7029871940612793, + "eval_global_dataset_runtime": 19.3902, + "eval_global_dataset_samples_per_second": 21.454, + "eval_global_dataset_steps_per_second": 0.361, + "step": 160 + }, + { + "epoch": 0.12490302560124127, + "grad_norm": 12.19548511505127, + "learning_rate": 6.53936545240893e-06, + "loss": 4.0703, + "step": 161 + }, + { + "epoch": 0.1256788207913111, + "grad_norm": 11.81810188293457, + "learning_rate": 6.580493537015276e-06, + "loss": 3.7131, + "step": 162 + }, + { + "epoch": 0.1264546159813809, + "grad_norm": 12.847527503967285, + "learning_rate": 6.621621621621621e-06, + "loss": 3.8675, + "step": 163 + }, + { + "epoch": 0.12723041117145073, + "grad_norm": 12.332374572753906, + "learning_rate": 6.6627497062279665e-06, + "loss": 3.6557, + "step": 164 + }, + { + "epoch": 0.12800620636152055, + "grad_norm": 15.913554191589355, + "learning_rate": 6.703877790834312e-06, + "loss": 5.5114, + "step": 165 + }, + { + "epoch": 0.1287820015515904, + "grad_norm": 12.6423978805542, + "learning_rate": 6.745005875440658e-06, + "loss": 3.5924, + "step": 166 + }, + { + "epoch": 0.1295577967416602, + "grad_norm": 11.64156723022461, + "learning_rate": 6.786133960047002e-06, + "loss": 3.7331, + "step": 167 + }, + { + "epoch": 0.13033359193173003, + "grad_norm": 14.570131301879883, + "learning_rate": 6.8272620446533485e-06, + "loss": 5.2668, + "step": 168 + }, + { + "epoch": 0.13110938712179984, + "grad_norm": 13.375134468078613, + "learning_rate": 6.868390129259695e-06, + "loss": 3.5033, + "step": 169 + }, + { + "epoch": 0.13188518231186966, + "grad_norm": 13.718034744262695, + "learning_rate": 6.909518213866039e-06, + "loss": 3.9921, + "step": 170 + }, + { + "epoch": 0.13266097750193948, + "grad_norm": 12.68748950958252, + "learning_rate": 6.950646298472385e-06, + "loss": 3.3935, + "step": 171 + }, + { + "epoch": 0.13343677269200932, + "grad_norm": 18.421518325805664, + "learning_rate": 6.99177438307873e-06, + "loss": 4.9198, + "step": 172 + }, + { + "epoch": 0.13421256788207914, + "grad_norm": 21.028749465942383, + "learning_rate": 7.032902467685076e-06, + "loss": 1.3145, + "step": 173 + }, + { + "epoch": 0.13498836307214895, + "grad_norm": 21.025897979736328, + "learning_rate": 7.074030552291421e-06, + "loss": 5.0971, + "step": 174 + }, + { + "epoch": 0.13576415826221877, + "grad_norm": 16.524944305419922, + "learning_rate": 7.115158636897766e-06, + "loss": 4.7722, + "step": 175 + }, + { + "epoch": 0.1365399534522886, + "grad_norm": 15.925724029541016, + "learning_rate": 7.1562867215041125e-06, + "loss": 4.6407, + "step": 176 + }, + { + "epoch": 0.1373157486423584, + "grad_norm": 19.057390213012695, + "learning_rate": 7.197414806110458e-06, + "loss": 5.1164, + "step": 177 + }, + { + "epoch": 0.13809154383242825, + "grad_norm": 15.260638236999512, + "learning_rate": 7.238542890716803e-06, + "loss": 3.751, + "step": 178 + }, + { + "epoch": 0.13886733902249807, + "grad_norm": 30.054716110229492, + "learning_rate": 7.279670975323148e-06, + "loss": 5.4906, + "step": 179 + }, + { + "epoch": 0.13964313421256788, + "grad_norm": 25.419301986694336, + "learning_rate": 7.3207990599294945e-06, + "loss": 5.1472, + "step": 180 + }, + { + "epoch": 0.13964313421256788, + "eval_Qnli-dev_cosine_accuracy": 0.615234375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.9118244647979736, + "eval_Qnli-dev_cosine_ap": 0.5757630053560099, + "eval_Qnli-dev_cosine_f1": 0.6347469220246238, + "eval_Qnli-dev_cosine_f1_threshold": 0.708939254283905, + "eval_Qnli-dev_cosine_precision": 0.4686868686868687, + "eval_Qnli-dev_cosine_recall": 0.9830508474576272, + "eval_Qnli-dev_dot_accuracy": 0.57421875, + "eval_Qnli-dev_dot_accuracy_threshold": 521.3780517578125, + "eval_Qnli-dev_dot_ap": 0.5092251488391427, + "eval_Qnli-dev_dot_f1": 0.6317204301075269, + "eval_Qnli-dev_dot_f1_threshold": 351.0001220703125, + "eval_Qnli-dev_dot_precision": 0.4625984251968504, + "eval_Qnli-dev_dot_recall": 0.9957627118644068, + "eval_Qnli-dev_euclidean_accuracy": 0.609375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 10.459676742553711, + "eval_Qnli-dev_euclidean_ap": 0.5817988420936584, + "eval_Qnli-dev_euclidean_f1": 0.6318607764390897, + "eval_Qnli-dev_euclidean_f1_threshold": 21.422916412353516, + "eval_Qnli-dev_euclidean_precision": 0.461839530332681, + "eval_Qnli-dev_euclidean_recall": 1.0, + "eval_Qnli-dev_manhattan_accuracy": 0.619140625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 212.72662353515625, + "eval_Qnli-dev_manhattan_ap": 0.5901977757397192, + "eval_Qnli-dev_manhattan_f1": 0.6324324324324324, + "eval_Qnli-dev_manhattan_f1_threshold": 349.28887939453125, + "eval_Qnli-dev_manhattan_precision": 0.4642857142857143, + "eval_Qnli-dev_manhattan_recall": 0.9915254237288136, + "eval_Qnli-dev_max_accuracy": 0.619140625, + "eval_Qnli-dev_max_accuracy_threshold": 521.3780517578125, + "eval_Qnli-dev_max_ap": 0.5901977757397192, + "eval_Qnli-dev_max_f1": 0.6347469220246238, + "eval_Qnli-dev_max_f1_threshold": 351.0001220703125, + "eval_Qnli-dev_max_precision": 0.4686868686868687, + "eval_Qnli-dev_max_recall": 1.0, + "eval_allNLI-dev_cosine_accuracy": 0.681640625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9692702293395996, + "eval_allNLI-dev_cosine_ap": 0.43588706986821285, + "eval_allNLI-dev_cosine_f1": 0.5211505922165821, + "eval_allNLI-dev_cosine_f1_threshold": 0.8656014800071716, + "eval_allNLI-dev_cosine_precision": 0.3684210526315789, + "eval_allNLI-dev_cosine_recall": 0.8901734104046243, + "eval_allNLI-dev_dot_accuracy": 0.66796875, + "eval_allNLI-dev_dot_accuracy_threshold": 785.7974243164062, + "eval_allNLI-dev_dot_ap": 0.36610448753393265, + "eval_allNLI-dev_dot_f1": 0.5118343195266272, + "eval_allNLI-dev_dot_f1_threshold": 447.95782470703125, + "eval_allNLI-dev_dot_precision": 0.34393638170974156, + "eval_allNLI-dev_dot_recall": 1.0, + "eval_allNLI-dev_euclidean_accuracy": 0.677734375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 6.728440284729004, + "eval_allNLI-dev_euclidean_ap": 0.4317993856992305, + "eval_allNLI-dev_euclidean_f1": 0.5214626391096979, + "eval_allNLI-dev_euclidean_f1_threshold": 15.216902732849121, + "eval_allNLI-dev_euclidean_precision": 0.35964912280701755, + "eval_allNLI-dev_euclidean_recall": 0.9479768786127167, + "eval_allNLI-dev_manhattan_accuracy": 0.673828125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 101.38760375976562, + "eval_allNLI-dev_manhattan_ap": 0.43314824920268913, + "eval_allNLI-dev_manhattan_f1": 0.5157421289355323, + "eval_allNLI-dev_manhattan_f1_threshold": 286.70037841796875, + "eval_allNLI-dev_manhattan_precision": 0.3481781376518219, + "eval_allNLI-dev_manhattan_recall": 0.9942196531791907, + "eval_allNLI-dev_max_accuracy": 0.681640625, + "eval_allNLI-dev_max_accuracy_threshold": 785.7974243164062, + "eval_allNLI-dev_max_ap": 0.43588706986821285, + "eval_allNLI-dev_max_f1": 0.5214626391096979, + "eval_allNLI-dev_max_f1_threshold": 447.95782470703125, + "eval_allNLI-dev_max_precision": 0.3684210526315789, + "eval_allNLI-dev_max_recall": 1.0, + "eval_sequential_score": 0.5901977757397192, + "eval_sts-test_pearson_cosine": 0.20493152812453436, + "eval_sts-test_pearson_dot": 0.21775282373018834, + "eval_sts-test_pearson_euclidean": 0.2301348094949043, + "eval_sts-test_pearson_manhattan": 0.23840753470415393, + "eval_sts-test_pearson_max": 0.23840753470415393, + "eval_sts-test_spearman_cosine": 0.28828077894040277, + "eval_sts-test_spearman_dot": 0.23061179744080743, + "eval_sts-test_spearman_euclidean": 0.26600149771047865, + "eval_sts-test_spearman_manhattan": 0.26965490118794977, + "eval_sts-test_spearman_max": 0.28828077894040277, + "eval_vitaminc-pairs_loss": 3.7227895259857178, + "eval_vitaminc-pairs_runtime": 5.4557, + "eval_vitaminc-pairs_samples_per_second": 23.462, + "eval_vitaminc-pairs_steps_per_second": 0.367, + "step": 180 + }, + { + "epoch": 0.13964313421256788, + "eval_negation-triplets_loss": 4.880922794342041, + "eval_negation-triplets_runtime": 1.0642, + "eval_negation-triplets_samples_per_second": 120.279, + "eval_negation-triplets_steps_per_second": 1.879, + "step": 180 + }, + { + "epoch": 0.13964313421256788, + "eval_scitail-pairs-pos_loss": 1.6082611083984375, + "eval_scitail-pairs-pos_runtime": 1.2328, + "eval_scitail-pairs-pos_samples_per_second": 103.83, + "eval_scitail-pairs-pos_steps_per_second": 1.622, + "step": 180 + }, + { + "epoch": 0.13964313421256788, + "eval_scitail-pairs-qa_loss": 2.3117823600769043, + "eval_scitail-pairs-qa_runtime": 0.763, + "eval_scitail-pairs-qa_samples_per_second": 167.752, + "eval_scitail-pairs-qa_steps_per_second": 2.621, + "step": 180 + }, + { + "epoch": 0.13964313421256788, + "eval_xsum-pairs_loss": 4.128580093383789, + "eval_xsum-pairs_runtime": 3.9087, + "eval_xsum-pairs_samples_per_second": 32.748, + "eval_xsum-pairs_steps_per_second": 0.512, + "step": 180 + }, + { + "epoch": 0.13964313421256788, + "eval_sciq_pairs_loss": 0.4913565516471863, + "eval_sciq_pairs_runtime": 5.8016, + "eval_sciq_pairs_samples_per_second": 22.063, + "eval_sciq_pairs_steps_per_second": 0.345, + "step": 180 + }, + { + "epoch": 0.13964313421256788, + "eval_qasc_pairs_loss": 4.587019920349121, + "eval_qasc_pairs_runtime": 0.8933, + "eval_qasc_pairs_samples_per_second": 143.294, + "eval_qasc_pairs_steps_per_second": 2.239, + "step": 180 + }, + { + "epoch": 0.13964313421256788, + "eval_openbookqa_pairs_loss": 5.624441146850586, + "eval_openbookqa_pairs_runtime": 0.852, + "eval_openbookqa_pairs_samples_per_second": 150.235, + "eval_openbookqa_pairs_steps_per_second": 2.347, + "step": 180 + }, + { + "epoch": 0.13964313421256788, + "eval_msmarco_pairs_loss": 5.057087421417236, + "eval_msmarco_pairs_runtime": 1.965, + "eval_msmarco_pairs_samples_per_second": 65.14, + "eval_msmarco_pairs_steps_per_second": 1.018, + "step": 180 + }, + { + "epoch": 0.13964313421256788, + "eval_nq_pairs_loss": 5.1600823402404785, + "eval_nq_pairs_runtime": 3.3352, + "eval_nq_pairs_samples_per_second": 38.378, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 180 + }, + { + "epoch": 0.13964313421256788, + "eval_trivia_pairs_loss": 5.498032093048096, + "eval_trivia_pairs_runtime": 4.4801, + "eval_trivia_pairs_samples_per_second": 28.57, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 180 + }, + { + "epoch": 0.13964313421256788, + "eval_gooaq_pairs_loss": 4.9159464836120605, + "eval_gooaq_pairs_runtime": 1.4649, + "eval_gooaq_pairs_samples_per_second": 87.38, + "eval_gooaq_pairs_steps_per_second": 1.365, + "step": 180 + }, + { + "epoch": 0.13964313421256788, + "eval_paws-pos_loss": 0.44706642627716064, + "eval_paws-pos_runtime": 1.0313, + "eval_paws-pos_samples_per_second": 124.114, + "eval_paws-pos_steps_per_second": 1.939, + "step": 180 + }, + { + "epoch": 0.13964313421256788, + "eval_global_dataset_loss": 3.263042688369751, + "eval_global_dataset_runtime": 19.4013, + "eval_global_dataset_samples_per_second": 21.442, + "eval_global_dataset_steps_per_second": 0.361, + "step": 180 + }, + { + "epoch": 0.1404189294026377, + "grad_norm": 17.023818969726562, + "learning_rate": 7.361927144535839e-06, + "loss": 4.17, + "step": 181 + }, + { + "epoch": 0.14119472459270752, + "grad_norm": 22.472793579101562, + "learning_rate": 7.403055229142185e-06, + "loss": 4.8977, + "step": 182 + }, + { + "epoch": 0.14197051978277736, + "grad_norm": 14.055580139160156, + "learning_rate": 7.444183313748531e-06, + "loss": 3.5031, + "step": 183 + }, + { + "epoch": 0.14274631497284718, + "grad_norm": 14.841517448425293, + "learning_rate": 7.485311398354876e-06, + "loss": 3.4286, + "step": 184 + }, + { + "epoch": 0.143522110162917, + "grad_norm": 21.218738555908203, + "learning_rate": 7.526439482961222e-06, + "loss": 3.2536, + "step": 185 + }, + { + "epoch": 0.1442979053529868, + "grad_norm": 14.635804176330566, + "learning_rate": 7.567567567567567e-06, + "loss": 4.1632, + "step": 186 + }, + { + "epoch": 0.14507370054305663, + "grad_norm": 16.717111587524414, + "learning_rate": 7.608695652173912e-06, + "loss": 3.6512, + "step": 187 + }, + { + "epoch": 0.14584949573312644, + "grad_norm": 20.993993759155273, + "learning_rate": 7.649823736780258e-06, + "loss": 3.3453, + "step": 188 + }, + { + "epoch": 0.1466252909231963, + "grad_norm": 21.484180450439453, + "learning_rate": 7.690951821386603e-06, + "loss": 4.9785, + "step": 189 + }, + { + "epoch": 0.1474010861132661, + "grad_norm": 14.600364685058594, + "learning_rate": 7.732079905992948e-06, + "loss": 3.1781, + "step": 190 + }, + { + "epoch": 0.14817688130333592, + "grad_norm": 17.323484420776367, + "learning_rate": 7.773207990599295e-06, + "loss": 3.6681, + "step": 191 + }, + { + "epoch": 0.14895267649340574, + "grad_norm": 17.703405380249023, + "learning_rate": 7.814336075205639e-06, + "loss": 4.5109, + "step": 192 + }, + { + "epoch": 0.14972847168347556, + "grad_norm": 15.650117874145508, + "learning_rate": 7.855464159811986e-06, + "loss": 4.4412, + "step": 193 + }, + { + "epoch": 0.15050426687354537, + "grad_norm": 30.876625061035156, + "learning_rate": 7.896592244418331e-06, + "loss": 5.1287, + "step": 194 + }, + { + "epoch": 0.15128006206361522, + "grad_norm": 18.507810592651367, + "learning_rate": 7.937720329024676e-06, + "loss": 4.2127, + "step": 195 + }, + { + "epoch": 0.15205585725368503, + "grad_norm": 14.450047492980957, + "learning_rate": 7.978848413631022e-06, + "loss": 3.1766, + "step": 196 + }, + { + "epoch": 0.15283165244375485, + "grad_norm": 20.77067756652832, + "learning_rate": 8.019976498237367e-06, + "loss": 2.5901, + "step": 197 + }, + { + "epoch": 0.15360744763382467, + "grad_norm": 12.4572114944458, + "learning_rate": 8.061104582843712e-06, + "loss": 3.5886, + "step": 198 + }, + { + "epoch": 0.15438324282389448, + "grad_norm": 15.76103687286377, + "learning_rate": 8.102232667450058e-06, + "loss": 2.9085, + "step": 199 + }, + { + "epoch": 0.1551590380139643, + "grad_norm": 14.595601081848145, + "learning_rate": 8.143360752056405e-06, + "loss": 2.4958, + "step": 200 + }, + { + "epoch": 0.1551590380139643, + "eval_Qnli-dev_cosine_accuracy": 0.615234375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8780630826950073, + "eval_Qnli-dev_cosine_ap": 0.6112556879875849, + "eval_Qnli-dev_cosine_f1": 0.6363636363636364, + "eval_Qnli-dev_cosine_f1_threshold": 0.7175576090812683, + "eval_Qnli-dev_cosine_precision": 0.4714285714285714, + "eval_Qnli-dev_cosine_recall": 0.9788135593220338, + "eval_Qnli-dev_dot_accuracy": 0.580078125, + "eval_Qnli-dev_dot_accuracy_threshold": 530.33447265625, + "eval_Qnli-dev_dot_ap": 0.5372984783278565, + "eval_Qnli-dev_dot_f1": 0.6305555555555555, + "eval_Qnli-dev_dot_f1_threshold": 403.0582275390625, + "eval_Qnli-dev_dot_precision": 0.4690082644628099, + "eval_Qnli-dev_dot_recall": 0.961864406779661, + "eval_Qnli-dev_euclidean_accuracy": 0.619140625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.641845703125, + "eval_Qnli-dev_euclidean_ap": 0.6133386902087727, + "eval_Qnli-dev_euclidean_f1": 0.6331521739130436, + "eval_Qnli-dev_euclidean_f1_threshold": 19.538082122802734, + "eval_Qnli-dev_euclidean_precision": 0.466, + "eval_Qnli-dev_euclidean_recall": 0.9872881355932204, + "eval_Qnli-dev_manhattan_accuracy": 0.619140625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 190.95291137695312, + "eval_Qnli-dev_manhattan_ap": 0.6200633454752746, + "eval_Qnli-dev_manhattan_f1": 0.6364883401920438, + "eval_Qnli-dev_manhattan_f1_threshold": 324.4531555175781, + "eval_Qnli-dev_manhattan_precision": 0.47058823529411764, + "eval_Qnli-dev_manhattan_recall": 0.9830508474576272, + "eval_Qnli-dev_max_accuracy": 0.619140625, + "eval_Qnli-dev_max_accuracy_threshold": 530.33447265625, + "eval_Qnli-dev_max_ap": 0.6200633454752746, + "eval_Qnli-dev_max_f1": 0.6364883401920438, + "eval_Qnli-dev_max_f1_threshold": 403.0582275390625, + "eval_Qnli-dev_max_precision": 0.4714285714285714, + "eval_Qnli-dev_max_recall": 0.9872881355932204, + "eval_allNLI-dev_cosine_accuracy": 0.6875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.976324200630188, + "eval_allNLI-dev_cosine_ap": 0.45960393014401535, + "eval_allNLI-dev_cosine_f1": 0.5253164556962024, + "eval_allNLI-dev_cosine_f1_threshold": 0.8385776877403259, + "eval_allNLI-dev_cosine_precision": 0.3616557734204793, + "eval_allNLI-dev_cosine_recall": 0.9595375722543352, + "eval_allNLI-dev_dot_accuracy": 0.677734375, + "eval_allNLI-dev_dot_accuracy_threshold": 772.0188598632812, + "eval_allNLI-dev_dot_ap": 0.3804150461800332, + "eval_allNLI-dev_dot_f1": 0.5080763582966226, + "eval_allNLI-dev_dot_f1_threshold": 444.7850646972656, + "eval_allNLI-dev_dot_precision": 0.3405511811023622, + "eval_allNLI-dev_dot_recall": 1.0, + "eval_allNLI-dev_euclidean_accuracy": 0.685546875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 5.875195503234863, + "eval_allNLI-dev_euclidean_ap": 0.4551932697343415, + "eval_allNLI-dev_euclidean_f1": 0.5257903494176372, + "eval_allNLI-dev_euclidean_f1_threshold": 13.960739135742188, + "eval_allNLI-dev_euclidean_precision": 0.3691588785046729, + "eval_allNLI-dev_euclidean_recall": 0.9132947976878613, + "eval_allNLI-dev_manhattan_accuracy": 0.681640625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 96.69795227050781, + "eval_allNLI-dev_manhattan_ap": 0.4537023325460285, + "eval_allNLI-dev_manhattan_f1": 0.5204081632653061, + "eval_allNLI-dev_manhattan_f1_threshold": 222.21884155273438, + "eval_allNLI-dev_manhattan_precision": 0.3686746987951807, + "eval_allNLI-dev_manhattan_recall": 0.884393063583815, + "eval_allNLI-dev_max_accuracy": 0.6875, + "eval_allNLI-dev_max_accuracy_threshold": 772.0188598632812, + "eval_allNLI-dev_max_ap": 0.45960393014401535, + "eval_allNLI-dev_max_f1": 0.5257903494176372, + "eval_allNLI-dev_max_f1_threshold": 444.7850646972656, + "eval_allNLI-dev_max_precision": 0.3691588785046729, + "eval_allNLI-dev_max_recall": 1.0, + "eval_sequential_score": 0.6200633454752746, + "eval_sts-test_pearson_cosine": 0.24924277969472153, + "eval_sts-test_pearson_dot": 0.20632094538884468, + "eval_sts-test_pearson_euclidean": 0.2775041116953867, + "eval_sts-test_pearson_manhattan": 0.29073367102351505, + "eval_sts-test_pearson_max": 0.29073367102351505, + "eval_sts-test_spearman_cosine": 0.3391141842675074, + "eval_sts-test_spearman_dot": 0.21011509320588695, + "eval_sts-test_spearman_euclidean": 0.31565870563443316, + "eval_sts-test_spearman_manhattan": 0.3219084430440491, + "eval_sts-test_spearman_max": 0.3391141842675074, + "eval_vitaminc-pairs_loss": 3.716855764389038, + "eval_vitaminc-pairs_runtime": 5.4756, + "eval_vitaminc-pairs_samples_per_second": 23.377, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 200 + }, + { + "epoch": 0.1551590380139643, + "eval_negation-triplets_loss": 4.766229152679443, + "eval_negation-triplets_runtime": 1.0705, + "eval_negation-triplets_samples_per_second": 119.57, + "eval_negation-triplets_steps_per_second": 1.868, + "step": 200 + }, + { + "epoch": 0.1551590380139643, + "eval_scitail-pairs-pos_loss": 0.9834614992141724, + "eval_scitail-pairs-pos_runtime": 1.2447, + "eval_scitail-pairs-pos_samples_per_second": 102.839, + "eval_scitail-pairs-pos_steps_per_second": 1.607, + "step": 200 + }, + { + "epoch": 0.1551590380139643, + "eval_scitail-pairs-qa_loss": 1.3808467388153076, + "eval_scitail-pairs-qa_runtime": 0.7767, + "eval_scitail-pairs-qa_samples_per_second": 164.794, + "eval_scitail-pairs-qa_steps_per_second": 2.575, + "step": 200 + }, + { + "epoch": 0.1551590380139643, + "eval_xsum-pairs_loss": 3.506582260131836, + "eval_xsum-pairs_runtime": 3.9348, + "eval_xsum-pairs_samples_per_second": 32.53, + "eval_xsum-pairs_steps_per_second": 0.508, + "step": 200 + }, + { + "epoch": 0.1551590380139643, + "eval_sciq_pairs_loss": 0.38454076647758484, + "eval_sciq_pairs_runtime": 5.8783, + "eval_sciq_pairs_samples_per_second": 21.775, + "eval_sciq_pairs_steps_per_second": 0.34, + "step": 200 + }, + { + "epoch": 0.1551590380139643, + "eval_qasc_pairs_loss": 3.758909225463867, + "eval_qasc_pairs_runtime": 0.9063, + "eval_qasc_pairs_samples_per_second": 141.232, + "eval_qasc_pairs_steps_per_second": 2.207, + "step": 200 + }, + { + "epoch": 0.1551590380139643, + "eval_openbookqa_pairs_loss": 4.948967456817627, + "eval_openbookqa_pairs_runtime": 0.8606, + "eval_openbookqa_pairs_samples_per_second": 148.737, + "eval_openbookqa_pairs_steps_per_second": 2.324, + "step": 200 + }, + { + "epoch": 0.1551590380139643, + "eval_msmarco_pairs_loss": 4.785572528839111, + "eval_msmarco_pairs_runtime": 1.9606, + "eval_msmarco_pairs_samples_per_second": 65.286, + "eval_msmarco_pairs_steps_per_second": 1.02, + "step": 200 + }, + { + "epoch": 0.1551590380139643, + "eval_nq_pairs_loss": 4.675125598907471, + "eval_nq_pairs_runtime": 3.3155, + "eval_nq_pairs_samples_per_second": 38.607, + "eval_nq_pairs_steps_per_second": 0.603, + "step": 200 + }, + { + "epoch": 0.1551590380139643, + "eval_trivia_pairs_loss": 4.418967247009277, + "eval_trivia_pairs_runtime": 4.478, + "eval_trivia_pairs_samples_per_second": 28.584, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 200 + }, + { + "epoch": 0.1551590380139643, + "eval_gooaq_pairs_loss": 4.486730575561523, + "eval_gooaq_pairs_runtime": 1.4616, + "eval_gooaq_pairs_samples_per_second": 87.578, + "eval_gooaq_pairs_steps_per_second": 1.368, + "step": 200 + }, + { + "epoch": 0.1551590380139643, + "eval_paws-pos_loss": 0.4021705687046051, + "eval_paws-pos_runtime": 1.0276, + "eval_paws-pos_samples_per_second": 124.562, + "eval_paws-pos_steps_per_second": 1.946, + "step": 200 + }, + { + "epoch": 0.1551590380139643, + "eval_global_dataset_loss": 2.6410105228424072, + "eval_global_dataset_runtime": 19.3933, + "eval_global_dataset_samples_per_second": 21.451, + "eval_global_dataset_steps_per_second": 0.361, + "step": 200 + }, + { + "epoch": 0.15593483320403415, + "grad_norm": 36.21416091918945, + "learning_rate": 8.184488836662748e-06, + "loss": 5.5704, + "step": 201 + }, + { + "epoch": 0.15671062839410396, + "grad_norm": 15.138063430786133, + "learning_rate": 8.225616921269095e-06, + "loss": 2.8612, + "step": 202 + }, + { + "epoch": 0.15748642358417378, + "grad_norm": 27.27367401123047, + "learning_rate": 8.26674500587544e-06, + "loss": 4.8846, + "step": 203 + }, + { + "epoch": 0.1582622187742436, + "grad_norm": 19.24480438232422, + "learning_rate": 8.307873090481786e-06, + "loss": 3.1182, + "step": 204 + }, + { + "epoch": 0.1590380139643134, + "grad_norm": 18.584644317626953, + "learning_rate": 8.349001175088131e-06, + "loss": 1.9789, + "step": 205 + }, + { + "epoch": 0.15981380915438323, + "grad_norm": 15.61733341217041, + "learning_rate": 8.390129259694476e-06, + "loss": 2.9743, + "step": 206 + }, + { + "epoch": 0.16058960434445307, + "grad_norm": 16.221357345581055, + "learning_rate": 8.431257344300822e-06, + "loss": 3.6543, + "step": 207 + }, + { + "epoch": 0.1613653995345229, + "grad_norm": 13.963678359985352, + "learning_rate": 8.472385428907167e-06, + "loss": 3.117, + "step": 208 + }, + { + "epoch": 0.1621411947245927, + "grad_norm": 18.764089584350586, + "learning_rate": 8.513513513513514e-06, + "loss": 1.5977, + "step": 209 + }, + { + "epoch": 0.16291698991466252, + "grad_norm": 29.377960205078125, + "learning_rate": 8.554641598119857e-06, + "loss": 5.1443, + "step": 210 + }, + { + "epoch": 0.16369278510473234, + "grad_norm": 13.549287796020508, + "learning_rate": 8.595769682726204e-06, + "loss": 0.668, + "step": 211 + }, + { + "epoch": 0.16446858029480219, + "grad_norm": 14.000697135925293, + "learning_rate": 8.636897767332548e-06, + "loss": 0.7522, + "step": 212 + }, + { + "epoch": 0.165244375484872, + "grad_norm": 18.982425689697266, + "learning_rate": 8.678025851938895e-06, + "loss": 2.3447, + "step": 213 + }, + { + "epoch": 0.16602017067494182, + "grad_norm": 13.098822593688965, + "learning_rate": 8.71915393654524e-06, + "loss": 0.5818, + "step": 214 + }, + { + "epoch": 0.16679596586501164, + "grad_norm": 17.622783660888672, + "learning_rate": 8.760282021151586e-06, + "loss": 3.556, + "step": 215 + }, + { + "epoch": 0.16757176105508145, + "grad_norm": 24.308530807495117, + "learning_rate": 8.801410105757933e-06, + "loss": 2.8425, + "step": 216 + }, + { + "epoch": 0.16834755624515127, + "grad_norm": 41.18635559082031, + "learning_rate": 8.842538190364276e-06, + "loss": 5.4947, + "step": 217 + }, + { + "epoch": 0.16912335143522111, + "grad_norm": 38.11179733276367, + "learning_rate": 8.883666274970621e-06, + "loss": 4.2956, + "step": 218 + }, + { + "epoch": 0.16989914662529093, + "grad_norm": 20.423124313354492, + "learning_rate": 8.924794359576967e-06, + "loss": 2.1325, + "step": 219 + }, + { + "epoch": 0.17067494181536075, + "grad_norm": 19.0192813873291, + "learning_rate": 8.965922444183314e-06, + "loss": 1.4286, + "step": 220 + }, + { + "epoch": 0.17067494181536075, + "eval_Qnli-dev_cosine_accuracy": 0.625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8608779907226562, + "eval_Qnli-dev_cosine_ap": 0.6155996680907755, + "eval_Qnli-dev_cosine_f1": 0.6555183946488294, + "eval_Qnli-dev_cosine_f1_threshold": 0.7760155200958252, + "eval_Qnli-dev_cosine_precision": 0.5414364640883977, + "eval_Qnli-dev_cosine_recall": 0.8305084745762712, + "eval_Qnli-dev_dot_accuracy": 0.62109375, + "eval_Qnli-dev_dot_accuracy_threshold": 477.1441650390625, + "eval_Qnli-dev_dot_ap": 0.5823923648803983, + "eval_Qnli-dev_dot_f1": 0.6345029239766081, + "eval_Qnli-dev_dot_f1_threshold": 358.5615234375, + "eval_Qnli-dev_dot_precision": 0.484375, + "eval_Qnli-dev_dot_recall": 0.9194915254237288, + "eval_Qnli-dev_euclidean_accuracy": 0.625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.561055183410645, + "eval_Qnli-dev_euclidean_ap": 0.6128136787518578, + "eval_Qnli-dev_euclidean_f1": 0.6476510067114094, + "eval_Qnli-dev_euclidean_f1_threshold": 15.50450325012207, + "eval_Qnli-dev_euclidean_precision": 0.5361111111111111, + "eval_Qnli-dev_euclidean_recall": 0.8177966101694916, + "eval_Qnli-dev_manhattan_accuracy": 0.619140625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 215.1610565185547, + "eval_Qnli-dev_manhattan_ap": 0.621124218438047, + "eval_Qnli-dev_manhattan_f1": 0.645484949832776, + "eval_Qnli-dev_manhattan_f1_threshold": 273.8544921875, + "eval_Qnli-dev_manhattan_precision": 0.5331491712707183, + "eval_Qnli-dev_manhattan_recall": 0.8177966101694916, + "eval_Qnli-dev_max_accuracy": 0.625, + "eval_Qnli-dev_max_accuracy_threshold": 477.1441650390625, + "eval_Qnli-dev_max_ap": 0.621124218438047, + "eval_Qnli-dev_max_f1": 0.6555183946488294, + "eval_Qnli-dev_max_f1_threshold": 358.5615234375, + "eval_Qnli-dev_max_precision": 0.5414364640883977, + "eval_Qnli-dev_max_recall": 0.9194915254237288, + "eval_allNLI-dev_cosine_accuracy": 0.685546875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9713066816329956, + "eval_allNLI-dev_cosine_ap": 0.48223591780287633, + "eval_allNLI-dev_cosine_f1": 0.5408805031446541, + "eval_allNLI-dev_cosine_f1_threshold": 0.8952139616012573, + "eval_allNLI-dev_cosine_precision": 0.4243421052631579, + "eval_allNLI-dev_cosine_recall": 0.7456647398843931, + "eval_allNLI-dev_dot_accuracy": 0.677734375, + "eval_allNLI-dev_dot_accuracy_threshold": 741.37939453125, + "eval_allNLI-dev_dot_ap": 0.3936517100476376, + "eval_allNLI-dev_dot_f1": 0.5074626865671642, + "eval_allNLI-dev_dot_f1_threshold": 416.87164306640625, + "eval_allNLI-dev_dot_precision": 0.3420523138832998, + "eval_allNLI-dev_dot_recall": 0.9826589595375722, + "eval_allNLI-dev_euclidean_accuracy": 0.685546875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 6.748725891113281, + "eval_allNLI-dev_euclidean_ap": 0.47719987976648454, + "eval_allNLI-dev_euclidean_f1": 0.5415860735009671, + "eval_allNLI-dev_euclidean_f1_threshold": 12.946854591369629, + "eval_allNLI-dev_euclidean_precision": 0.4069767441860465, + "eval_allNLI-dev_euclidean_recall": 0.8092485549132948, + "eval_allNLI-dev_manhattan_accuracy": 0.685546875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 110.15467834472656, + "eval_allNLI-dev_manhattan_ap": 0.4846184039440549, + "eval_allNLI-dev_manhattan_f1": 0.5335753176043557, + "eval_allNLI-dev_manhattan_f1_threshold": 234.3075408935547, + "eval_allNLI-dev_manhattan_precision": 0.3888888888888889, + "eval_allNLI-dev_manhattan_recall": 0.8497109826589595, + "eval_allNLI-dev_max_accuracy": 0.685546875, + "eval_allNLI-dev_max_accuracy_threshold": 741.37939453125, + "eval_allNLI-dev_max_ap": 0.4846184039440549, + "eval_allNLI-dev_max_f1": 0.5415860735009671, + "eval_allNLI-dev_max_f1_threshold": 416.87164306640625, + "eval_allNLI-dev_max_precision": 0.4243421052631579, + "eval_allNLI-dev_max_recall": 0.9826589595375722, + "eval_sequential_score": 0.621124218438047, + "eval_sts-test_pearson_cosine": 0.31919410971368356, + "eval_sts-test_pearson_dot": 0.21556048485576235, + "eval_sts-test_pearson_euclidean": 0.35723202726898373, + "eval_sts-test_pearson_manhattan": 0.36765421659339287, + "eval_sts-test_pearson_max": 0.36765421659339287, + "eval_sts-test_spearman_cosine": 0.41124188237768894, + "eval_sts-test_spearman_dot": 0.2248200891915325, + "eval_sts-test_spearman_euclidean": 0.389006828577642, + "eval_sts-test_spearman_manhattan": 0.39845426736998035, + "eval_sts-test_spearman_max": 0.41124188237768894, + "eval_vitaminc-pairs_loss": 4.022590160369873, + "eval_vitaminc-pairs_runtime": 5.4704, + "eval_vitaminc-pairs_samples_per_second": 23.399, + "eval_vitaminc-pairs_steps_per_second": 0.366, + "step": 220 + }, + { + "epoch": 0.17067494181536075, + "eval_negation-triplets_loss": 4.421084403991699, + "eval_negation-triplets_runtime": 1.0619, + "eval_negation-triplets_samples_per_second": 120.534, + "eval_negation-triplets_steps_per_second": 1.883, + "step": 220 + }, + { + "epoch": 0.17067494181536075, + "eval_scitail-pairs-pos_loss": 0.7828177809715271, + "eval_scitail-pairs-pos_runtime": 1.2253, + "eval_scitail-pairs-pos_samples_per_second": 104.468, + "eval_scitail-pairs-pos_steps_per_second": 1.632, + "step": 220 + }, + { + "epoch": 0.17067494181536075, + "eval_scitail-pairs-qa_loss": 0.9572672843933105, + "eval_scitail-pairs-qa_runtime": 0.7677, + "eval_scitail-pairs-qa_samples_per_second": 166.739, + "eval_scitail-pairs-qa_steps_per_second": 2.605, + "step": 220 + }, + { + "epoch": 0.17067494181536075, + "eval_xsum-pairs_loss": 3.0789806842803955, + "eval_xsum-pairs_runtime": 3.9039, + "eval_xsum-pairs_samples_per_second": 32.788, + "eval_xsum-pairs_steps_per_second": 0.512, + "step": 220 + }, + { + "epoch": 0.17067494181536075, + "eval_sciq_pairs_loss": 0.39149752259254456, + "eval_sciq_pairs_runtime": 5.8117, + "eval_sciq_pairs_samples_per_second": 22.024, + "eval_sciq_pairs_steps_per_second": 0.344, + "step": 220 + }, + { + "epoch": 0.17067494181536075, + "eval_qasc_pairs_loss": 3.374030828475952, + "eval_qasc_pairs_runtime": 0.9033, + "eval_qasc_pairs_samples_per_second": 141.696, + "eval_qasc_pairs_steps_per_second": 2.214, + "step": 220 + }, + { + "epoch": 0.17067494181536075, + "eval_openbookqa_pairs_loss": 5.161572456359863, + "eval_openbookqa_pairs_runtime": 0.8619, + "eval_openbookqa_pairs_samples_per_second": 148.506, + "eval_openbookqa_pairs_steps_per_second": 2.32, + "step": 220 + }, + { + "epoch": 0.17067494181536075, + "eval_msmarco_pairs_loss": 4.489230632781982, + "eval_msmarco_pairs_runtime": 1.9668, + "eval_msmarco_pairs_samples_per_second": 65.079, + "eval_msmarco_pairs_steps_per_second": 1.017, + "step": 220 + }, + { + "epoch": 0.17067494181536075, + "eval_nq_pairs_loss": 4.549116611480713, + "eval_nq_pairs_runtime": 3.3346, + "eval_nq_pairs_samples_per_second": 38.385, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 220 + }, + { + "epoch": 0.17067494181536075, + "eval_trivia_pairs_loss": 5.130815029144287, + "eval_trivia_pairs_runtime": 4.4818, + "eval_trivia_pairs_samples_per_second": 28.56, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 220 + }, + { + "epoch": 0.17067494181536075, + "eval_gooaq_pairs_loss": 3.8106689453125, + "eval_gooaq_pairs_runtime": 1.4609, + "eval_gooaq_pairs_samples_per_second": 87.615, + "eval_gooaq_pairs_steps_per_second": 1.369, + "step": 220 + }, + { + "epoch": 0.17067494181536075, + "eval_paws-pos_loss": 0.1346072554588318, + "eval_paws-pos_runtime": 1.0308, + "eval_paws-pos_samples_per_second": 124.178, + "eval_paws-pos_steps_per_second": 1.94, + "step": 220 + }, + { + "epoch": 0.17067494181536075, + "eval_global_dataset_loss": 2.3529915809631348, + "eval_global_dataset_runtime": 19.4389, + "eval_global_dataset_samples_per_second": 21.4, + "eval_global_dataset_steps_per_second": 0.36, + "step": 220 + }, + { + "epoch": 0.17145073700543056, + "grad_norm": 24.33696174621582, + "learning_rate": 9.007050528789659e-06, + "loss": 2.2834, + "step": 221 + }, + { + "epoch": 0.17222653219550038, + "grad_norm": 19.657947540283203, + "learning_rate": 9.048178613396003e-06, + "loss": 3.4539, + "step": 222 + }, + { + "epoch": 0.1730023273855702, + "grad_norm": 22.255239486694336, + "learning_rate": 9.08930669800235e-06, + "loss": 2.7994, + "step": 223 + }, + { + "epoch": 0.17377812257564004, + "grad_norm": 20.619884490966797, + "learning_rate": 9.130434782608695e-06, + "loss": 2.463, + "step": 224 + }, + { + "epoch": 0.17455391776570986, + "grad_norm": 22.244462966918945, + "learning_rate": 9.17156286721504e-06, + "loss": 4.1749, + "step": 225 + }, + { + "epoch": 0.17532971295577968, + "grad_norm": 25.1329402923584, + "learning_rate": 9.212690951821385e-06, + "loss": 3.7001, + "step": 226 + }, + { + "epoch": 0.1761055081458495, + "grad_norm": 35.8182258605957, + "learning_rate": 9.253819036427732e-06, + "loss": 5.6003, + "step": 227 + }, + { + "epoch": 0.1768813033359193, + "grad_norm": 33.197845458984375, + "learning_rate": 9.294947121034076e-06, + "loss": 5.5236, + "step": 228 + }, + { + "epoch": 0.17765709852598913, + "grad_norm": 14.53373908996582, + "learning_rate": 9.336075205640421e-06, + "loss": 2.8411, + "step": 229 + }, + { + "epoch": 0.17843289371605897, + "grad_norm": 8.925052642822266, + "learning_rate": 9.377203290246768e-06, + "loss": 0.3611, + "step": 230 + }, + { + "epoch": 0.1792086889061288, + "grad_norm": 19.080123901367188, + "learning_rate": 9.418331374853114e-06, + "loss": 3.4145, + "step": 231 + }, + { + "epoch": 0.1799844840961986, + "grad_norm": 16.67955780029297, + "learning_rate": 9.459459459459459e-06, + "loss": 2.7527, + "step": 232 + }, + { + "epoch": 0.18076027928626842, + "grad_norm": 13.526026725769043, + "learning_rate": 9.500587544065804e-06, + "loss": 1.977, + "step": 233 + }, + { + "epoch": 0.18153607447633824, + "grad_norm": 12.823989868164062, + "learning_rate": 9.54171562867215e-06, + "loss": 1.9694, + "step": 234 + }, + { + "epoch": 0.18231186966640806, + "grad_norm": 14.984912872314453, + "learning_rate": 9.582843713278495e-06, + "loss": 2.4804, + "step": 235 + }, + { + "epoch": 0.1830876648564779, + "grad_norm": 13.724047660827637, + "learning_rate": 9.62397179788484e-06, + "loss": 2.7869, + "step": 236 + }, + { + "epoch": 0.18386346004654772, + "grad_norm": 14.626879692077637, + "learning_rate": 9.665099882491187e-06, + "loss": 2.0562, + "step": 237 + }, + { + "epoch": 0.18463925523661753, + "grad_norm": 8.06078815460205, + "learning_rate": 9.706227967097532e-06, + "loss": 0.3609, + "step": 238 + }, + { + "epoch": 0.18541505042668735, + "grad_norm": 13.897974014282227, + "learning_rate": 9.747356051703878e-06, + "loss": 1.9205, + "step": 239 + }, + { + "epoch": 0.18619084561675717, + "grad_norm": 6.743955612182617, + "learning_rate": 9.788484136310221e-06, + "loss": 0.2964, + "step": 240 + }, + { + "epoch": 0.18619084561675717, + "eval_Qnli-dev_cosine_accuracy": 0.62890625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8693021535873413, + "eval_Qnli-dev_cosine_ap": 0.611781004288293, + "eval_Qnli-dev_cosine_f1": 0.6366559485530547, + "eval_Qnli-dev_cosine_f1_threshold": 0.7878068685531616, + "eval_Qnli-dev_cosine_precision": 0.5129533678756477, + "eval_Qnli-dev_cosine_recall": 0.8389830508474576, + "eval_Qnli-dev_dot_accuracy": 0.595703125, + "eval_Qnli-dev_dot_accuracy_threshold": 521.8904418945312, + "eval_Qnli-dev_dot_ap": 0.5496339935508336, + "eval_Qnli-dev_dot_f1": 0.6397694524495677, + "eval_Qnli-dev_dot_f1_threshold": 375.82940673828125, + "eval_Qnli-dev_dot_precision": 0.4847161572052402, + "eval_Qnli-dev_dot_recall": 0.940677966101695, + "eval_Qnli-dev_euclidean_accuracy": 0.626953125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.786832809448242, + "eval_Qnli-dev_euclidean_ap": 0.616692733684346, + "eval_Qnli-dev_euclidean_f1": 0.6368159203980099, + "eval_Qnli-dev_euclidean_f1_threshold": 15.11730670928955, + "eval_Qnli-dev_euclidean_precision": 0.5231607629427792, + "eval_Qnli-dev_euclidean_recall": 0.8135593220338984, + "eval_Qnli-dev_manhattan_accuracy": 0.6171875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 198.48147583007812, + "eval_Qnli-dev_manhattan_ap": 0.6210006336689851, + "eval_Qnli-dev_manhattan_f1": 0.6374367622259697, + "eval_Qnli-dev_manhattan_f1_threshold": 262.7891845703125, + "eval_Qnli-dev_manhattan_precision": 0.5294117647058824, + "eval_Qnli-dev_manhattan_recall": 0.8008474576271186, + "eval_Qnli-dev_max_accuracy": 0.62890625, + "eval_Qnli-dev_max_accuracy_threshold": 521.8904418945312, + "eval_Qnli-dev_max_ap": 0.6210006336689851, + "eval_Qnli-dev_max_f1": 0.6397694524495677, + "eval_Qnli-dev_max_f1_threshold": 375.82940673828125, + "eval_Qnli-dev_max_precision": 0.5294117647058824, + "eval_Qnli-dev_max_recall": 0.940677966101695, + "eval_allNLI-dev_cosine_accuracy": 0.6875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9714217782020569, + "eval_allNLI-dev_cosine_ap": 0.5054463009748529, + "eval_allNLI-dev_cosine_f1": 0.5534591194968553, + "eval_allNLI-dev_cosine_f1_threshold": 0.891494631767273, + "eval_allNLI-dev_cosine_precision": 0.4342105263157895, + "eval_allNLI-dev_cosine_recall": 0.7630057803468208, + "eval_allNLI-dev_dot_accuracy": 0.6796875, + "eval_allNLI-dev_dot_accuracy_threshold": 697.1558837890625, + "eval_allNLI-dev_dot_ap": 0.40155312352801564, + "eval_allNLI-dev_dot_f1": 0.5088235294117648, + "eval_allNLI-dev_dot_f1_threshold": 376.5243225097656, + "eval_allNLI-dev_dot_precision": 0.34122287968441817, + "eval_allNLI-dev_dot_recall": 1.0, + "eval_allNLI-dev_euclidean_accuracy": 0.685546875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 6.3458075523376465, + "eval_allNLI-dev_euclidean_ap": 0.5028194115939991, + "eval_allNLI-dev_euclidean_f1": 0.5518590998043053, + "eval_allNLI-dev_euclidean_f1_threshold": 12.797416687011719, + "eval_allNLI-dev_euclidean_precision": 0.4171597633136095, + "eval_allNLI-dev_euclidean_recall": 0.815028901734104, + "eval_allNLI-dev_manhattan_accuracy": 0.689453125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 145.72430419921875, + "eval_allNLI-dev_manhattan_ap": 0.5074540422577104, + "eval_allNLI-dev_manhattan_f1": 0.5382932166301969, + "eval_allNLI-dev_manhattan_f1_threshold": 210.015625, + "eval_allNLI-dev_manhattan_precision": 0.43309859154929575, + "eval_allNLI-dev_manhattan_recall": 0.7109826589595376, + "eval_allNLI-dev_max_accuracy": 0.689453125, + "eval_allNLI-dev_max_accuracy_threshold": 697.1558837890625, + "eval_allNLI-dev_max_ap": 0.5074540422577104, + "eval_allNLI-dev_max_f1": 0.5534591194968553, + "eval_allNLI-dev_max_f1_threshold": 376.5243225097656, + "eval_allNLI-dev_max_precision": 0.4342105263157895, + "eval_allNLI-dev_max_recall": 1.0, + "eval_sequential_score": 0.6210006336689851, + "eval_sts-test_pearson_cosine": 0.437261385872042, + "eval_sts-test_pearson_dot": 0.27944802216402886, + "eval_sts-test_pearson_euclidean": 0.46065093061659046, + "eval_sts-test_pearson_manhattan": 0.4698077558920942, + "eval_sts-test_pearson_max": 0.4698077558920942, + "eval_sts-test_spearman_cosine": 0.5068103175719304, + "eval_sts-test_spearman_dot": 0.28160415924837434, + "eval_sts-test_spearman_euclidean": 0.4813224156515044, + "eval_sts-test_spearman_manhattan": 0.4888016492165877, + "eval_sts-test_spearman_max": 0.5068103175719304, + "eval_vitaminc-pairs_loss": 3.8961689472198486, + "eval_vitaminc-pairs_runtime": 5.4898, + "eval_vitaminc-pairs_samples_per_second": 23.316, + "eval_vitaminc-pairs_steps_per_second": 0.364, + "step": 240 + }, + { + "epoch": 0.18619084561675717, + "eval_negation-triplets_loss": 3.9956982135772705, + "eval_negation-triplets_runtime": 1.0718, + "eval_negation-triplets_samples_per_second": 119.43, + "eval_negation-triplets_steps_per_second": 1.866, + "step": 240 + }, + { + "epoch": 0.18619084561675717, + "eval_scitail-pairs-pos_loss": 0.6250349283218384, + "eval_scitail-pairs-pos_runtime": 1.2283, + "eval_scitail-pairs-pos_samples_per_second": 104.211, + "eval_scitail-pairs-pos_steps_per_second": 1.628, + "step": 240 + }, + { + "epoch": 0.18619084561675717, + "eval_scitail-pairs-qa_loss": 0.8983888030052185, + "eval_scitail-pairs-qa_runtime": 0.7728, + "eval_scitail-pairs-qa_samples_per_second": 165.636, + "eval_scitail-pairs-qa_steps_per_second": 2.588, + "step": 240 + }, + { + "epoch": 0.18619084561675717, + "eval_xsum-pairs_loss": 2.681138038635254, + "eval_xsum-pairs_runtime": 3.921, + "eval_xsum-pairs_samples_per_second": 32.645, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 240 + }, + { + "epoch": 0.18619084561675717, + "eval_sciq_pairs_loss": 0.3189742863178253, + "eval_sciq_pairs_runtime": 5.8916, + "eval_sciq_pairs_samples_per_second": 21.726, + "eval_sciq_pairs_steps_per_second": 0.339, + "step": 240 + }, + { + "epoch": 0.18619084561675717, + "eval_qasc_pairs_loss": 2.76663875579834, + "eval_qasc_pairs_runtime": 0.9119, + "eval_qasc_pairs_samples_per_second": 140.374, + "eval_qasc_pairs_steps_per_second": 2.193, + "step": 240 + }, + { + "epoch": 0.18619084561675717, + "eval_openbookqa_pairs_loss": 4.003782749176025, + "eval_openbookqa_pairs_runtime": 0.8645, + "eval_openbookqa_pairs_samples_per_second": 148.058, + "eval_openbookqa_pairs_steps_per_second": 2.313, + "step": 240 + }, + { + "epoch": 0.18619084561675717, + "eval_msmarco_pairs_loss": 3.789357900619507, + "eval_msmarco_pairs_runtime": 1.9837, + "eval_msmarco_pairs_samples_per_second": 64.525, + "eval_msmarco_pairs_steps_per_second": 1.008, + "step": 240 + }, + { + "epoch": 0.18619084561675717, + "eval_nq_pairs_loss": 3.925625801086426, + "eval_nq_pairs_runtime": 3.3218, + "eval_nq_pairs_samples_per_second": 38.534, + "eval_nq_pairs_steps_per_second": 0.602, + "step": 240 + }, + { + "epoch": 0.18619084561675717, + "eval_trivia_pairs_loss": 3.9500904083251953, + "eval_trivia_pairs_runtime": 4.4861, + "eval_trivia_pairs_samples_per_second": 28.533, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 240 + }, + { + "epoch": 0.18619084561675717, + "eval_gooaq_pairs_loss": 3.2876720428466797, + "eval_gooaq_pairs_runtime": 1.4821, + "eval_gooaq_pairs_samples_per_second": 86.362, + "eval_gooaq_pairs_steps_per_second": 1.349, + "step": 240 + }, + { + "epoch": 0.18619084561675717, + "eval_paws-pos_loss": 0.12235681712627411, + "eval_paws-pos_runtime": 1.0363, + "eval_paws-pos_samples_per_second": 123.516, + "eval_paws-pos_steps_per_second": 1.93, + "step": 240 + }, + { + "epoch": 0.18619084561675717, + "eval_global_dataset_loss": 1.9579764604568481, + "eval_global_dataset_runtime": 19.4, + "eval_global_dataset_samples_per_second": 21.443, + "eval_global_dataset_steps_per_second": 0.361, + "step": 240 + }, + { + "epoch": 0.186966640806827, + "grad_norm": 17.613515853881836, + "learning_rate": 9.829612220916568e-06, + "loss": 3.2402, + "step": 241 + }, + { + "epoch": 0.18774243599689683, + "grad_norm": 19.704069137573242, + "learning_rate": 9.870740305522913e-06, + "loss": 3.1076, + "step": 242 + }, + { + "epoch": 0.18851823118696664, + "grad_norm": 14.683489799499512, + "learning_rate": 9.911868390129259e-06, + "loss": 1.8656, + "step": 243 + }, + { + "epoch": 0.18929402637703646, + "grad_norm": 15.879508018493652, + "learning_rate": 9.952996474735606e-06, + "loss": 2.7912, + "step": 244 + }, + { + "epoch": 0.19006982156710628, + "grad_norm": 5.585937023162842, + "learning_rate": 9.99412455934195e-06, + "loss": 0.2569, + "step": 245 + }, + { + "epoch": 0.1908456167571761, + "grad_norm": 18.365917205810547, + "learning_rate": 1.0035252643948295e-05, + "loss": 2.1095, + "step": 246 + }, + { + "epoch": 0.19162141194724594, + "grad_norm": 14.385225296020508, + "learning_rate": 1.007638072855464e-05, + "loss": 1.5261, + "step": 247 + }, + { + "epoch": 0.19239720713731576, + "grad_norm": 23.97905921936035, + "learning_rate": 1.0117508813160987e-05, + "loss": 3.4551, + "step": 248 + }, + { + "epoch": 0.19317300232738557, + "grad_norm": 15.02582836151123, + "learning_rate": 1.0158636897767332e-05, + "loss": 2.2465, + "step": 249 + }, + { + "epoch": 0.1939487975174554, + "grad_norm": 32.1656494140625, + "learning_rate": 1.0199764982373677e-05, + "loss": 4.7511, + "step": 250 + }, + { + "epoch": 0.1947245927075252, + "grad_norm": 22.766891479492188, + "learning_rate": 1.0240893066980021e-05, + "loss": 2.9648, + "step": 251 + }, + { + "epoch": 0.19550038789759502, + "grad_norm": 22.01688575744629, + "learning_rate": 1.0282021151586368e-05, + "loss": 3.4702, + "step": 252 + }, + { + "epoch": 0.19627618308766487, + "grad_norm": 19.137022018432617, + "learning_rate": 1.0323149236192713e-05, + "loss": 2.2448, + "step": 253 + }, + { + "epoch": 0.19705197827773469, + "grad_norm": 26.455888748168945, + "learning_rate": 1.0364277320799059e-05, + "loss": 4.5872, + "step": 254 + }, + { + "epoch": 0.1978277734678045, + "grad_norm": 15.777259826660156, + "learning_rate": 1.0405405405405406e-05, + "loss": 1.8122, + "step": 255 + }, + { + "epoch": 0.19860356865787432, + "grad_norm": 17.67413902282715, + "learning_rate": 1.0446533490011751e-05, + "loss": 2.8551, + "step": 256 + }, + { + "epoch": 0.19937936384794414, + "grad_norm": 17.045608520507812, + "learning_rate": 1.0487661574618094e-05, + "loss": 2.6086, + "step": 257 + }, + { + "epoch": 0.20015515903801395, + "grad_norm": 14.350496292114258, + "learning_rate": 1.052878965922444e-05, + "loss": 1.232, + "step": 258 + }, + { + "epoch": 0.2009309542280838, + "grad_norm": 18.945785522460938, + "learning_rate": 1.0569917743830787e-05, + "loss": 2.0938, + "step": 259 + }, + { + "epoch": 0.2017067494181536, + "grad_norm": 21.14448356628418, + "learning_rate": 1.0611045828437132e-05, + "loss": 3.2889, + "step": 260 + }, + { + "epoch": 0.2017067494181536, + "eval_Qnli-dev_cosine_accuracy": 0.62890625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8959039449691772, + "eval_Qnli-dev_cosine_ap": 0.6317320089748972, + "eval_Qnli-dev_cosine_f1": 0.6438569206842925, + "eval_Qnli-dev_cosine_f1_threshold": 0.7514594793319702, + "eval_Qnli-dev_cosine_precision": 0.5085995085995086, + "eval_Qnli-dev_cosine_recall": 0.8771186440677966, + "eval_Qnli-dev_dot_accuracy": 0.615234375, + "eval_Qnli-dev_dot_accuracy_threshold": 501.94647216796875, + "eval_Qnli-dev_dot_ap": 0.570773669147705, + "eval_Qnli-dev_dot_f1": 0.6416184971098267, + "eval_Qnli-dev_dot_f1_threshold": 352.76312255859375, + "eval_Qnli-dev_dot_precision": 0.4868421052631579, + "eval_Qnli-dev_dot_recall": 0.940677966101695, + "eval_Qnli-dev_euclidean_accuracy": 0.62890625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.027973175048828, + "eval_Qnli-dev_euclidean_ap": 0.6338956796033433, + "eval_Qnli-dev_euclidean_f1": 0.6456692913385828, + "eval_Qnli-dev_euclidean_f1_threshold": 16.208221435546875, + "eval_Qnli-dev_euclidean_precision": 0.5137844611528822, + "eval_Qnli-dev_euclidean_recall": 0.8686440677966102, + "eval_Qnli-dev_manhattan_accuracy": 0.6171875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 177.6362762451172, + "eval_Qnli-dev_manhattan_ap": 0.635486065397315, + "eval_Qnli-dev_manhattan_f1": 0.6525037936267072, + "eval_Qnli-dev_manhattan_f1_threshold": 298.76824951171875, + "eval_Qnli-dev_manhattan_precision": 0.508274231678487, + "eval_Qnli-dev_manhattan_recall": 0.9110169491525424, + "eval_Qnli-dev_max_accuracy": 0.62890625, + "eval_Qnli-dev_max_accuracy_threshold": 501.94647216796875, + "eval_Qnli-dev_max_ap": 0.635486065397315, + "eval_Qnli-dev_max_f1": 0.6525037936267072, + "eval_Qnli-dev_max_f1_threshold": 352.76312255859375, + "eval_Qnli-dev_max_precision": 0.5137844611528822, + "eval_Qnli-dev_max_recall": 0.940677966101695, + "eval_allNLI-dev_cosine_accuracy": 0.693359375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9633911848068237, + "eval_allNLI-dev_cosine_ap": 0.5232737415129294, + "eval_allNLI-dev_cosine_f1": 0.5643153526970954, + "eval_allNLI-dev_cosine_f1_threshold": 0.8676438927650452, + "eval_allNLI-dev_cosine_precision": 0.4401294498381877, + "eval_allNLI-dev_cosine_recall": 0.7861271676300579, + "eval_allNLI-dev_dot_accuracy": 0.67578125, + "eval_allNLI-dev_dot_accuracy_threshold": 628.2297973632812, + "eval_allNLI-dev_dot_ap": 0.41885658826681826, + "eval_allNLI-dev_dot_f1": 0.5112781954887218, + "eval_allNLI-dev_dot_f1_threshold": 387.4698181152344, + "eval_allNLI-dev_dot_precision": 0.34552845528455284, + "eval_allNLI-dev_dot_recall": 0.9826589595375722, + "eval_allNLI-dev_euclidean_accuracy": 0.69140625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 6.63711404800415, + "eval_allNLI-dev_euclidean_ap": 0.5165862059273876, + "eval_allNLI-dev_euclidean_f1": 0.5619469026548674, + "eval_allNLI-dev_euclidean_f1_threshold": 12.216785430908203, + "eval_allNLI-dev_euclidean_precision": 0.4551971326164875, + "eval_allNLI-dev_euclidean_recall": 0.7341040462427746, + "eval_allNLI-dev_manhattan_accuracy": 0.689453125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 113.34144592285156, + "eval_allNLI-dev_manhattan_ap": 0.514531606503361, + "eval_allNLI-dev_manhattan_f1": 0.5508982035928144, + "eval_allNLI-dev_manhattan_f1_threshold": 228.50057983398438, + "eval_allNLI-dev_manhattan_precision": 0.42073170731707316, + "eval_allNLI-dev_manhattan_recall": 0.7976878612716763, + "eval_allNLI-dev_max_accuracy": 0.693359375, + "eval_allNLI-dev_max_accuracy_threshold": 628.2297973632812, + "eval_allNLI-dev_max_ap": 0.5232737415129294, + "eval_allNLI-dev_max_f1": 0.5643153526970954, + "eval_allNLI-dev_max_f1_threshold": 387.4698181152344, + "eval_allNLI-dev_max_precision": 0.4551971326164875, + "eval_allNLI-dev_max_recall": 0.9826589595375722, + "eval_sequential_score": 0.635486065397315, + "eval_sts-test_pearson_cosine": 0.5473549878890289, + "eval_sts-test_pearson_dot": 0.4025186478072074, + "eval_sts-test_pearson_euclidean": 0.5655101153873763, + "eval_sts-test_pearson_manhattan": 0.5691444231082895, + "eval_sts-test_pearson_max": 0.5691444231082895, + "eval_sts-test_spearman_cosine": 0.6075610251541963, + "eval_sts-test_spearman_dot": 0.3974190005719933, + "eval_sts-test_spearman_euclidean": 0.583691060338115, + "eval_sts-test_spearman_manhattan": 0.5868551535235325, + "eval_sts-test_spearman_max": 0.6075610251541963, + "eval_vitaminc-pairs_loss": 4.039915561676025, + "eval_vitaminc-pairs_runtime": 5.4967, + "eval_vitaminc-pairs_samples_per_second": 23.287, + "eval_vitaminc-pairs_steps_per_second": 0.364, + "step": 260 + }, + { + "epoch": 0.2017067494181536, + "eval_negation-triplets_loss": 3.594820976257324, + "eval_negation-triplets_runtime": 1.1206, + "eval_negation-triplets_samples_per_second": 114.228, + "eval_negation-triplets_steps_per_second": 1.785, + "step": 260 + }, + { + "epoch": 0.2017067494181536, + "eval_scitail-pairs-pos_loss": 0.4375041425228119, + "eval_scitail-pairs-pos_runtime": 1.2948, + "eval_scitail-pairs-pos_samples_per_second": 98.861, + "eval_scitail-pairs-pos_steps_per_second": 1.545, + "step": 260 + }, + { + "epoch": 0.2017067494181536, + "eval_scitail-pairs-qa_loss": 0.6644041538238525, + "eval_scitail-pairs-qa_runtime": 0.7832, + "eval_scitail-pairs-qa_samples_per_second": 163.441, + "eval_scitail-pairs-qa_steps_per_second": 2.554, + "step": 260 + }, + { + "epoch": 0.2017067494181536, + "eval_xsum-pairs_loss": 2.3631696701049805, + "eval_xsum-pairs_runtime": 3.941, + "eval_xsum-pairs_samples_per_second": 32.479, + "eval_xsum-pairs_steps_per_second": 0.507, + "step": 260 + }, + { + "epoch": 0.2017067494181536, + "eval_sciq_pairs_loss": 0.32018786668777466, + "eval_sciq_pairs_runtime": 5.9413, + "eval_sciq_pairs_samples_per_second": 21.544, + "eval_sciq_pairs_steps_per_second": 0.337, + "step": 260 + }, + { + "epoch": 0.2017067494181536, + "eval_qasc_pairs_loss": 2.350424289703369, + "eval_qasc_pairs_runtime": 0.9926, + "eval_qasc_pairs_samples_per_second": 128.957, + "eval_qasc_pairs_steps_per_second": 2.015, + "step": 260 + }, + { + "epoch": 0.2017067494181536, + "eval_openbookqa_pairs_loss": 3.6394641399383545, + "eval_openbookqa_pairs_runtime": 0.8846, + "eval_openbookqa_pairs_samples_per_second": 144.694, + "eval_openbookqa_pairs_steps_per_second": 2.261, + "step": 260 + }, + { + "epoch": 0.2017067494181536, + "eval_msmarco_pairs_loss": 3.4378509521484375, + "eval_msmarco_pairs_runtime": 1.99, + "eval_msmarco_pairs_samples_per_second": 64.32, + "eval_msmarco_pairs_steps_per_second": 1.005, + "step": 260 + }, + { + "epoch": 0.2017067494181536, + "eval_nq_pairs_loss": 3.741243362426758, + "eval_nq_pairs_runtime": 3.3459, + "eval_nq_pairs_samples_per_second": 38.256, + "eval_nq_pairs_steps_per_second": 0.598, + "step": 260 + }, + { + "epoch": 0.2017067494181536, + "eval_trivia_pairs_loss": 3.907733201980591, + "eval_trivia_pairs_runtime": 4.4944, + "eval_trivia_pairs_samples_per_second": 28.48, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 260 + }, + { + "epoch": 0.2017067494181536, + "eval_gooaq_pairs_loss": 3.0075690746307373, + "eval_gooaq_pairs_runtime": 1.5055, + "eval_gooaq_pairs_samples_per_second": 85.024, + "eval_gooaq_pairs_steps_per_second": 1.328, + "step": 260 + }, + { + "epoch": 0.2017067494181536, + "eval_paws-pos_loss": 0.10642223060131073, + "eval_paws-pos_runtime": 1.0293, + "eval_paws-pos_samples_per_second": 124.359, + "eval_paws-pos_steps_per_second": 1.943, + "step": 260 + }, + { + "epoch": 0.2017067494181536, + "eval_global_dataset_loss": 1.825014591217041, + "eval_global_dataset_runtime": 19.4234, + "eval_global_dataset_samples_per_second": 21.417, + "eval_global_dataset_steps_per_second": 0.36, + "step": 260 + }, + { + "epoch": 0.20248254460822343, + "grad_norm": 15.610722541809082, + "learning_rate": 1.0652173913043477e-05, + "loss": 1.776, + "step": 261 + }, + { + "epoch": 0.20325833979829325, + "grad_norm": 15.1345853805542, + "learning_rate": 1.0693301997649823e-05, + "loss": 1.7895, + "step": 262 + }, + { + "epoch": 0.20403413498836306, + "grad_norm": 26.102519989013672, + "learning_rate": 1.0734430082256168e-05, + "loss": 4.0261, + "step": 263 + }, + { + "epoch": 0.20480993017843288, + "grad_norm": 16.856807708740234, + "learning_rate": 1.0775558166862513e-05, + "loss": 2.254, + "step": 264 + }, + { + "epoch": 0.20558572536850273, + "grad_norm": 25.3414363861084, + "learning_rate": 1.0816686251468858e-05, + "loss": 3.6334, + "step": 265 + }, + { + "epoch": 0.20636152055857254, + "grad_norm": 24.477399826049805, + "learning_rate": 1.0857814336075205e-05, + "loss": 4.2001, + "step": 266 + }, + { + "epoch": 0.20713731574864236, + "grad_norm": 14.106359481811523, + "learning_rate": 1.089894242068155e-05, + "loss": 1.7653, + "step": 267 + }, + { + "epoch": 0.20791311093871218, + "grad_norm": 17.06173324584961, + "learning_rate": 1.0940070505287894e-05, + "loss": 3.1158, + "step": 268 + }, + { + "epoch": 0.208688906128782, + "grad_norm": 5.879356384277344, + "learning_rate": 1.0981198589894241e-05, + "loss": 0.2744, + "step": 269 + }, + { + "epoch": 0.20946470131885184, + "grad_norm": 15.685729026794434, + "learning_rate": 1.1022326674500587e-05, + "loss": 2.6071, + "step": 270 + }, + { + "epoch": 0.21024049650892165, + "grad_norm": 18.253646850585938, + "learning_rate": 1.1063454759106932e-05, + "loss": 2.8655, + "step": 271 + }, + { + "epoch": 0.21101629169899147, + "grad_norm": 12.888394355773926, + "learning_rate": 1.1104582843713279e-05, + "loss": 1.5014, + "step": 272 + }, + { + "epoch": 0.2117920868890613, + "grad_norm": 13.16964054107666, + "learning_rate": 1.1145710928319624e-05, + "loss": 2.5295, + "step": 273 + }, + { + "epoch": 0.2125678820791311, + "grad_norm": 25.27763557434082, + "learning_rate": 1.1186839012925968e-05, + "loss": 3.8334, + "step": 274 + }, + { + "epoch": 0.21334367726920092, + "grad_norm": 17.442773818969727, + "learning_rate": 1.1227967097532313e-05, + "loss": 2.7213, + "step": 275 + }, + { + "epoch": 0.21411947245927077, + "grad_norm": 13.96842098236084, + "learning_rate": 1.126909518213866e-05, + "loss": 1.5855, + "step": 276 + }, + { + "epoch": 0.21489526764934058, + "grad_norm": 15.856791496276855, + "learning_rate": 1.1310223266745005e-05, + "loss": 1.6089, + "step": 277 + }, + { + "epoch": 0.2156710628394104, + "grad_norm": 17.355192184448242, + "learning_rate": 1.135135135135135e-05, + "loss": 3.5855, + "step": 278 + }, + { + "epoch": 0.21644685802948022, + "grad_norm": 22.566591262817383, + "learning_rate": 1.1392479435957698e-05, + "loss": 3.4454, + "step": 279 + }, + { + "epoch": 0.21722265321955003, + "grad_norm": 11.724306106567383, + "learning_rate": 1.1433607520564041e-05, + "loss": 1.7671, + "step": 280 + }, + { + "epoch": 0.21722265321955003, + "eval_Qnli-dev_cosine_accuracy": 0.642578125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8725194931030273, + "eval_Qnli-dev_cosine_ap": 0.6480191892664191, + "eval_Qnli-dev_cosine_f1": 0.6541471048513302, + "eval_Qnli-dev_cosine_f1_threshold": 0.7736532688140869, + "eval_Qnli-dev_cosine_precision": 0.5186104218362283, + "eval_Qnli-dev_cosine_recall": 0.885593220338983, + "eval_Qnli-dev_dot_accuracy": 0.623046875, + "eval_Qnli-dev_dot_accuracy_threshold": 561.6754150390625, + "eval_Qnli-dev_dot_ap": 0.594531019849786, + "eval_Qnli-dev_dot_f1": 0.6424418604651162, + "eval_Qnli-dev_dot_f1_threshold": 397.423095703125, + "eval_Qnli-dev_dot_precision": 0.4889380530973451, + "eval_Qnli-dev_dot_recall": 0.9364406779661016, + "eval_Qnli-dev_euclidean_accuracy": 0.63671875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.079384803771973, + "eval_Qnli-dev_euclidean_ap": 0.650209205985763, + "eval_Qnli-dev_euclidean_f1": 0.6486486486486487, + "eval_Qnli-dev_euclidean_f1_threshold": 16.04608154296875, + "eval_Qnli-dev_euclidean_precision": 0.5190839694656488, + "eval_Qnli-dev_euclidean_recall": 0.864406779661017, + "eval_Qnli-dev_manhattan_accuracy": 0.634765625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 235.5088653564453, + "eval_Qnli-dev_manhattan_ap": 0.6486808711899925, + "eval_Qnli-dev_manhattan_f1": 0.6571879936808848, + "eval_Qnli-dev_manhattan_f1_threshold": 284.24969482421875, + "eval_Qnli-dev_manhattan_precision": 0.5239294710327456, + "eval_Qnli-dev_manhattan_recall": 0.8813559322033898, + "eval_Qnli-dev_max_accuracy": 0.642578125, + "eval_Qnli-dev_max_accuracy_threshold": 561.6754150390625, + "eval_Qnli-dev_max_ap": 0.650209205985763, + "eval_Qnli-dev_max_f1": 0.6571879936808848, + "eval_Qnli-dev_max_f1_threshold": 397.423095703125, + "eval_Qnli-dev_max_precision": 0.5239294710327456, + "eval_Qnli-dev_max_recall": 0.9364406779661016, + "eval_allNLI-dev_cosine_accuracy": 0.701171875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.943870484828949, + "eval_allNLI-dev_cosine_ap": 0.5396954038256323, + "eval_allNLI-dev_cosine_f1": 0.5720620842572062, + "eval_allNLI-dev_cosine_f1_threshold": 0.875219464302063, + "eval_allNLI-dev_cosine_precision": 0.46402877697841727, + "eval_allNLI-dev_cosine_recall": 0.7456647398843931, + "eval_allNLI-dev_dot_accuracy": 0.673828125, + "eval_allNLI-dev_dot_accuracy_threshold": 625.1502685546875, + "eval_allNLI-dev_dot_ap": 0.3989484021171416, + "eval_allNLI-dev_dot_f1": 0.5117004680187207, + "eval_allNLI-dev_dot_f1_threshold": 430.39263916015625, + "eval_allNLI-dev_dot_precision": 0.3504273504273504, + "eval_allNLI-dev_dot_recall": 0.9479768786127167, + "eval_allNLI-dev_euclidean_accuracy": 0.69921875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 8.466145515441895, + "eval_allNLI-dev_euclidean_ap": 0.5366056492160509, + "eval_allNLI-dev_euclidean_f1": 0.5630630630630631, + "eval_allNLI-dev_euclidean_f1_threshold": 12.355962753295898, + "eval_allNLI-dev_euclidean_precision": 0.4612546125461255, + "eval_allNLI-dev_euclidean_recall": 0.7225433526011561, + "eval_allNLI-dev_manhattan_accuracy": 0.69140625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 146.6441650390625, + "eval_allNLI-dev_manhattan_ap": 0.5328335627356967, + "eval_allNLI-dev_manhattan_f1": 0.5636743215031316, + "eval_allNLI-dev_manhattan_f1_threshold": 231.57371520996094, + "eval_allNLI-dev_manhattan_precision": 0.4411764705882353, + "eval_allNLI-dev_manhattan_recall": 0.7803468208092486, + "eval_allNLI-dev_max_accuracy": 0.701171875, + "eval_allNLI-dev_max_accuracy_threshold": 625.1502685546875, + "eval_allNLI-dev_max_ap": 0.5396954038256323, + "eval_allNLI-dev_max_f1": 0.5720620842572062, + "eval_allNLI-dev_max_f1_threshold": 430.39263916015625, + "eval_allNLI-dev_max_precision": 0.46402877697841727, + "eval_allNLI-dev_max_recall": 0.9479768786127167, + "eval_sequential_score": 0.650209205985763, + "eval_sts-test_pearson_cosine": 0.6252071183660979, + "eval_sts-test_pearson_dot": 0.4643201935347477, + "eval_sts-test_pearson_euclidean": 0.6463557760014038, + "eval_sts-test_pearson_manhattan": 0.6502994079294852, + "eval_sts-test_pearson_max": 0.6502994079294852, + "eval_sts-test_spearman_cosine": 0.6841740211234761, + "eval_sts-test_spearman_dot": 0.46714038290647875, + "eval_sts-test_spearman_euclidean": 0.6609735799290716, + "eval_sts-test_spearman_manhattan": 0.6662960264513962, + "eval_sts-test_spearman_max": 0.6841740211234761, + "eval_vitaminc-pairs_loss": 3.895996570587158, + "eval_vitaminc-pairs_runtime": 5.474, + "eval_vitaminc-pairs_samples_per_second": 23.383, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 280 + }, + { + "epoch": 0.21722265321955003, + "eval_negation-triplets_loss": 3.2737627029418945, + "eval_negation-triplets_runtime": 1.0689, + "eval_negation-triplets_samples_per_second": 119.754, + "eval_negation-triplets_steps_per_second": 1.871, + "step": 280 + }, + { + "epoch": 0.21722265321955003, + "eval_scitail-pairs-pos_loss": 0.3453359305858612, + "eval_scitail-pairs-pos_runtime": 1.2021, + "eval_scitail-pairs-pos_samples_per_second": 106.484, + "eval_scitail-pairs-pos_steps_per_second": 1.664, + "step": 280 + }, + { + "epoch": 0.21722265321955003, + "eval_scitail-pairs-qa_loss": 0.6447652578353882, + "eval_scitail-pairs-qa_runtime": 0.764, + "eval_scitail-pairs-qa_samples_per_second": 167.533, + "eval_scitail-pairs-qa_steps_per_second": 2.618, + "step": 280 + }, + { + "epoch": 0.21722265321955003, + "eval_xsum-pairs_loss": 1.9839868545532227, + "eval_xsum-pairs_runtime": 3.9145, + "eval_xsum-pairs_samples_per_second": 32.699, + "eval_xsum-pairs_steps_per_second": 0.511, + "step": 280 + }, + { + "epoch": 0.21722265321955003, + "eval_sciq_pairs_loss": 0.27117013931274414, + "eval_sciq_pairs_runtime": 5.8521, + "eval_sciq_pairs_samples_per_second": 21.872, + "eval_sciq_pairs_steps_per_second": 0.342, + "step": 280 + }, + { + "epoch": 0.21722265321955003, + "eval_qasc_pairs_loss": 2.157790422439575, + "eval_qasc_pairs_runtime": 0.9131, + "eval_qasc_pairs_samples_per_second": 140.189, + "eval_qasc_pairs_steps_per_second": 2.19, + "step": 280 + }, + { + "epoch": 0.21722265321955003, + "eval_openbookqa_pairs_loss": 3.370903253555298, + "eval_openbookqa_pairs_runtime": 0.855, + "eval_openbookqa_pairs_samples_per_second": 149.715, + "eval_openbookqa_pairs_steps_per_second": 2.339, + "step": 280 + }, + { + "epoch": 0.21722265321955003, + "eval_msmarco_pairs_loss": 2.6094882488250732, + "eval_msmarco_pairs_runtime": 1.9593, + "eval_msmarco_pairs_samples_per_second": 65.33, + "eval_msmarco_pairs_steps_per_second": 1.021, + "step": 280 + }, + { + "epoch": 0.21722265321955003, + "eval_nq_pairs_loss": 2.909904718399048, + "eval_nq_pairs_runtime": 3.3195, + "eval_nq_pairs_samples_per_second": 38.56, + "eval_nq_pairs_steps_per_second": 0.602, + "step": 280 + }, + { + "epoch": 0.21722265321955003, + "eval_trivia_pairs_loss": 2.8713767528533936, + "eval_trivia_pairs_runtime": 4.4956, + "eval_trivia_pairs_samples_per_second": 28.472, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 280 + }, + { + "epoch": 0.21722265321955003, + "eval_gooaq_pairs_loss": 2.3393373489379883, + "eval_gooaq_pairs_runtime": 1.4723, + "eval_gooaq_pairs_samples_per_second": 86.941, + "eval_gooaq_pairs_steps_per_second": 1.358, + "step": 280 + }, + { + "epoch": 0.21722265321955003, + "eval_paws-pos_loss": 0.11137357354164124, + "eval_paws-pos_runtime": 1.0289, + "eval_paws-pos_samples_per_second": 124.404, + "eval_paws-pos_steps_per_second": 1.944, + "step": 280 + }, + { + "epoch": 0.21722265321955003, + "eval_global_dataset_loss": 1.5072476863861084, + "eval_global_dataset_runtime": 19.4173, + "eval_global_dataset_samples_per_second": 21.424, + "eval_global_dataset_steps_per_second": 0.361, + "step": 280 + }, + { + "epoch": 0.21799844840961985, + "grad_norm": 15.085926055908203, + "learning_rate": 1.1474735605170386e-05, + "loss": 2.5975, + "step": 281 + }, + { + "epoch": 0.2187742435996897, + "grad_norm": 13.402788162231445, + "learning_rate": 1.1515863689776732e-05, + "loss": 0.9452, + "step": 282 + }, + { + "epoch": 0.2195500387897595, + "grad_norm": 14.641647338867188, + "learning_rate": 1.1556991774383079e-05, + "loss": 2.2051, + "step": 283 + }, + { + "epoch": 0.22032583397982933, + "grad_norm": 5.166534900665283, + "learning_rate": 1.1598119858989424e-05, + "loss": 0.2627, + "step": 284 + }, + { + "epoch": 0.22110162916989914, + "grad_norm": 16.76167106628418, + "learning_rate": 1.1639247943595768e-05, + "loss": 1.7162, + "step": 285 + }, + { + "epoch": 0.22187742435996896, + "grad_norm": 14.418363571166992, + "learning_rate": 1.1680376028202113e-05, + "loss": 1.3192, + "step": 286 + }, + { + "epoch": 0.22265321955003878, + "grad_norm": 24.48800277709961, + "learning_rate": 1.172150411280846e-05, + "loss": 3.5457, + "step": 287 + }, + { + "epoch": 0.22342901474010862, + "grad_norm": 15.297579765319824, + "learning_rate": 1.1762632197414805e-05, + "loss": 1.3446, + "step": 288 + }, + { + "epoch": 0.22420480993017844, + "grad_norm": 23.22272491455078, + "learning_rate": 1.180376028202115e-05, + "loss": 2.9348, + "step": 289 + }, + { + "epoch": 0.22498060512024826, + "grad_norm": 17.5606632232666, + "learning_rate": 1.1844888366627497e-05, + "loss": 1.6742, + "step": 290 + }, + { + "epoch": 0.22575640031031807, + "grad_norm": 15.780293464660645, + "learning_rate": 1.1886016451233841e-05, + "loss": 1.5353, + "step": 291 + }, + { + "epoch": 0.2265321955003879, + "grad_norm": 13.763825416564941, + "learning_rate": 1.1927144535840186e-05, + "loss": 1.1798, + "step": 292 + }, + { + "epoch": 0.2273079906904577, + "grad_norm": 16.144515991210938, + "learning_rate": 1.1968272620446532e-05, + "loss": 1.3208, + "step": 293 + }, + { + "epoch": 0.22808378588052755, + "grad_norm": 14.696866035461426, + "learning_rate": 1.2009400705052879e-05, + "loss": 1.0856, + "step": 294 + }, + { + "epoch": 0.22885958107059737, + "grad_norm": 20.401079177856445, + "learning_rate": 1.2050528789659224e-05, + "loss": 2.5945, + "step": 295 + }, + { + "epoch": 0.22963537626066718, + "grad_norm": 14.700536727905273, + "learning_rate": 1.209165687426557e-05, + "loss": 1.2534, + "step": 296 + }, + { + "epoch": 0.230411171450737, + "grad_norm": 29.061100006103516, + "learning_rate": 1.2132784958871913e-05, + "loss": 4.5888, + "step": 297 + }, + { + "epoch": 0.23118696664080682, + "grad_norm": 14.214822769165039, + "learning_rate": 1.217391304347826e-05, + "loss": 1.5575, + "step": 298 + }, + { + "epoch": 0.23196276183087663, + "grad_norm": 13.572636604309082, + "learning_rate": 1.2215041128084605e-05, + "loss": 1.9698, + "step": 299 + }, + { + "epoch": 0.23273855702094648, + "grad_norm": 17.725801467895508, + "learning_rate": 1.225616921269095e-05, + "loss": 2.5515, + "step": 300 + }, + { + "epoch": 0.23273855702094648, + "eval_Qnli-dev_cosine_accuracy": 0.646484375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.895124077796936, + "eval_Qnli-dev_cosine_ap": 0.6528783052084683, + "eval_Qnli-dev_cosine_f1": 0.6590909090909091, + "eval_Qnli-dev_cosine_f1_threshold": 0.7926325798034668, + "eval_Qnli-dev_cosine_precision": 0.5342105263157895, + "eval_Qnli-dev_cosine_recall": 0.8601694915254238, + "eval_Qnli-dev_dot_accuracy": 0.619140625, + "eval_Qnli-dev_dot_accuracy_threshold": 549.4300537109375, + "eval_Qnli-dev_dot_ap": 0.5900595040342307, + "eval_Qnli-dev_dot_f1": 0.6406926406926408, + "eval_Qnli-dev_dot_f1_threshold": 392.31390380859375, + "eval_Qnli-dev_dot_precision": 0.48577680525164113, + "eval_Qnli-dev_dot_recall": 0.940677966101695, + "eval_Qnli-dev_euclidean_accuracy": 0.64453125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.800490379333496, + "eval_Qnli-dev_euclidean_ap": 0.6558463098189781, + "eval_Qnli-dev_euclidean_f1": 0.6517189835575486, + "eval_Qnli-dev_euclidean_f1_threshold": 17.142818450927734, + "eval_Qnli-dev_euclidean_precision": 0.5034642032332564, + "eval_Qnli-dev_euclidean_recall": 0.923728813559322, + "eval_Qnli-dev_manhattan_accuracy": 0.634765625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 238.0390625, + "eval_Qnli-dev_manhattan_ap": 0.654006014265549, + "eval_Qnli-dev_manhattan_f1": 0.6537313432835821, + "eval_Qnli-dev_manhattan_f1_threshold": 303.4346618652344, + "eval_Qnli-dev_manhattan_precision": 0.5046082949308756, + "eval_Qnli-dev_manhattan_recall": 0.9279661016949152, + "eval_Qnli-dev_max_accuracy": 0.646484375, + "eval_Qnli-dev_max_accuracy_threshold": 549.4300537109375, + "eval_Qnli-dev_max_ap": 0.6558463098189781, + "eval_Qnli-dev_max_f1": 0.6590909090909091, + "eval_Qnli-dev_max_f1_threshold": 392.31390380859375, + "eval_Qnli-dev_max_precision": 0.5342105263157895, + "eval_Qnli-dev_max_recall": 0.940677966101695, + "eval_allNLI-dev_cosine_accuracy": 0.703125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.932201623916626, + "eval_allNLI-dev_cosine_ap": 0.5579356554444865, + "eval_allNLI-dev_cosine_f1": 0.5611814345991561, + "eval_allNLI-dev_cosine_f1_threshold": 0.8481569290161133, + "eval_allNLI-dev_cosine_precision": 0.4418604651162791, + "eval_allNLI-dev_cosine_recall": 0.7687861271676301, + "eval_allNLI-dev_dot_accuracy": 0.66796875, + "eval_allNLI-dev_dot_accuracy_threshold": 608.14404296875, + "eval_allNLI-dev_dot_ap": 0.4095525177799477, + "eval_allNLI-dev_dot_f1": 0.5193199381761978, + "eval_allNLI-dev_dot_f1_threshold": 398.32855224609375, + "eval_allNLI-dev_dot_precision": 0.35443037974683544, + "eval_allNLI-dev_dot_recall": 0.9710982658959537, + "eval_allNLI-dev_euclidean_accuracy": 0.705078125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 9.391706466674805, + "eval_allNLI-dev_euclidean_ap": 0.552362196578927, + "eval_allNLI-dev_euclidean_f1": 0.5685483870967741, + "eval_allNLI-dev_euclidean_f1_threshold": 13.739648818969727, + "eval_allNLI-dev_euclidean_precision": 0.43653250773993807, + "eval_allNLI-dev_euclidean_recall": 0.815028901734104, + "eval_allNLI-dev_manhattan_accuracy": 0.703125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 178.74974060058594, + "eval_allNLI-dev_manhattan_ap": 0.5496116812305261, + "eval_allNLI-dev_manhattan_f1": 0.576271186440678, + "eval_allNLI-dev_manhattan_f1_threshold": 236.75367736816406, + "eval_allNLI-dev_manhattan_precision": 0.45484949832775917, + "eval_allNLI-dev_manhattan_recall": 0.7861271676300579, + "eval_allNLI-dev_max_accuracy": 0.705078125, + "eval_allNLI-dev_max_accuracy_threshold": 608.14404296875, + "eval_allNLI-dev_max_ap": 0.5579356554444865, + "eval_allNLI-dev_max_f1": 0.576271186440678, + "eval_allNLI-dev_max_f1_threshold": 398.32855224609375, + "eval_allNLI-dev_max_precision": 0.45484949832775917, + "eval_allNLI-dev_max_recall": 0.9710982658959537, + "eval_sequential_score": 0.6558463098189781, + "eval_sts-test_pearson_cosine": 0.6778653196849711, + "eval_sts-test_pearson_dot": 0.5256481184350303, + "eval_sts-test_pearson_euclidean": 0.7047971077534723, + "eval_sts-test_pearson_manhattan": 0.7076251525623551, + "eval_sts-test_pearson_max": 0.7076251525623551, + "eval_sts-test_spearman_cosine": 0.7378726259611652, + "eval_sts-test_spearman_dot": 0.5280127939502546, + "eval_sts-test_spearman_euclidean": 0.7186699492006384, + "eval_sts-test_spearman_manhattan": 0.7217660204687409, + "eval_sts-test_spearman_max": 0.7378726259611652, + "eval_vitaminc-pairs_loss": 3.769603967666626, + "eval_vitaminc-pairs_runtime": 5.507, + "eval_vitaminc-pairs_samples_per_second": 23.243, + "eval_vitaminc-pairs_steps_per_second": 0.363, + "step": 300 + }, + { + "epoch": 0.23273855702094648, + "eval_negation-triplets_loss": 3.0029027462005615, + "eval_negation-triplets_runtime": 1.0811, + "eval_negation-triplets_samples_per_second": 118.398, + "eval_negation-triplets_steps_per_second": 1.85, + "step": 300 + }, + { + "epoch": 0.23273855702094648, + "eval_scitail-pairs-pos_loss": 0.2898590862751007, + "eval_scitail-pairs-pos_runtime": 1.3081, + "eval_scitail-pairs-pos_samples_per_second": 97.854, + "eval_scitail-pairs-pos_steps_per_second": 1.529, + "step": 300 + }, + { + "epoch": 0.23273855702094648, + "eval_scitail-pairs-qa_loss": 0.48321533203125, + "eval_scitail-pairs-qa_runtime": 0.8138, + "eval_scitail-pairs-qa_samples_per_second": 157.278, + "eval_scitail-pairs-qa_steps_per_second": 2.457, + "step": 300 + }, + { + "epoch": 0.23273855702094648, + "eval_xsum-pairs_loss": 1.8343725204467773, + "eval_xsum-pairs_runtime": 3.9405, + "eval_xsum-pairs_samples_per_second": 32.483, + "eval_xsum-pairs_steps_per_second": 0.508, + "step": 300 + }, + { + "epoch": 0.23273855702094648, + "eval_sciq_pairs_loss": 0.22236409783363342, + "eval_sciq_pairs_runtime": 5.8846, + "eval_sciq_pairs_samples_per_second": 21.752, + "eval_sciq_pairs_steps_per_second": 0.34, + "step": 300 + }, + { + "epoch": 0.23273855702094648, + "eval_qasc_pairs_loss": 2.026122808456421, + "eval_qasc_pairs_runtime": 0.9092, + "eval_qasc_pairs_samples_per_second": 140.781, + "eval_qasc_pairs_steps_per_second": 2.2, + "step": 300 + }, + { + "epoch": 0.23273855702094648, + "eval_openbookqa_pairs_loss": 3.074580669403076, + "eval_openbookqa_pairs_runtime": 0.8609, + "eval_openbookqa_pairs_samples_per_second": 148.688, + "eval_openbookqa_pairs_steps_per_second": 2.323, + "step": 300 + }, + { + "epoch": 0.23273855702094648, + "eval_msmarco_pairs_loss": 2.1841843128204346, + "eval_msmarco_pairs_runtime": 1.9699, + "eval_msmarco_pairs_samples_per_second": 64.979, + "eval_msmarco_pairs_steps_per_second": 1.015, + "step": 300 + }, + { + "epoch": 0.23273855702094648, + "eval_nq_pairs_loss": 2.703659772872925, + "eval_nq_pairs_runtime": 3.3561, + "eval_nq_pairs_samples_per_second": 38.139, + "eval_nq_pairs_steps_per_second": 0.596, + "step": 300 + }, + { + "epoch": 0.23273855702094648, + "eval_trivia_pairs_loss": 2.679865837097168, + "eval_trivia_pairs_runtime": 4.4691, + "eval_trivia_pairs_samples_per_second": 28.641, + "eval_trivia_pairs_steps_per_second": 0.448, + "step": 300 + }, + { + "epoch": 0.23273855702094648, + "eval_gooaq_pairs_loss": 2.084993839263916, + "eval_gooaq_pairs_runtime": 1.4582, + "eval_gooaq_pairs_samples_per_second": 87.778, + "eval_gooaq_pairs_steps_per_second": 1.372, + "step": 300 + }, + { + "epoch": 0.23273855702094648, + "eval_paws-pos_loss": 0.10109881311655045, + "eval_paws-pos_runtime": 1.0293, + "eval_paws-pos_samples_per_second": 124.361, + "eval_paws-pos_steps_per_second": 1.943, + "step": 300 + }, + { + "epoch": 0.23273855702094648, + "eval_global_dataset_loss": 1.313809871673584, + "eval_global_dataset_runtime": 19.3851, + "eval_global_dataset_samples_per_second": 21.46, + "eval_global_dataset_steps_per_second": 0.361, + "step": 300 + }, + { + "epoch": 0.2335143522110163, + "grad_norm": 24.429811477661133, + "learning_rate": 1.2297297297297297e-05, + "loss": 3.6509, + "step": 301 + }, + { + "epoch": 0.2342901474010861, + "grad_norm": 14.498895645141602, + "learning_rate": 1.2338425381903641e-05, + "loss": 1.2198, + "step": 302 + }, + { + "epoch": 0.23506594259115593, + "grad_norm": 5.096879959106445, + "learning_rate": 1.2379553466509986e-05, + "loss": 0.192, + "step": 303 + }, + { + "epoch": 0.23584173778122575, + "grad_norm": 26.32643699645996, + "learning_rate": 1.2420681551116333e-05, + "loss": 3.4814, + "step": 304 + }, + { + "epoch": 0.2366175329712956, + "grad_norm": 12.396186828613281, + "learning_rate": 1.2461809635722679e-05, + "loss": 1.3822, + "step": 305 + }, + { + "epoch": 0.2373933281613654, + "grad_norm": 5.122003555297852, + "learning_rate": 1.2502937720329024e-05, + "loss": 0.2507, + "step": 306 + }, + { + "epoch": 0.23816912335143522, + "grad_norm": 13.125459671020508, + "learning_rate": 1.2544065804935369e-05, + "loss": 1.9133, + "step": 307 + }, + { + "epoch": 0.23894491854150504, + "grad_norm": Infinity, + "learning_rate": 1.2544065804935369e-05, + "loss": 0.7287, + "step": 308 + }, + { + "epoch": 0.23972071373157486, + "grad_norm": 16.207378387451172, + "learning_rate": 1.2585193889541714e-05, + "loss": 1.6418, + "step": 309 + }, + { + "epoch": 0.24049650892164467, + "grad_norm": 18.47272300720215, + "learning_rate": 1.262632197414806e-05, + "loss": 2.1207, + "step": 310 + }, + { + "epoch": 0.24127230411171452, + "grad_norm": 13.787567138671875, + "learning_rate": 1.2667450058754405e-05, + "loss": 1.178, + "step": 311 + }, + { + "epoch": 0.24204809930178434, + "grad_norm": 26.077777862548828, + "learning_rate": 1.2708578143360752e-05, + "loss": 3.5281, + "step": 312 + }, + { + "epoch": 0.24282389449185415, + "grad_norm": 16.01590347290039, + "learning_rate": 1.2749706227967097e-05, + "loss": 2.2668, + "step": 313 + }, + { + "epoch": 0.24359968968192397, + "grad_norm": 15.849452018737793, + "learning_rate": 1.2790834312573443e-05, + "loss": 1.5443, + "step": 314 + }, + { + "epoch": 0.2443754848719938, + "grad_norm": 3.772362470626831, + "learning_rate": 1.2831962397179786e-05, + "loss": 0.1658, + "step": 315 + }, + { + "epoch": 0.2451512800620636, + "grad_norm": 13.069218635559082, + "learning_rate": 1.2873090481786133e-05, + "loss": 1.0262, + "step": 316 + }, + { + "epoch": 0.24592707525213345, + "grad_norm": 13.916678428649902, + "learning_rate": 1.2914218566392478e-05, + "loss": 1.0937, + "step": 317 + }, + { + "epoch": 0.24670287044220326, + "grad_norm": 11.905617713928223, + "learning_rate": 1.2955346650998824e-05, + "loss": 0.9618, + "step": 318 + }, + { + "epoch": 0.24747866563227308, + "grad_norm": 14.229816436767578, + "learning_rate": 1.299647473560517e-05, + "loss": 1.2004, + "step": 319 + }, + { + "epoch": 0.2482544608223429, + "grad_norm": 17.332258224487305, + "learning_rate": 1.3037602820211514e-05, + "loss": 1.9246, + "step": 320 + }, + { + "epoch": 0.2482544608223429, + "eval_Qnli-dev_cosine_accuracy": 0.654296875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8905587196350098, + "eval_Qnli-dev_cosine_ap": 0.6566239311243058, + "eval_Qnli-dev_cosine_f1": 0.6535947712418301, + "eval_Qnli-dev_cosine_f1_threshold": 0.7684817314147949, + "eval_Qnli-dev_cosine_precision": 0.5319148936170213, + "eval_Qnli-dev_cosine_recall": 0.847457627118644, + "eval_Qnli-dev_dot_accuracy": 0.650390625, + "eval_Qnli-dev_dot_accuracy_threshold": 500.9089660644531, + "eval_Qnli-dev_dot_ap": 0.614512025409862, + "eval_Qnli-dev_dot_f1": 0.6458333333333333, + "eval_Qnli-dev_dot_f1_threshold": 391.3675842285156, + "eval_Qnli-dev_dot_precision": 0.49770642201834864, + "eval_Qnli-dev_dot_recall": 0.9194915254237288, + "eval_Qnli-dev_euclidean_accuracy": 0.654296875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.877283096313477, + "eval_Qnli-dev_euclidean_ap": 0.6557982087274061, + "eval_Qnli-dev_euclidean_f1": 0.6550632911392406, + "eval_Qnli-dev_euclidean_f1_threshold": 16.702594757080078, + "eval_Qnli-dev_euclidean_precision": 0.5227272727272727, + "eval_Qnli-dev_euclidean_recall": 0.8771186440677966, + "eval_Qnli-dev_manhattan_accuracy": 0.6484375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 231.19979858398438, + "eval_Qnli-dev_manhattan_ap": 0.656366538019928, + "eval_Qnli-dev_manhattan_f1": 0.6552706552706552, + "eval_Qnli-dev_manhattan_f1_threshold": 338.2535705566406, + "eval_Qnli-dev_manhattan_precision": 0.49356223175965663, + "eval_Qnli-dev_manhattan_recall": 0.9745762711864406, + "eval_Qnli-dev_max_accuracy": 0.654296875, + "eval_Qnli-dev_max_accuracy_threshold": 500.9089660644531, + "eval_Qnli-dev_max_ap": 0.6566239311243058, + "eval_Qnli-dev_max_f1": 0.6552706552706552, + "eval_Qnli-dev_max_f1_threshold": 391.3675842285156, + "eval_Qnli-dev_max_precision": 0.5319148936170213, + "eval_Qnli-dev_max_recall": 0.9745762711864406, + "eval_allNLI-dev_cosine_accuracy": 0.701171875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9404147863388062, + "eval_allNLI-dev_cosine_ap": 0.5579463691828763, + "eval_allNLI-dev_cosine_f1": 0.5701357466063348, + "eval_allNLI-dev_cosine_f1_threshold": 0.8488144874572754, + "eval_allNLI-dev_cosine_precision": 0.4684014869888476, + "eval_allNLI-dev_cosine_recall": 0.7283236994219653, + "eval_allNLI-dev_dot_accuracy": 0.671875, + "eval_allNLI-dev_dot_accuracy_threshold": 551.6342163085938, + "eval_allNLI-dev_dot_ap": 0.43378774860429486, + "eval_allNLI-dev_dot_f1": 0.5163297045101088, + "eval_allNLI-dev_dot_f1_threshold": 377.14434814453125, + "eval_allNLI-dev_dot_precision": 0.35319148936170214, + "eval_allNLI-dev_dot_recall": 0.9595375722543352, + "eval_allNLI-dev_euclidean_accuracy": 0.703125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 8.960792541503906, + "eval_allNLI-dev_euclidean_ap": 0.5546275802890028, + "eval_allNLI-dev_euclidean_f1": 0.587962962962963, + "eval_allNLI-dev_euclidean_f1_threshold": 12.992522239685059, + "eval_allNLI-dev_euclidean_precision": 0.49034749034749037, + "eval_allNLI-dev_euclidean_recall": 0.7341040462427746, + "eval_allNLI-dev_manhattan_accuracy": 0.705078125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 164.56690979003906, + "eval_allNLI-dev_manhattan_ap": 0.5514099685309427, + "eval_allNLI-dev_manhattan_f1": 0.5848214285714286, + "eval_allNLI-dev_manhattan_f1_threshold": 235.60809326171875, + "eval_allNLI-dev_manhattan_precision": 0.4763636363636364, + "eval_allNLI-dev_manhattan_recall": 0.7572254335260116, + "eval_allNLI-dev_max_accuracy": 0.705078125, + "eval_allNLI-dev_max_accuracy_threshold": 551.6342163085938, + "eval_allNLI-dev_max_ap": 0.5579463691828763, + "eval_allNLI-dev_max_f1": 0.587962962962963, + "eval_allNLI-dev_max_f1_threshold": 377.14434814453125, + "eval_allNLI-dev_max_precision": 0.49034749034749037, + "eval_allNLI-dev_max_recall": 0.9595375722543352, + "eval_sequential_score": 0.6566239311243058, + "eval_sts-test_pearson_cosine": 0.7117567998089547, + "eval_sts-test_pearson_dot": 0.5733361964341069, + "eval_sts-test_pearson_euclidean": 0.7403010779801018, + "eval_sts-test_pearson_manhattan": 0.7435587607796431, + "eval_sts-test_pearson_max": 0.7435587607796431, + "eval_sts-test_spearman_cosine": 0.7663937236274388, + "eval_sts-test_spearman_dot": 0.568597265393501, + "eval_sts-test_spearman_euclidean": 0.7497347533156606, + "eval_sts-test_spearman_manhattan": 0.7521505159080911, + "eval_sts-test_spearman_max": 0.7663937236274388, + "eval_vitaminc-pairs_loss": 3.890385866165161, + "eval_vitaminc-pairs_runtime": 5.4819, + "eval_vitaminc-pairs_samples_per_second": 23.35, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 320 + }, + { + "epoch": 0.2482544608223429, + "eval_negation-triplets_loss": 2.72622013092041, + "eval_negation-triplets_runtime": 1.1205, + "eval_negation-triplets_samples_per_second": 114.23, + "eval_negation-triplets_steps_per_second": 1.785, + "step": 320 + }, + { + "epoch": 0.2482544608223429, + "eval_scitail-pairs-pos_loss": 0.2762339115142822, + "eval_scitail-pairs-pos_runtime": 1.2581, + "eval_scitail-pairs-pos_samples_per_second": 101.741, + "eval_scitail-pairs-pos_steps_per_second": 1.59, + "step": 320 + }, + { + "epoch": 0.2482544608223429, + "eval_scitail-pairs-qa_loss": 0.45161712169647217, + "eval_scitail-pairs-qa_runtime": 0.7692, + "eval_scitail-pairs-qa_samples_per_second": 166.412, + "eval_scitail-pairs-qa_steps_per_second": 2.6, + "step": 320 + }, + { + "epoch": 0.2482544608223429, + "eval_xsum-pairs_loss": 1.5600008964538574, + "eval_xsum-pairs_runtime": 3.9187, + "eval_xsum-pairs_samples_per_second": 32.664, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 320 + }, + { + "epoch": 0.2482544608223429, + "eval_sciq_pairs_loss": 0.19931760430335999, + "eval_sciq_pairs_runtime": 5.9074, + "eval_sciq_pairs_samples_per_second": 21.668, + "eval_sciq_pairs_steps_per_second": 0.339, + "step": 320 + }, + { + "epoch": 0.2482544608223429, + "eval_qasc_pairs_loss": 1.9438467025756836, + "eval_qasc_pairs_runtime": 0.9129, + "eval_qasc_pairs_samples_per_second": 140.208, + "eval_qasc_pairs_steps_per_second": 2.191, + "step": 320 + }, + { + "epoch": 0.2482544608223429, + "eval_openbookqa_pairs_loss": 2.985114812850952, + "eval_openbookqa_pairs_runtime": 0.8708, + "eval_openbookqa_pairs_samples_per_second": 146.986, + "eval_openbookqa_pairs_steps_per_second": 2.297, + "step": 320 + }, + { + "epoch": 0.2482544608223429, + "eval_msmarco_pairs_loss": 1.8927481174468994, + "eval_msmarco_pairs_runtime": 1.9678, + "eval_msmarco_pairs_samples_per_second": 65.047, + "eval_msmarco_pairs_steps_per_second": 1.016, + "step": 320 + }, + { + "epoch": 0.2482544608223429, + "eval_nq_pairs_loss": 2.5300557613372803, + "eval_nq_pairs_runtime": 3.327, + "eval_nq_pairs_samples_per_second": 38.473, + "eval_nq_pairs_steps_per_second": 0.601, + "step": 320 + }, + { + "epoch": 0.2482544608223429, + "eval_trivia_pairs_loss": 2.241274356842041, + "eval_trivia_pairs_runtime": 4.4728, + "eval_trivia_pairs_samples_per_second": 28.617, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 320 + }, + { + "epoch": 0.2482544608223429, + "eval_gooaq_pairs_loss": 1.903867483139038, + "eval_gooaq_pairs_runtime": 1.4588, + "eval_gooaq_pairs_samples_per_second": 87.746, + "eval_gooaq_pairs_steps_per_second": 1.371, + "step": 320 + }, + { + "epoch": 0.2482544608223429, + "eval_paws-pos_loss": 0.07195574045181274, + "eval_paws-pos_runtime": 1.0296, + "eval_paws-pos_samples_per_second": 124.318, + "eval_paws-pos_steps_per_second": 1.942, + "step": 320 + }, + { + "epoch": 0.2482544608223429, + "eval_global_dataset_loss": 1.1946886777877808, + "eval_global_dataset_runtime": 19.4048, + "eval_global_dataset_samples_per_second": 21.438, + "eval_global_dataset_steps_per_second": 0.361, + "step": 320 + }, + { + "epoch": 0.24903025601241272, + "grad_norm": 20.10428237915039, + "learning_rate": 1.307873090481786e-05, + "loss": 2.431, + "step": 321 + }, + { + "epoch": 0.24980605120248253, + "grad_norm": 23.540996551513672, + "learning_rate": 1.3119858989424205e-05, + "loss": 2.7341, + "step": 322 + }, + { + "epoch": 0.2505818463925524, + "grad_norm": 17.5711727142334, + "learning_rate": 1.3160987074030552e-05, + "loss": 2.1992, + "step": 323 + }, + { + "epoch": 0.2513576415826222, + "grad_norm": 15.386451721191406, + "learning_rate": 1.3202115158636897e-05, + "loss": 1.7715, + "step": 324 + }, + { + "epoch": 0.252133436772692, + "grad_norm": 15.816802978515625, + "learning_rate": 1.3243243243243242e-05, + "loss": 2.2684, + "step": 325 + }, + { + "epoch": 0.2529092319627618, + "grad_norm": 13.765118598937988, + "learning_rate": 1.3284371327849586e-05, + "loss": 0.9866, + "step": 326 + }, + { + "epoch": 0.25368502715283164, + "grad_norm": 14.085391998291016, + "learning_rate": 1.3325499412455933e-05, + "loss": 1.662, + "step": 327 + }, + { + "epoch": 0.25446082234290146, + "grad_norm": 17.159772872924805, + "learning_rate": 1.3366627497062278e-05, + "loss": 2.5149, + "step": 328 + }, + { + "epoch": 0.2552366175329713, + "grad_norm": 21.812040328979492, + "learning_rate": 1.3407755581668624e-05, + "loss": 2.1649, + "step": 329 + }, + { + "epoch": 0.2560124127230411, + "grad_norm": 14.715442657470703, + "learning_rate": 1.344888366627497e-05, + "loss": 1.346, + "step": 330 + }, + { + "epoch": 0.25678820791311097, + "grad_norm": 11.09142017364502, + "learning_rate": 1.3490011750881316e-05, + "loss": 0.5911, + "step": 331 + }, + { + "epoch": 0.2575640031031808, + "grad_norm": 10.195703506469727, + "learning_rate": 1.353113983548766e-05, + "loss": 0.5095, + "step": 332 + }, + { + "epoch": 0.2583397982932506, + "grad_norm": 18.701921463012695, + "learning_rate": 1.3572267920094005e-05, + "loss": 1.5759, + "step": 333 + }, + { + "epoch": 0.2591155934833204, + "grad_norm": 18.979888916015625, + "learning_rate": 1.3613396004700352e-05, + "loss": 1.9982, + "step": 334 + }, + { + "epoch": 0.25989138867339023, + "grad_norm": 12.601350784301758, + "learning_rate": 1.3654524089306697e-05, + "loss": 0.9951, + "step": 335 + }, + { + "epoch": 0.26066718386346005, + "grad_norm": 13.246020317077637, + "learning_rate": 1.3695652173913042e-05, + "loss": 1.0595, + "step": 336 + }, + { + "epoch": 0.26144297905352987, + "grad_norm": 14.633255958557129, + "learning_rate": 1.373678025851939e-05, + "loss": 1.1949, + "step": 337 + }, + { + "epoch": 0.2622187742435997, + "grad_norm": 23.022418975830078, + "learning_rate": 1.3777908343125733e-05, + "loss": 2.5501, + "step": 338 + }, + { + "epoch": 0.2629945694336695, + "grad_norm": 11.299046516418457, + "learning_rate": 1.3819036427732078e-05, + "loss": 0.6162, + "step": 339 + }, + { + "epoch": 0.2637703646237393, + "grad_norm": 17.452680587768555, + "learning_rate": 1.3860164512338423e-05, + "loss": 2.5935, + "step": 340 + }, + { + "epoch": 0.2637703646237393, + "eval_Qnli-dev_cosine_accuracy": 0.654296875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8862133026123047, + "eval_Qnli-dev_cosine_ap": 0.66756403549097, + "eval_Qnli-dev_cosine_f1": 0.6584234930448223, + "eval_Qnli-dev_cosine_f1_threshold": 0.7246769666671753, + "eval_Qnli-dev_cosine_precision": 0.5182481751824818, + "eval_Qnli-dev_cosine_recall": 0.902542372881356, + "eval_Qnli-dev_dot_accuracy": 0.6171875, + "eval_Qnli-dev_dot_accuracy_threshold": 484.943359375, + "eval_Qnli-dev_dot_ap": 0.5978698374693151, + "eval_Qnli-dev_dot_f1": 0.6473029045643155, + "eval_Qnli-dev_dot_f1_threshold": 315.24456787109375, + "eval_Qnli-dev_dot_precision": 0.4804928131416838, + "eval_Qnli-dev_dot_recall": 0.9915254237288136, + "eval_Qnli-dev_euclidean_accuracy": 0.66015625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.338907241821289, + "eval_Qnli-dev_euclidean_ap": 0.671692842222398, + "eval_Qnli-dev_euclidean_f1": 0.667741935483871, + "eval_Qnli-dev_euclidean_f1_threshold": 16.789207458496094, + "eval_Qnli-dev_euclidean_precision": 0.5390625, + "eval_Qnli-dev_euclidean_recall": 0.8771186440677966, + "eval_Qnli-dev_manhattan_accuracy": 0.65234375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 236.0693817138672, + "eval_Qnli-dev_manhattan_ap": 0.6684108051825812, + "eval_Qnli-dev_manhattan_f1": 0.6613672496025437, + "eval_Qnli-dev_manhattan_f1_threshold": 300.67083740234375, + "eval_Qnli-dev_manhattan_precision": 0.5292620865139949, + "eval_Qnli-dev_manhattan_recall": 0.8813559322033898, + "eval_Qnli-dev_max_accuracy": 0.66015625, + "eval_Qnli-dev_max_accuracy_threshold": 484.943359375, + "eval_Qnli-dev_max_ap": 0.671692842222398, + "eval_Qnli-dev_max_f1": 0.667741935483871, + "eval_Qnli-dev_max_f1_threshold": 315.24456787109375, + "eval_Qnli-dev_max_precision": 0.5390625, + "eval_Qnli-dev_max_recall": 0.9915254237288136, + "eval_allNLI-dev_cosine_accuracy": 0.705078125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9073655605316162, + "eval_allNLI-dev_cosine_ap": 0.5564225469981687, + "eval_allNLI-dev_cosine_f1": 0.5771543086172345, + "eval_allNLI-dev_cosine_f1_threshold": 0.8149856328964233, + "eval_allNLI-dev_cosine_precision": 0.44171779141104295, + "eval_allNLI-dev_cosine_recall": 0.8323699421965318, + "eval_allNLI-dev_dot_accuracy": 0.673828125, + "eval_allNLI-dev_dot_accuracy_threshold": 569.36572265625, + "eval_allNLI-dev_dot_ap": 0.4263791502380346, + "eval_allNLI-dev_dot_f1": 0.5185185185185185, + "eval_allNLI-dev_dot_f1_threshold": 372.38739013671875, + "eval_allNLI-dev_dot_precision": 0.35368421052631577, + "eval_allNLI-dev_dot_recall": 0.9710982658959537, + "eval_allNLI-dev_euclidean_accuracy": 0.701171875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 8.666521072387695, + "eval_allNLI-dev_euclidean_ap": 0.5527711745598307, + "eval_allNLI-dev_euclidean_f1": 0.58, + "eval_allNLI-dev_euclidean_f1_threshold": 14.50294303894043, + "eval_allNLI-dev_euclidean_precision": 0.4434250764525994, + "eval_allNLI-dev_euclidean_recall": 0.838150289017341, + "eval_allNLI-dev_manhattan_accuracy": 0.69921875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 172.1617431640625, + "eval_allNLI-dev_manhattan_ap": 0.549699669805767, + "eval_allNLI-dev_manhattan_f1": 0.57847533632287, + "eval_allNLI-dev_manhattan_f1_threshold": 237.099609375, + "eval_allNLI-dev_manhattan_precision": 0.4725274725274725, + "eval_allNLI-dev_manhattan_recall": 0.7456647398843931, + "eval_allNLI-dev_max_accuracy": 0.705078125, + "eval_allNLI-dev_max_accuracy_threshold": 569.36572265625, + "eval_allNLI-dev_max_ap": 0.5564225469981687, + "eval_allNLI-dev_max_f1": 0.58, + "eval_allNLI-dev_max_f1_threshold": 372.38739013671875, + "eval_allNLI-dev_max_precision": 0.4725274725274725, + "eval_allNLI-dev_max_recall": 0.9710982658959537, + "eval_sequential_score": 0.671692842222398, + "eval_sts-test_pearson_cosine": 0.7243956628134518, + "eval_sts-test_pearson_dot": 0.5891127817185653, + "eval_sts-test_pearson_euclidean": 0.7563072643530611, + "eval_sts-test_pearson_manhattan": 0.7595492842385803, + "eval_sts-test_pearson_max": 0.7595492842385803, + "eval_sts-test_spearman_cosine": 0.7786661305463707, + "eval_sts-test_spearman_dot": 0.5749737383753772, + "eval_sts-test_spearman_euclidean": 0.7658272621607465, + "eval_sts-test_spearman_manhattan": 0.7665132094953538, + "eval_sts-test_spearman_max": 0.7786661305463707, + "eval_vitaminc-pairs_loss": 4.09152364730835, + "eval_vitaminc-pairs_runtime": 5.4593, + "eval_vitaminc-pairs_samples_per_second": 23.446, + "eval_vitaminc-pairs_steps_per_second": 0.366, + "step": 340 + }, + { + "epoch": 0.2637703646237393, + "eval_negation-triplets_loss": 2.6371383666992188, + "eval_negation-triplets_runtime": 1.0786, + "eval_negation-triplets_samples_per_second": 118.673, + "eval_negation-triplets_steps_per_second": 1.854, + "step": 340 + }, + { + "epoch": 0.2637703646237393, + "eval_scitail-pairs-pos_loss": 0.2726733982563019, + "eval_scitail-pairs-pos_runtime": 1.3003, + "eval_scitail-pairs-pos_samples_per_second": 98.437, + "eval_scitail-pairs-pos_steps_per_second": 1.538, + "step": 340 + }, + { + "epoch": 0.2637703646237393, + "eval_scitail-pairs-qa_loss": 0.29476043581962585, + "eval_scitail-pairs-qa_runtime": 0.7756, + "eval_scitail-pairs-qa_samples_per_second": 165.044, + "eval_scitail-pairs-qa_steps_per_second": 2.579, + "step": 340 + }, + { + "epoch": 0.2637703646237393, + "eval_xsum-pairs_loss": 1.1797327995300293, + "eval_xsum-pairs_runtime": 3.926, + "eval_xsum-pairs_samples_per_second": 32.603, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 340 + }, + { + "epoch": 0.2637703646237393, + "eval_sciq_pairs_loss": 0.20037926733493805, + "eval_sciq_pairs_runtime": 5.9805, + "eval_sciq_pairs_samples_per_second": 21.403, + "eval_sciq_pairs_steps_per_second": 0.334, + "step": 340 + }, + { + "epoch": 0.2637703646237393, + "eval_qasc_pairs_loss": 1.7151902914047241, + "eval_qasc_pairs_runtime": 0.9167, + "eval_qasc_pairs_samples_per_second": 139.638, + "eval_qasc_pairs_steps_per_second": 2.182, + "step": 340 + }, + { + "epoch": 0.2637703646237393, + "eval_openbookqa_pairs_loss": 2.896383285522461, + "eval_openbookqa_pairs_runtime": 0.8663, + "eval_openbookqa_pairs_samples_per_second": 147.758, + "eval_openbookqa_pairs_steps_per_second": 2.309, + "step": 340 + }, + { + "epoch": 0.2637703646237393, + "eval_msmarco_pairs_loss": 1.7021561861038208, + "eval_msmarco_pairs_runtime": 1.9693, + "eval_msmarco_pairs_samples_per_second": 64.998, + "eval_msmarco_pairs_steps_per_second": 1.016, + "step": 340 + }, + { + "epoch": 0.2637703646237393, + "eval_nq_pairs_loss": 2.386425495147705, + "eval_nq_pairs_runtime": 3.329, + "eval_nq_pairs_samples_per_second": 38.45, + "eval_nq_pairs_steps_per_second": 0.601, + "step": 340 + }, + { + "epoch": 0.2637703646237393, + "eval_trivia_pairs_loss": 2.106788158416748, + "eval_trivia_pairs_runtime": 4.4813, + "eval_trivia_pairs_samples_per_second": 28.563, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 340 + }, + { + "epoch": 0.2637703646237393, + "eval_gooaq_pairs_loss": 1.6904075145721436, + "eval_gooaq_pairs_runtime": 1.462, + "eval_gooaq_pairs_samples_per_second": 87.549, + "eval_gooaq_pairs_steps_per_second": 1.368, + "step": 340 + }, + { + "epoch": 0.2637703646237393, + "eval_paws-pos_loss": 0.07298789173364639, + "eval_paws-pos_runtime": 1.0282, + "eval_paws-pos_samples_per_second": 124.488, + "eval_paws-pos_steps_per_second": 1.945, + "step": 340 + }, + { + "epoch": 0.2637703646237393, + "eval_global_dataset_loss": 1.1202222108840942, + "eval_global_dataset_runtime": 19.7634, + "eval_global_dataset_samples_per_second": 21.049, + "eval_global_dataset_steps_per_second": 0.354, + "step": 340 + }, + { + "epoch": 0.26454615981380913, + "grad_norm": 13.917789459228516, + "learning_rate": 1.390129259694477e-05, + "loss": 1.0405, + "step": 341 + }, + { + "epoch": 0.26532195500387895, + "grad_norm": 3.5778942108154297, + "learning_rate": 1.3942420681551116e-05, + "loss": 0.1426, + "step": 342 + }, + { + "epoch": 0.2660977501939488, + "grad_norm": 2.9975743293762207, + "learning_rate": 1.398354876615746e-05, + "loss": 0.1581, + "step": 343 + }, + { + "epoch": 0.26687354538401864, + "grad_norm": 16.48363494873047, + "learning_rate": 1.4024676850763806e-05, + "loss": 1.2941, + "step": 344 + }, + { + "epoch": 0.26764934057408846, + "grad_norm": 18.412607192993164, + "learning_rate": 1.4065804935370152e-05, + "loss": 1.4073, + "step": 345 + }, + { + "epoch": 0.2684251357641583, + "grad_norm": 23.597875595092773, + "learning_rate": 1.4106933019976497e-05, + "loss": 2.9241, + "step": 346 + }, + { + "epoch": 0.2692009309542281, + "grad_norm": 14.105793952941895, + "learning_rate": 1.4148061104582842e-05, + "loss": 1.0363, + "step": 347 + }, + { + "epoch": 0.2699767261442979, + "grad_norm": 10.320302963256836, + "learning_rate": 1.4189189189189189e-05, + "loss": 0.4006, + "step": 348 + }, + { + "epoch": 0.2707525213343677, + "grad_norm": 3.639824390411377, + "learning_rate": 1.4230317273795533e-05, + "loss": 0.1023, + "step": 349 + }, + { + "epoch": 0.27152831652443754, + "grad_norm": 17.6627140045166, + "learning_rate": 1.4271445358401878e-05, + "loss": 1.387, + "step": 350 + }, + { + "epoch": 0.27230411171450736, + "grad_norm": 13.692511558532715, + "learning_rate": 1.4312573443008225e-05, + "loss": 1.1226, + "step": 351 + }, + { + "epoch": 0.2730799069045772, + "grad_norm": 12.948782920837402, + "learning_rate": 1.435370152761457e-05, + "loss": 1.0783, + "step": 352 + }, + { + "epoch": 0.273855702094647, + "grad_norm": 10.763011932373047, + "learning_rate": 1.4394829612220916e-05, + "loss": 0.5612, + "step": 353 + }, + { + "epoch": 0.2746314972847168, + "grad_norm": 9.04108715057373, + "learning_rate": 1.4435957696827261e-05, + "loss": 0.341, + "step": 354 + }, + { + "epoch": 0.2754072924747867, + "grad_norm": 19.667123794555664, + "learning_rate": 1.4477085781433606e-05, + "loss": 2.1785, + "step": 355 + }, + { + "epoch": 0.2761830876648565, + "grad_norm": 13.326544761657715, + "learning_rate": 1.4518213866039951e-05, + "loss": 1.0447, + "step": 356 + }, + { + "epoch": 0.2769588828549263, + "grad_norm": 13.251660346984863, + "learning_rate": 1.4559341950646297e-05, + "loss": 1.7571, + "step": 357 + }, + { + "epoch": 0.27773467804499613, + "grad_norm": 11.811800003051758, + "learning_rate": 1.4600470035252644e-05, + "loss": 0.7628, + "step": 358 + }, + { + "epoch": 0.27851047323506595, + "grad_norm": 12.644298553466797, + "learning_rate": 1.4641598119858989e-05, + "loss": 1.484, + "step": 359 + }, + { + "epoch": 0.27928626842513576, + "grad_norm": 13.762871742248535, + "learning_rate": 1.4682726204465333e-05, + "loss": 0.9251, + "step": 360 + }, + { + "epoch": 0.27928626842513576, + "eval_Qnli-dev_cosine_accuracy": 0.666015625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8717567324638367, + "eval_Qnli-dev_cosine_ap": 0.6860759366130562, + "eval_Qnli-dev_cosine_f1": 0.6689536878216124, + "eval_Qnli-dev_cosine_f1_threshold": 0.7581139206886292, + "eval_Qnli-dev_cosine_precision": 0.5619596541786743, + "eval_Qnli-dev_cosine_recall": 0.826271186440678, + "eval_Qnli-dev_dot_accuracy": 0.63671875, + "eval_Qnli-dev_dot_accuracy_threshold": 459.53399658203125, + "eval_Qnli-dev_dot_ap": 0.6218381128546775, + "eval_Qnli-dev_dot_f1": 0.6439716312056738, + "eval_Qnli-dev_dot_f1_threshold": 301.81005859375, + "eval_Qnli-dev_dot_precision": 0.4840085287846482, + "eval_Qnli-dev_dot_recall": 0.961864406779661, + "eval_Qnli-dev_euclidean_accuracy": 0.669921875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.433991432189941, + "eval_Qnli-dev_euclidean_ap": 0.6869080605731943, + "eval_Qnli-dev_euclidean_f1": 0.6656101426307447, + "eval_Qnli-dev_euclidean_f1_threshold": 17.318620681762695, + "eval_Qnli-dev_euclidean_precision": 0.5316455696202531, + "eval_Qnli-dev_euclidean_recall": 0.8898305084745762, + "eval_Qnli-dev_manhattan_accuracy": 0.6640625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 228.75628662109375, + "eval_Qnli-dev_manhattan_ap": 0.68965585683084, + "eval_Qnli-dev_manhattan_f1": 0.6632478632478632, + "eval_Qnli-dev_manhattan_f1_threshold": 284.19061279296875, + "eval_Qnli-dev_manhattan_precision": 0.5558739255014327, + "eval_Qnli-dev_manhattan_recall": 0.8220338983050848, + "eval_Qnli-dev_max_accuracy": 0.669921875, + "eval_Qnli-dev_max_accuracy_threshold": 459.53399658203125, + "eval_Qnli-dev_max_ap": 0.68965585683084, + "eval_Qnli-dev_max_f1": 0.6689536878216124, + "eval_Qnli-dev_max_f1_threshold": 301.81005859375, + "eval_Qnli-dev_max_precision": 0.5619596541786743, + "eval_Qnli-dev_max_recall": 0.961864406779661, + "eval_allNLI-dev_cosine_accuracy": 0.70703125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9161433577537537, + "eval_allNLI-dev_cosine_ap": 0.5573013541231226, + "eval_allNLI-dev_cosine_f1": 0.5746102449888643, + "eval_allNLI-dev_cosine_f1_threshold": 0.8298524618148804, + "eval_allNLI-dev_cosine_precision": 0.4673913043478261, + "eval_allNLI-dev_cosine_recall": 0.7456647398843931, + "eval_allNLI-dev_dot_accuracy": 0.671875, + "eval_allNLI-dev_dot_accuracy_threshold": 528.9375610351562, + "eval_allNLI-dev_dot_ap": 0.44070356109899267, + "eval_allNLI-dev_dot_f1": 0.5288461538461539, + "eval_allNLI-dev_dot_f1_threshold": 365.5830993652344, + "eval_allNLI-dev_dot_precision": 0.36585365853658536, + "eval_allNLI-dev_dot_recall": 0.953757225433526, + "eval_allNLI-dev_euclidean_accuracy": 0.70703125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 9.844797134399414, + "eval_allNLI-dev_euclidean_ap": 0.5560731306129769, + "eval_allNLI-dev_euclidean_f1": 0.5802197802197803, + "eval_allNLI-dev_euclidean_f1_threshold": 13.859457015991211, + "eval_allNLI-dev_euclidean_precision": 0.46808510638297873, + "eval_allNLI-dev_euclidean_recall": 0.7630057803468208, + "eval_allNLI-dev_manhattan_accuracy": 0.705078125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 175.87850952148438, + "eval_allNLI-dev_manhattan_ap": 0.5563049257109853, + "eval_allNLI-dev_manhattan_f1": 0.5781584582441114, + "eval_allNLI-dev_manhattan_f1_threshold": 247.3190460205078, + "eval_allNLI-dev_manhattan_precision": 0.45918367346938777, + "eval_allNLI-dev_manhattan_recall": 0.7803468208092486, + "eval_allNLI-dev_max_accuracy": 0.70703125, + "eval_allNLI-dev_max_accuracy_threshold": 528.9375610351562, + "eval_allNLI-dev_max_ap": 0.5573013541231226, + "eval_allNLI-dev_max_f1": 0.5802197802197803, + "eval_allNLI-dev_max_f1_threshold": 365.5830993652344, + "eval_allNLI-dev_max_precision": 0.46808510638297873, + "eval_allNLI-dev_max_recall": 0.953757225433526, + "eval_sequential_score": 0.68965585683084, + "eval_sts-test_pearson_cosine": 0.721042503051432, + "eval_sts-test_pearson_dot": 0.607093809147636, + "eval_sts-test_pearson_euclidean": 0.7426227812990502, + "eval_sts-test_pearson_manhattan": 0.7503986889702454, + "eval_sts-test_pearson_max": 0.7503986889702454, + "eval_sts-test_spearman_cosine": 0.7696001515061325, + "eval_sts-test_spearman_dot": 0.5927313521543144, + "eval_sts-test_spearman_euclidean": 0.7520645039836967, + "eval_sts-test_spearman_manhattan": 0.7576274437847389, + "eval_sts-test_spearman_max": 0.7696001515061325, + "eval_vitaminc-pairs_loss": 3.9781370162963867, + "eval_vitaminc-pairs_runtime": 5.512, + "eval_vitaminc-pairs_samples_per_second": 23.222, + "eval_vitaminc-pairs_steps_per_second": 0.363, + "step": 360 + }, + { + "epoch": 0.27928626842513576, + "eval_negation-triplets_loss": 2.5823302268981934, + "eval_negation-triplets_runtime": 1.0851, + "eval_negation-triplets_samples_per_second": 117.963, + "eval_negation-triplets_steps_per_second": 1.843, + "step": 360 + }, + { + "epoch": 0.27928626842513576, + "eval_scitail-pairs-pos_loss": 0.26392969489097595, + "eval_scitail-pairs-pos_runtime": 1.2797, + "eval_scitail-pairs-pos_samples_per_second": 100.027, + "eval_scitail-pairs-pos_steps_per_second": 1.563, + "step": 360 + }, + { + "epoch": 0.27928626842513576, + "eval_scitail-pairs-qa_loss": 0.1483456939458847, + "eval_scitail-pairs-qa_runtime": 0.7754, + "eval_scitail-pairs-qa_samples_per_second": 165.081, + "eval_scitail-pairs-qa_steps_per_second": 2.579, + "step": 360 + }, + { + "epoch": 0.27928626842513576, + "eval_xsum-pairs_loss": 1.1921199560165405, + "eval_xsum-pairs_runtime": 3.921, + "eval_xsum-pairs_samples_per_second": 32.645, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 360 + }, + { + "epoch": 0.27928626842513576, + "eval_sciq_pairs_loss": 0.19694603979587555, + "eval_sciq_pairs_runtime": 5.9451, + "eval_sciq_pairs_samples_per_second": 21.53, + "eval_sciq_pairs_steps_per_second": 0.336, + "step": 360 + }, + { + "epoch": 0.27928626842513576, + "eval_qasc_pairs_loss": 1.6314117908477783, + "eval_qasc_pairs_runtime": 0.915, + "eval_qasc_pairs_samples_per_second": 139.89, + "eval_qasc_pairs_steps_per_second": 2.186, + "step": 360 + }, + { + "epoch": 0.27928626842513576, + "eval_openbookqa_pairs_loss": 2.9225094318389893, + "eval_openbookqa_pairs_runtime": 0.8645, + "eval_openbookqa_pairs_samples_per_second": 148.068, + "eval_openbookqa_pairs_steps_per_second": 2.314, + "step": 360 + }, + { + "epoch": 0.27928626842513576, + "eval_msmarco_pairs_loss": 1.7400954961776733, + "eval_msmarco_pairs_runtime": 1.9712, + "eval_msmarco_pairs_samples_per_second": 64.934, + "eval_msmarco_pairs_steps_per_second": 1.015, + "step": 360 + }, + { + "epoch": 0.27928626842513576, + "eval_nq_pairs_loss": 2.5981578826904297, + "eval_nq_pairs_runtime": 3.321, + "eval_nq_pairs_samples_per_second": 38.543, + "eval_nq_pairs_steps_per_second": 0.602, + "step": 360 + }, + { + "epoch": 0.27928626842513576, + "eval_trivia_pairs_loss": 2.081812858581543, + "eval_trivia_pairs_runtime": 4.4685, + "eval_trivia_pairs_samples_per_second": 28.645, + "eval_trivia_pairs_steps_per_second": 0.448, + "step": 360 + }, + { + "epoch": 0.27928626842513576, + "eval_gooaq_pairs_loss": 1.7672526836395264, + "eval_gooaq_pairs_runtime": 1.458, + "eval_gooaq_pairs_samples_per_second": 87.79, + "eval_gooaq_pairs_steps_per_second": 1.372, + "step": 360 + }, + { + "epoch": 0.27928626842513576, + "eval_paws-pos_loss": 0.06698527932167053, + "eval_paws-pos_runtime": 1.0435, + "eval_paws-pos_samples_per_second": 122.667, + "eval_paws-pos_steps_per_second": 1.917, + "step": 360 + }, + { + "epoch": 0.27928626842513576, + "eval_global_dataset_loss": 1.085707664489746, + "eval_global_dataset_runtime": 19.4244, + "eval_global_dataset_samples_per_second": 21.416, + "eval_global_dataset_steps_per_second": 0.36, + "step": 360 + }, + { + "epoch": 0.2800620636152056, + "grad_norm": 23.30377769470215, + "learning_rate": 1.4723854289071678e-05, + "loss": 2.7937, + "step": 361 + }, + { + "epoch": 0.2808378588052754, + "grad_norm": 1.3426634073257446, + "learning_rate": 1.4764982373678025e-05, + "loss": 0.0538, + "step": 362 + }, + { + "epoch": 0.2816136539953452, + "grad_norm": 1.4219021797180176, + "learning_rate": 1.480611045828437e-05, + "loss": 0.0386, + "step": 363 + }, + { + "epoch": 0.28238944918541503, + "grad_norm": 13.968167304992676, + "learning_rate": 1.4847238542890715e-05, + "loss": 0.9797, + "step": 364 + }, + { + "epoch": 0.28316524437548485, + "grad_norm": 15.638184547424316, + "learning_rate": 1.4888366627497062e-05, + "loss": 1.1493, + "step": 365 + }, + { + "epoch": 0.2839410395655547, + "grad_norm": 22.125505447387695, + "learning_rate": 1.4929494712103406e-05, + "loss": 2.3626, + "step": 366 + }, + { + "epoch": 0.28471683475562454, + "grad_norm": 23.756380081176758, + "learning_rate": 1.4970622796709751e-05, + "loss": 2.1746, + "step": 367 + }, + { + "epoch": 0.28549262994569435, + "grad_norm": 14.218887329101562, + "learning_rate": 1.5011750881316097e-05, + "loss": 1.6042, + "step": 368 + }, + { + "epoch": 0.28626842513576417, + "grad_norm": 24.62739372253418, + "learning_rate": 1.5052878965922444e-05, + "loss": 2.3684, + "step": 369 + }, + { + "epoch": 0.287044220325834, + "grad_norm": 16.05254554748535, + "learning_rate": 1.5094007050528789e-05, + "loss": 2.2978, + "step": 370 + }, + { + "epoch": 0.2878200155159038, + "grad_norm": 21.618303298950195, + "learning_rate": 1.5135135135135134e-05, + "loss": 2.4572, + "step": 371 + }, + { + "epoch": 0.2885958107059736, + "grad_norm": 15.635889053344727, + "learning_rate": 1.5176263219741478e-05, + "loss": 1.3686, + "step": 372 + }, + { + "epoch": 0.28937160589604344, + "grad_norm": 2.211442470550537, + "learning_rate": 1.5217391304347825e-05, + "loss": 0.0768, + "step": 373 + }, + { + "epoch": 0.29014740108611325, + "grad_norm": 17.59386444091797, + "learning_rate": 1.525851938895417e-05, + "loss": 2.0374, + "step": 374 + }, + { + "epoch": 0.29092319627618307, + "grad_norm": 13.077046394348145, + "learning_rate": 1.5299647473560515e-05, + "loss": 0.8662, + "step": 375 + }, + { + "epoch": 0.2916989914662529, + "grad_norm": 17.446516036987305, + "learning_rate": 1.534077555816686e-05, + "loss": 2.145, + "step": 376 + }, + { + "epoch": 0.2924747866563227, + "grad_norm": 19.174341201782227, + "learning_rate": 1.5381903642773206e-05, + "loss": 2.2876, + "step": 377 + }, + { + "epoch": 0.2932505818463926, + "grad_norm": 17.09736442565918, + "learning_rate": 1.542303172737955e-05, + "loss": 2.3548, + "step": 378 + }, + { + "epoch": 0.2940263770364624, + "grad_norm": 11.424643516540527, + "learning_rate": 1.5464159811985896e-05, + "loss": 0.8359, + "step": 379 + }, + { + "epoch": 0.2948021722265322, + "grad_norm": 11.573511123657227, + "learning_rate": 1.5505287896592242e-05, + "loss": 1.1812, + "step": 380 + }, + { + "epoch": 0.2948021722265322, + "eval_Qnli-dev_cosine_accuracy": 0.6640625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.9027425050735474, + "eval_Qnli-dev_cosine_ap": 0.6850466461497646, + "eval_Qnli-dev_cosine_f1": 0.6601626016260163, + "eval_Qnli-dev_cosine_f1_threshold": 0.7795690298080444, + "eval_Qnli-dev_cosine_precision": 0.5356200527704486, + "eval_Qnli-dev_cosine_recall": 0.8601694915254238, + "eval_Qnli-dev_dot_accuracy": 0.65234375, + "eval_Qnli-dev_dot_accuracy_threshold": 509.9356689453125, + "eval_Qnli-dev_dot_ap": 0.6203506873347397, + "eval_Qnli-dev_dot_f1": 0.6473988439306358, + "eval_Qnli-dev_dot_f1_threshold": 384.783203125, + "eval_Qnli-dev_dot_precision": 0.49122807017543857, + "eval_Qnli-dev_dot_recall": 0.9491525423728814, + "eval_Qnli-dev_euclidean_accuracy": 0.6640625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.7774658203125, + "eval_Qnli-dev_euclidean_ap": 0.6843744819455305, + "eval_Qnli-dev_euclidean_f1": 0.6635367762128326, + "eval_Qnli-dev_euclidean_f1_threshold": 16.764877319335938, + "eval_Qnli-dev_euclidean_precision": 0.5260545905707196, + "eval_Qnli-dev_euclidean_recall": 0.8983050847457628, + "eval_Qnli-dev_manhattan_accuracy": 0.66796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 223.02120971679688, + "eval_Qnli-dev_manhattan_ap": 0.6851966405501015, + "eval_Qnli-dev_manhattan_f1": 0.665625, + "eval_Qnli-dev_manhattan_f1_threshold": 294.8950500488281, + "eval_Qnli-dev_manhattan_precision": 0.5272277227722773, + "eval_Qnli-dev_manhattan_recall": 0.902542372881356, + "eval_Qnli-dev_max_accuracy": 0.66796875, + "eval_Qnli-dev_max_accuracy_threshold": 509.9356689453125, + "eval_Qnli-dev_max_ap": 0.6851966405501015, + "eval_Qnli-dev_max_f1": 0.665625, + "eval_Qnli-dev_max_f1_threshold": 384.783203125, + "eval_Qnli-dev_max_precision": 0.5356200527704486, + "eval_Qnli-dev_max_recall": 0.9491525423728814, + "eval_allNLI-dev_cosine_accuracy": 0.705078125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9008569717407227, + "eval_allNLI-dev_cosine_ap": 0.5531787103766541, + "eval_allNLI-dev_cosine_f1": 0.5574387947269304, + "eval_allNLI-dev_cosine_f1_threshold": 0.7870105504989624, + "eval_allNLI-dev_cosine_precision": 0.4134078212290503, + "eval_allNLI-dev_cosine_recall": 0.8554913294797688, + "eval_allNLI-dev_dot_accuracy": 0.673828125, + "eval_allNLI-dev_dot_accuracy_threshold": 554.8272705078125, + "eval_allNLI-dev_dot_ap": 0.44573561475351475, + "eval_allNLI-dev_dot_f1": 0.5312499999999999, + "eval_allNLI-dev_dot_f1_threshold": 415.10894775390625, + "eval_allNLI-dev_dot_precision": 0.37965260545905705, + "eval_allNLI-dev_dot_recall": 0.884393063583815, + "eval_allNLI-dev_euclidean_accuracy": 0.70703125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.18548583984375, + "eval_allNLI-dev_euclidean_ap": 0.5500950729176494, + "eval_allNLI-dev_euclidean_f1": 0.5668016194331984, + "eval_allNLI-dev_euclidean_f1_threshold": 15.04577922821045, + "eval_allNLI-dev_euclidean_precision": 0.43613707165109034, + "eval_allNLI-dev_euclidean_recall": 0.8092485549132948, + "eval_allNLI-dev_manhattan_accuracy": 0.703125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 193.99375915527344, + "eval_allNLI-dev_manhattan_ap": 0.5517183166468855, + "eval_allNLI-dev_manhattan_f1": 0.5609756097560975, + "eval_allNLI-dev_manhattan_f1_threshold": 266.77337646484375, + "eval_allNLI-dev_manhattan_precision": 0.43260188087774293, + "eval_allNLI-dev_manhattan_recall": 0.7976878612716763, + "eval_allNLI-dev_max_accuracy": 0.70703125, + "eval_allNLI-dev_max_accuracy_threshold": 554.8272705078125, + "eval_allNLI-dev_max_ap": 0.5531787103766541, + "eval_allNLI-dev_max_f1": 0.5668016194331984, + "eval_allNLI-dev_max_f1_threshold": 415.10894775390625, + "eval_allNLI-dev_max_precision": 0.43613707165109034, + "eval_allNLI-dev_max_recall": 0.884393063583815, + "eval_sequential_score": 0.6851966405501015, + "eval_sts-test_pearson_cosine": 0.7583482295015535, + "eval_sts-test_pearson_dot": 0.6652657285284738, + "eval_sts-test_pearson_euclidean": 0.7747623802939627, + "eval_sts-test_pearson_manhattan": 0.7802942568431033, + "eval_sts-test_pearson_max": 0.7802942568431033, + "eval_sts-test_spearman_cosine": 0.7967169993412055, + "eval_sts-test_spearman_dot": 0.6482103411751988, + "eval_sts-test_spearman_euclidean": 0.7804728732225256, + "eval_sts-test_spearman_manhattan": 0.7852097494895438, + "eval_sts-test_spearman_max": 0.7967169993412055, + "eval_vitaminc-pairs_loss": 3.9984049797058105, + "eval_vitaminc-pairs_runtime": 5.4815, + "eval_vitaminc-pairs_samples_per_second": 23.351, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 380 + }, + { + "epoch": 0.2948021722265322, + "eval_negation-triplets_loss": 2.440089225769043, + "eval_negation-triplets_runtime": 1.1093, + "eval_negation-triplets_samples_per_second": 115.389, + "eval_negation-triplets_steps_per_second": 1.803, + "step": 380 + }, + { + "epoch": 0.2948021722265322, + "eval_scitail-pairs-pos_loss": 0.24433737993240356, + "eval_scitail-pairs-pos_runtime": 1.2503, + "eval_scitail-pairs-pos_samples_per_second": 102.378, + "eval_scitail-pairs-pos_steps_per_second": 1.6, + "step": 380 + }, + { + "epoch": 0.2948021722265322, + "eval_scitail-pairs-qa_loss": 0.2128182053565979, + "eval_scitail-pairs-qa_runtime": 0.7635, + "eval_scitail-pairs-qa_samples_per_second": 167.654, + "eval_scitail-pairs-qa_steps_per_second": 2.62, + "step": 380 + }, + { + "epoch": 0.2948021722265322, + "eval_xsum-pairs_loss": 1.0225056409835815, + "eval_xsum-pairs_runtime": 3.9203, + "eval_xsum-pairs_samples_per_second": 32.65, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 380 + }, + { + "epoch": 0.2948021722265322, + "eval_sciq_pairs_loss": 0.1557704657316208, + "eval_sciq_pairs_runtime": 5.8827, + "eval_sciq_pairs_samples_per_second": 21.759, + "eval_sciq_pairs_steps_per_second": 0.34, + "step": 380 + }, + { + "epoch": 0.2948021722265322, + "eval_qasc_pairs_loss": 1.5930508375167847, + "eval_qasc_pairs_runtime": 0.9026, + "eval_qasc_pairs_samples_per_second": 141.812, + "eval_qasc_pairs_steps_per_second": 2.216, + "step": 380 + }, + { + "epoch": 0.2948021722265322, + "eval_openbookqa_pairs_loss": 2.7589190006256104, + "eval_openbookqa_pairs_runtime": 0.8539, + "eval_openbookqa_pairs_samples_per_second": 149.905, + "eval_openbookqa_pairs_steps_per_second": 2.342, + "step": 380 + }, + { + "epoch": 0.2948021722265322, + "eval_msmarco_pairs_loss": 1.4480749368667603, + "eval_msmarco_pairs_runtime": 1.962, + "eval_msmarco_pairs_samples_per_second": 65.24, + "eval_msmarco_pairs_steps_per_second": 1.019, + "step": 380 + }, + { + "epoch": 0.2948021722265322, + "eval_nq_pairs_loss": 2.1773812770843506, + "eval_nq_pairs_runtime": 3.3202, + "eval_nq_pairs_samples_per_second": 38.551, + "eval_nq_pairs_steps_per_second": 0.602, + "step": 380 + }, + { + "epoch": 0.2948021722265322, + "eval_trivia_pairs_loss": 1.9234707355499268, + "eval_trivia_pairs_runtime": 4.4787, + "eval_trivia_pairs_samples_per_second": 28.58, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 380 + }, + { + "epoch": 0.2948021722265322, + "eval_gooaq_pairs_loss": 1.5105968713760376, + "eval_gooaq_pairs_runtime": 1.4668, + "eval_gooaq_pairs_samples_per_second": 87.263, + "eval_gooaq_pairs_steps_per_second": 1.363, + "step": 380 + }, + { + "epoch": 0.2948021722265322, + "eval_paws-pos_loss": 0.06424460560083389, + "eval_paws-pos_runtime": 1.0306, + "eval_paws-pos_samples_per_second": 124.201, + "eval_paws-pos_steps_per_second": 1.941, + "step": 380 + }, + { + "epoch": 0.2948021722265322, + "eval_global_dataset_loss": 0.9622049927711487, + "eval_global_dataset_runtime": 19.4149, + "eval_global_dataset_samples_per_second": 21.427, + "eval_global_dataset_steps_per_second": 0.361, + "step": 380 + }, + { + "epoch": 0.295577967416602, + "grad_norm": 10.749715805053711, + "learning_rate": 1.554641598119859e-05, + "loss": 0.7859, + "step": 381 + }, + { + "epoch": 0.29635376260667184, + "grad_norm": 11.390009880065918, + "learning_rate": 1.5587544065804936e-05, + "loss": 0.9975, + "step": 382 + }, + { + "epoch": 0.29712955779674166, + "grad_norm": 9.9592866897583, + "learning_rate": 1.5628672150411278e-05, + "loss": 0.7058, + "step": 383 + }, + { + "epoch": 0.2979053529868115, + "grad_norm": 13.89486026763916, + "learning_rate": 1.5669800235017623e-05, + "loss": 1.2329, + "step": 384 + }, + { + "epoch": 0.2986811481768813, + "grad_norm": 12.893826484680176, + "learning_rate": 1.571092831962397e-05, + "loss": 1.3389, + "step": 385 + }, + { + "epoch": 0.2994569433669511, + "grad_norm": 17.778247833251953, + "learning_rate": 1.5752056404230317e-05, + "loss": 2.3682, + "step": 386 + }, + { + "epoch": 0.30023273855702093, + "grad_norm": 15.36402416229248, + "learning_rate": 1.5793184488836662e-05, + "loss": 2.6586, + "step": 387 + }, + { + "epoch": 0.30100853374709075, + "grad_norm": 10.452804565429688, + "learning_rate": 1.5834312573443007e-05, + "loss": 0.7775, + "step": 388 + }, + { + "epoch": 0.30178432893716056, + "grad_norm": 14.66141128540039, + "learning_rate": 1.5875440658049353e-05, + "loss": 2.5431, + "step": 389 + }, + { + "epoch": 0.30256012412723043, + "grad_norm": 14.61391544342041, + "learning_rate": 1.5916568742655698e-05, + "loss": 0.7388, + "step": 390 + }, + { + "epoch": 0.30333591931730025, + "grad_norm": 19.732927322387695, + "learning_rate": 1.5957696827262043e-05, + "loss": 2.1497, + "step": 391 + }, + { + "epoch": 0.30411171450737007, + "grad_norm": 12.427021980285645, + "learning_rate": 1.599882491186839e-05, + "loss": 0.9127, + "step": 392 + }, + { + "epoch": 0.3048875096974399, + "grad_norm": 19.227365493774414, + "learning_rate": 1.6039952996474734e-05, + "loss": 2.4302, + "step": 393 + }, + { + "epoch": 0.3056633048875097, + "grad_norm": 13.65877628326416, + "learning_rate": 1.608108108108108e-05, + "loss": 1.0674, + "step": 394 + }, + { + "epoch": 0.3064391000775795, + "grad_norm": 25.27019691467285, + "learning_rate": 1.6122209165687425e-05, + "loss": 3.9795, + "step": 395 + }, + { + "epoch": 0.30721489526764933, + "grad_norm": 19.40943717956543, + "learning_rate": 1.616333725029377e-05, + "loss": 2.3552, + "step": 396 + }, + { + "epoch": 0.30799069045771915, + "grad_norm": 2.705491542816162, + "learning_rate": 1.6204465334900115e-05, + "loss": 0.0934, + "step": 397 + }, + { + "epoch": 0.30876648564778897, + "grad_norm": 11.990962028503418, + "learning_rate": 1.624559341950646e-05, + "loss": 1.7597, + "step": 398 + }, + { + "epoch": 0.3095422808378588, + "grad_norm": 11.044981956481934, + "learning_rate": 1.628672150411281e-05, + "loss": 0.6345, + "step": 399 + }, + { + "epoch": 0.3103180760279286, + "grad_norm": 16.41925811767578, + "learning_rate": 1.632784958871915e-05, + "loss": 2.0666, + "step": 400 + }, + { + "epoch": 0.3103180760279286, + "eval_Qnli-dev_cosine_accuracy": 0.673828125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8843217492103577, + "eval_Qnli-dev_cosine_ap": 0.6894082661342746, + "eval_Qnli-dev_cosine_f1": 0.6605783866057838, + "eval_Qnli-dev_cosine_f1_threshold": 0.7470111846923828, + "eval_Qnli-dev_cosine_precision": 0.5154394299287411, + "eval_Qnli-dev_cosine_recall": 0.9194915254237288, + "eval_Qnli-dev_dot_accuracy": 0.62890625, + "eval_Qnli-dev_dot_accuracy_threshold": 506.17962646484375, + "eval_Qnli-dev_dot_ap": 0.6112775051641957, + "eval_Qnli-dev_dot_f1": 0.6456043956043955, + "eval_Qnli-dev_dot_f1_threshold": 336.57537841796875, + "eval_Qnli-dev_dot_precision": 0.47764227642276424, + "eval_Qnli-dev_dot_recall": 0.9957627118644068, + "eval_Qnli-dev_euclidean_accuracy": 0.677734375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.53748893737793, + "eval_Qnli-dev_euclidean_ap": 0.6900341176815967, + "eval_Qnli-dev_euclidean_f1": 0.6621848739495799, + "eval_Qnli-dev_euclidean_f1_threshold": 15.495328903198242, + "eval_Qnli-dev_euclidean_precision": 0.5487465181058496, + "eval_Qnli-dev_euclidean_recall": 0.8347457627118644, + "eval_Qnli-dev_manhattan_accuracy": 0.669921875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 213.0938720703125, + "eval_Qnli-dev_manhattan_ap": 0.6924573404388727, + "eval_Qnli-dev_manhattan_f1": 0.6600660066006601, + "eval_Qnli-dev_manhattan_f1_threshold": 278.85546875, + "eval_Qnli-dev_manhattan_precision": 0.5405405405405406, + "eval_Qnli-dev_manhattan_recall": 0.847457627118644, + "eval_Qnli-dev_max_accuracy": 0.677734375, + "eval_Qnli-dev_max_accuracy_threshold": 506.17962646484375, + "eval_Qnli-dev_max_ap": 0.6924573404388727, + "eval_Qnli-dev_max_f1": 0.6621848739495799, + "eval_Qnli-dev_max_f1_threshold": 336.57537841796875, + "eval_Qnli-dev_max_precision": 0.5487465181058496, + "eval_Qnli-dev_max_recall": 0.9957627118644068, + "eval_allNLI-dev_cosine_accuracy": 0.708984375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.917401909828186, + "eval_allNLI-dev_cosine_ap": 0.5655492068366275, + "eval_allNLI-dev_cosine_f1": 0.583143507972665, + "eval_allNLI-dev_cosine_f1_threshold": 0.83958500623703, + "eval_allNLI-dev_cosine_precision": 0.48120300751879697, + "eval_allNLI-dev_cosine_recall": 0.7398843930635838, + "eval_allNLI-dev_dot_accuracy": 0.6796875, + "eval_allNLI-dev_dot_accuracy_threshold": 591.0133056640625, + "eval_allNLI-dev_dot_ap": 0.46233126222631415, + "eval_allNLI-dev_dot_f1": 0.5340909090909091, + "eval_allNLI-dev_dot_f1_threshold": 436.5849304199219, + "eval_allNLI-dev_dot_precision": 0.3971830985915493, + "eval_allNLI-dev_dot_recall": 0.815028901734104, + "eval_allNLI-dev_euclidean_accuracy": 0.7109375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 9.90230941772461, + "eval_allNLI-dev_euclidean_ap": 0.5614934705862846, + "eval_allNLI-dev_euclidean_f1": 0.5873015873015873, + "eval_allNLI-dev_euclidean_f1_threshold": 12.479818344116211, + "eval_allNLI-dev_euclidean_precision": 0.5414634146341464, + "eval_allNLI-dev_euclidean_recall": 0.6416184971098265, + "eval_allNLI-dev_manhattan_accuracy": 0.703125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 177.12820434570312, + "eval_allNLI-dev_manhattan_ap": 0.5630971701533405, + "eval_allNLI-dev_manhattan_f1": 0.5906040268456376, + "eval_allNLI-dev_manhattan_f1_threshold": 244.14511108398438, + "eval_allNLI-dev_manhattan_precision": 0.48175182481751827, + "eval_allNLI-dev_manhattan_recall": 0.7630057803468208, + "eval_allNLI-dev_max_accuracy": 0.7109375, + "eval_allNLI-dev_max_accuracy_threshold": 591.0133056640625, + "eval_allNLI-dev_max_ap": 0.5655492068366275, + "eval_allNLI-dev_max_f1": 0.5906040268456376, + "eval_allNLI-dev_max_f1_threshold": 436.5849304199219, + "eval_allNLI-dev_max_precision": 0.5414634146341464, + "eval_allNLI-dev_max_recall": 0.815028901734104, + "eval_sequential_score": 0.6924573404388727, + "eval_sts-test_pearson_cosine": 0.7516962916367249, + "eval_sts-test_pearson_dot": 0.6565190583762249, + "eval_sts-test_pearson_euclidean": 0.7729158137339595, + "eval_sts-test_pearson_manhattan": 0.7778342890767319, + "eval_sts-test_pearson_max": 0.7778342890767319, + "eval_sts-test_spearman_cosine": 0.7946422413605775, + "eval_sts-test_spearman_dot": 0.6443914922843637, + "eval_sts-test_spearman_euclidean": 0.779666276487207, + "eval_sts-test_spearman_manhattan": 0.7822102066922625, + "eval_sts-test_spearman_max": 0.7946422413605775, + "eval_vitaminc-pairs_loss": 3.804929494857788, + "eval_vitaminc-pairs_runtime": 5.4301, + "eval_vitaminc-pairs_samples_per_second": 23.572, + "eval_vitaminc-pairs_steps_per_second": 0.368, + "step": 400 + }, + { + "epoch": 0.3103180760279286, + "eval_negation-triplets_loss": 2.425306797027588, + "eval_negation-triplets_runtime": 1.0649, + "eval_negation-triplets_samples_per_second": 120.194, + "eval_negation-triplets_steps_per_second": 1.878, + "step": 400 + }, + { + "epoch": 0.3103180760279286, + "eval_scitail-pairs-pos_loss": 0.24340492486953735, + "eval_scitail-pairs-pos_runtime": 1.2486, + "eval_scitail-pairs-pos_samples_per_second": 102.514, + "eval_scitail-pairs-pos_steps_per_second": 1.602, + "step": 400 + }, + { + "epoch": 0.3103180760279286, + "eval_scitail-pairs-qa_loss": 0.22168239951133728, + "eval_scitail-pairs-qa_runtime": 0.7662, + "eval_scitail-pairs-qa_samples_per_second": 167.053, + "eval_scitail-pairs-qa_steps_per_second": 2.61, + "step": 400 + }, + { + "epoch": 0.3103180760279286, + "eval_xsum-pairs_loss": 1.0453300476074219, + "eval_xsum-pairs_runtime": 3.925, + "eval_xsum-pairs_samples_per_second": 32.611, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 400 + }, + { + "epoch": 0.3103180760279286, + "eval_sciq_pairs_loss": 0.15400780737400055, + "eval_sciq_pairs_runtime": 5.901, + "eval_sciq_pairs_samples_per_second": 21.691, + "eval_sciq_pairs_steps_per_second": 0.339, + "step": 400 + }, + { + "epoch": 0.3103180760279286, + "eval_qasc_pairs_loss": 1.3813235759735107, + "eval_qasc_pairs_runtime": 0.9127, + "eval_qasc_pairs_samples_per_second": 140.242, + "eval_qasc_pairs_steps_per_second": 2.191, + "step": 400 + }, + { + "epoch": 0.3103180760279286, + "eval_openbookqa_pairs_loss": 2.546370029449463, + "eval_openbookqa_pairs_runtime": 0.8645, + "eval_openbookqa_pairs_samples_per_second": 148.067, + "eval_openbookqa_pairs_steps_per_second": 2.314, + "step": 400 + }, + { + "epoch": 0.3103180760279286, + "eval_msmarco_pairs_loss": 1.4181283712387085, + "eval_msmarco_pairs_runtime": 1.9865, + "eval_msmarco_pairs_samples_per_second": 64.434, + "eval_msmarco_pairs_steps_per_second": 1.007, + "step": 400 + }, + { + "epoch": 0.3103180760279286, + "eval_nq_pairs_loss": 2.1880569458007812, + "eval_nq_pairs_runtime": 3.3615, + "eval_nq_pairs_samples_per_second": 38.078, + "eval_nq_pairs_steps_per_second": 0.595, + "step": 400 + }, + { + "epoch": 0.3103180760279286, + "eval_trivia_pairs_loss": 1.8544824123382568, + "eval_trivia_pairs_runtime": 4.4873, + "eval_trivia_pairs_samples_per_second": 28.525, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 400 + }, + { + "epoch": 0.3103180760279286, + "eval_gooaq_pairs_loss": 1.456693410873413, + "eval_gooaq_pairs_runtime": 1.4656, + "eval_gooaq_pairs_samples_per_second": 87.335, + "eval_gooaq_pairs_steps_per_second": 1.365, + "step": 400 + }, + { + "epoch": 0.3103180760279286, + "eval_paws-pos_loss": 0.06181538105010986, + "eval_paws-pos_runtime": 1.0319, + "eval_paws-pos_samples_per_second": 124.046, + "eval_paws-pos_steps_per_second": 1.938, + "step": 400 + }, + { + "epoch": 0.3103180760279286, + "eval_global_dataset_loss": 0.9175137877464294, + "eval_global_dataset_runtime": 19.4041, + "eval_global_dataset_samples_per_second": 21.439, + "eval_global_dataset_steps_per_second": 0.361, + "step": 400 + }, + { + "epoch": 0.3110938712179985, + "grad_norm": 19.21514129638672, + "learning_rate": 1.6368977673325496e-05, + "loss": 2.4715, + "step": 401 + }, + { + "epoch": 0.3118696664080683, + "grad_norm": 9.850366592407227, + "learning_rate": 1.641010575793184e-05, + "loss": 0.6239, + "step": 402 + }, + { + "epoch": 0.3126454615981381, + "grad_norm": 1.7789502143859863, + "learning_rate": 1.645123384253819e-05, + "loss": 0.1026, + "step": 403 + }, + { + "epoch": 0.3134212567882079, + "grad_norm": 10.798182487487793, + "learning_rate": 1.6492361927144535e-05, + "loss": 0.5445, + "step": 404 + }, + { + "epoch": 0.31419705197827774, + "grad_norm": 11.48928165435791, + "learning_rate": 1.653349001175088e-05, + "loss": 0.8147, + "step": 405 + }, + { + "epoch": 0.31497284716834756, + "grad_norm": 12.467869758605957, + "learning_rate": 1.6574618096357226e-05, + "loss": 0.7878, + "step": 406 + }, + { + "epoch": 0.3157486423584174, + "grad_norm": 15.528653144836426, + "learning_rate": 1.661574618096357e-05, + "loss": 1.1501, + "step": 407 + }, + { + "epoch": 0.3165244375484872, + "grad_norm": 12.029793739318848, + "learning_rate": 1.6656874265569917e-05, + "loss": 0.9747, + "step": 408 + }, + { + "epoch": 0.317300232738557, + "grad_norm": 22.320796966552734, + "learning_rate": 1.6698002350176262e-05, + "loss": 2.3754, + "step": 409 + }, + { + "epoch": 0.3180760279286268, + "grad_norm": 25.08719825744629, + "learning_rate": 1.6739130434782607e-05, + "loss": 4.0373, + "step": 410 + }, + { + "epoch": 0.31885182311869664, + "grad_norm": 13.977923393249512, + "learning_rate": 1.6780258519388953e-05, + "loss": 0.8974, + "step": 411 + }, + { + "epoch": 0.31962761830876646, + "grad_norm": 13.509927749633789, + "learning_rate": 1.6821386603995298e-05, + "loss": 1.2489, + "step": 412 + }, + { + "epoch": 0.32040341349883633, + "grad_norm": 14.722395896911621, + "learning_rate": 1.6862514688601643e-05, + "loss": 0.902, + "step": 413 + }, + { + "epoch": 0.32117920868890615, + "grad_norm": 18.918495178222656, + "learning_rate": 1.690364277320799e-05, + "loss": 1.7322, + "step": 414 + }, + { + "epoch": 0.32195500387897596, + "grad_norm": 10.59786319732666, + "learning_rate": 1.6944770857814334e-05, + "loss": 0.6132, + "step": 415 + }, + { + "epoch": 0.3227307990690458, + "grad_norm": 1.7927511930465698, + "learning_rate": 1.6985898942420682e-05, + "loss": 0.0822, + "step": 416 + }, + { + "epoch": 0.3235065942591156, + "grad_norm": 11.786699295043945, + "learning_rate": 1.7027027027027028e-05, + "loss": 0.8815, + "step": 417 + }, + { + "epoch": 0.3242823894491854, + "grad_norm": 1.6473735570907593, + "learning_rate": 1.706815511163337e-05, + "loss": 0.0377, + "step": 418 + }, + { + "epoch": 0.32505818463925523, + "grad_norm": 13.123526573181152, + "learning_rate": 1.7109283196239715e-05, + "loss": 0.9151, + "step": 419 + }, + { + "epoch": 0.32583397982932505, + "grad_norm": 19.916244506835938, + "learning_rate": 1.7150411280846064e-05, + "loss": 2.1843, + "step": 420 + }, + { + "epoch": 0.32583397982932505, + "eval_Qnli-dev_cosine_accuracy": 0.673828125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8427678346633911, + "eval_Qnli-dev_cosine_ap": 0.6999623298919331, + "eval_Qnli-dev_cosine_f1": 0.669710806697108, + "eval_Qnli-dev_cosine_f1_threshold": 0.7154334187507629, + "eval_Qnli-dev_cosine_precision": 0.5225653206650831, + "eval_Qnli-dev_cosine_recall": 0.9322033898305084, + "eval_Qnli-dev_dot_accuracy": 0.6484375, + "eval_Qnli-dev_dot_accuracy_threshold": 490.5438232421875, + "eval_Qnli-dev_dot_ap": 0.6238031701462488, + "eval_Qnli-dev_dot_f1": 0.6474820143884892, + "eval_Qnli-dev_dot_f1_threshold": 345.4635009765625, + "eval_Qnli-dev_dot_precision": 0.49019607843137253, + "eval_Qnli-dev_dot_recall": 0.9533898305084746, + "eval_Qnli-dev_euclidean_accuracy": 0.671875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.324577331542969, + "eval_Qnli-dev_euclidean_ap": 0.7038822156764853, + "eval_Qnli-dev_euclidean_f1": 0.6708463949843261, + "eval_Qnli-dev_euclidean_f1_threshold": 17.482162475585938, + "eval_Qnli-dev_euclidean_precision": 0.5323383084577115, + "eval_Qnli-dev_euclidean_recall": 0.9067796610169492, + "eval_Qnli-dev_manhattan_accuracy": 0.66796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 222.56390380859375, + "eval_Qnli-dev_manhattan_ap": 0.6978719975511867, + "eval_Qnli-dev_manhattan_f1": 0.6697674418604651, + "eval_Qnli-dev_manhattan_f1_threshold": 310.0783386230469, + "eval_Qnli-dev_manhattan_precision": 0.5281173594132029, + "eval_Qnli-dev_manhattan_recall": 0.9152542372881356, + "eval_Qnli-dev_max_accuracy": 0.673828125, + "eval_Qnli-dev_max_accuracy_threshold": 490.5438232421875, + "eval_Qnli-dev_max_ap": 0.7038822156764853, + "eval_Qnli-dev_max_f1": 0.6708463949843261, + "eval_Qnli-dev_max_f1_threshold": 345.4635009765625, + "eval_Qnli-dev_max_precision": 0.5323383084577115, + "eval_Qnli-dev_max_recall": 0.9533898305084746, + "eval_allNLI-dev_cosine_accuracy": 0.705078125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8868391513824463, + "eval_allNLI-dev_cosine_ap": 0.566806919379863, + "eval_allNLI-dev_cosine_f1": 0.582441113490364, + "eval_allNLI-dev_cosine_f1_threshold": 0.8114951252937317, + "eval_allNLI-dev_cosine_precision": 0.46258503401360546, + "eval_allNLI-dev_cosine_recall": 0.7861271676300579, + "eval_allNLI-dev_dot_accuracy": 0.685546875, + "eval_allNLI-dev_dot_accuracy_threshold": 567.1017456054688, + "eval_allNLI-dev_dot_ap": 0.47532857442812637, + "eval_allNLI-dev_dot_f1": 0.5345794392523364, + "eval_allNLI-dev_dot_f1_threshold": 416.911376953125, + "eval_allNLI-dev_dot_precision": 0.39502762430939226, + "eval_allNLI-dev_dot_recall": 0.8265895953757225, + "eval_allNLI-dev_euclidean_accuracy": 0.701171875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.022714614868164, + "eval_allNLI-dev_euclidean_ap": 0.5651807817319943, + "eval_allNLI-dev_euclidean_f1": 0.5806451612903225, + "eval_allNLI-dev_euclidean_f1_threshold": 14.004497528076172, + "eval_allNLI-dev_euclidean_precision": 0.4827586206896552, + "eval_allNLI-dev_euclidean_recall": 0.7283236994219653, + "eval_allNLI-dev_manhattan_accuracy": 0.705078125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 205.31576538085938, + "eval_allNLI-dev_manhattan_ap": 0.5657819587898676, + "eval_allNLI-dev_manhattan_f1": 0.5851318944844125, + "eval_allNLI-dev_manhattan_f1_threshold": 241.74118041992188, + "eval_allNLI-dev_manhattan_precision": 0.5, + "eval_allNLI-dev_manhattan_recall": 0.7052023121387283, + "eval_allNLI-dev_max_accuracy": 0.705078125, + "eval_allNLI-dev_max_accuracy_threshold": 567.1017456054688, + "eval_allNLI-dev_max_ap": 0.566806919379863, + "eval_allNLI-dev_max_f1": 0.5851318944844125, + "eval_allNLI-dev_max_f1_threshold": 416.911376953125, + "eval_allNLI-dev_max_precision": 0.5, + "eval_allNLI-dev_max_recall": 0.8265895953757225, + "eval_sequential_score": 0.7038822156764853, + "eval_sts-test_pearson_cosine": 0.7538194564602132, + "eval_sts-test_pearson_dot": 0.6774233062016093, + "eval_sts-test_pearson_euclidean": 0.774731843315144, + "eval_sts-test_pearson_manhattan": 0.7787277584885369, + "eval_sts-test_pearson_max": 0.7787277584885369, + "eval_sts-test_spearman_cosine": 0.7933955607207358, + "eval_sts-test_spearman_dot": 0.6663946134802414, + "eval_sts-test_spearman_euclidean": 0.777507699393993, + "eval_sts-test_spearman_manhattan": 0.7817543703437296, + "eval_sts-test_spearman_max": 0.7933955607207358, + "eval_vitaminc-pairs_loss": 4.315428733825684, + "eval_vitaminc-pairs_runtime": 5.4742, + "eval_vitaminc-pairs_samples_per_second": 23.382, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 420 + }, + { + "epoch": 0.32583397982932505, + "eval_negation-triplets_loss": 2.3480582237243652, + "eval_negation-triplets_runtime": 1.0618, + "eval_negation-triplets_samples_per_second": 120.545, + "eval_negation-triplets_steps_per_second": 1.884, + "step": 420 + }, + { + "epoch": 0.32583397982932505, + "eval_scitail-pairs-pos_loss": 0.23376089334487915, + "eval_scitail-pairs-pos_runtime": 1.2206, + "eval_scitail-pairs-pos_samples_per_second": 104.867, + "eval_scitail-pairs-pos_steps_per_second": 1.639, + "step": 420 + }, + { + "epoch": 0.32583397982932505, + "eval_scitail-pairs-qa_loss": 0.20903728902339935, + "eval_scitail-pairs-qa_runtime": 0.7636, + "eval_scitail-pairs-qa_samples_per_second": 167.621, + "eval_scitail-pairs-qa_steps_per_second": 2.619, + "step": 420 + }, + { + "epoch": 0.32583397982932505, + "eval_xsum-pairs_loss": 0.8867725133895874, + "eval_xsum-pairs_runtime": 3.9137, + "eval_xsum-pairs_samples_per_second": 32.706, + "eval_xsum-pairs_steps_per_second": 0.511, + "step": 420 + }, + { + "epoch": 0.32583397982932505, + "eval_sciq_pairs_loss": 0.1393250972032547, + "eval_sciq_pairs_runtime": 5.9293, + "eval_sciq_pairs_samples_per_second": 21.588, + "eval_sciq_pairs_steps_per_second": 0.337, + "step": 420 + }, + { + "epoch": 0.32583397982932505, + "eval_qasc_pairs_loss": 1.2475699186325073, + "eval_qasc_pairs_runtime": 0.913, + "eval_qasc_pairs_samples_per_second": 140.203, + "eval_qasc_pairs_steps_per_second": 2.191, + "step": 420 + }, + { + "epoch": 0.32583397982932505, + "eval_openbookqa_pairs_loss": 2.4479591846466064, + "eval_openbookqa_pairs_runtime": 0.8791, + "eval_openbookqa_pairs_samples_per_second": 145.61, + "eval_openbookqa_pairs_steps_per_second": 2.275, + "step": 420 + }, + { + "epoch": 0.32583397982932505, + "eval_msmarco_pairs_loss": 1.2587966918945312, + "eval_msmarco_pairs_runtime": 1.9704, + "eval_msmarco_pairs_samples_per_second": 64.96, + "eval_msmarco_pairs_steps_per_second": 1.015, + "step": 420 + }, + { + "epoch": 0.32583397982932505, + "eval_nq_pairs_loss": 2.0437731742858887, + "eval_nq_pairs_runtime": 3.3317, + "eval_nq_pairs_samples_per_second": 38.418, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 420 + }, + { + "epoch": 0.32583397982932505, + "eval_trivia_pairs_loss": 1.710050106048584, + "eval_trivia_pairs_runtime": 4.4932, + "eval_trivia_pairs_samples_per_second": 28.487, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 420 + }, + { + "epoch": 0.32583397982932505, + "eval_gooaq_pairs_loss": 1.3402475118637085, + "eval_gooaq_pairs_runtime": 1.4653, + "eval_gooaq_pairs_samples_per_second": 87.355, + "eval_gooaq_pairs_steps_per_second": 1.365, + "step": 420 + }, + { + "epoch": 0.32583397982932505, + "eval_paws-pos_loss": 0.05687706917524338, + "eval_paws-pos_runtime": 1.0334, + "eval_paws-pos_samples_per_second": 123.861, + "eval_paws-pos_steps_per_second": 1.935, + "step": 420 + }, + { + "epoch": 0.32583397982932505, + "eval_global_dataset_loss": 0.944195568561554, + "eval_global_dataset_runtime": 19.4163, + "eval_global_dataset_samples_per_second": 21.425, + "eval_global_dataset_steps_per_second": 0.361, + "step": 420 + }, + { + "epoch": 0.32660977501939487, + "grad_norm": 11.245063781738281, + "learning_rate": 1.719153936545241e-05, + "loss": 0.8176, + "step": 421 + }, + { + "epoch": 0.3273855702094647, + "grad_norm": 12.543959617614746, + "learning_rate": 1.7232667450058754e-05, + "loss": 0.8948, + "step": 422 + }, + { + "epoch": 0.3281613653995345, + "grad_norm": 10.850048065185547, + "learning_rate": 1.7273795534665096e-05, + "loss": 0.578, + "step": 423 + }, + { + "epoch": 0.32893716058960437, + "grad_norm": 5.266476631164551, + "learning_rate": 1.7314923619271445e-05, + "loss": 0.1785, + "step": 424 + }, + { + "epoch": 0.3297129557796742, + "grad_norm": 19.835979461669922, + "learning_rate": 1.735605170387779e-05, + "loss": 2.0945, + "step": 425 + }, + { + "epoch": 0.330488750969744, + "grad_norm": 16.07326889038086, + "learning_rate": 1.7397179788484135e-05, + "loss": 1.0638, + "step": 426 + }, + { + "epoch": 0.3312645461598138, + "grad_norm": 17.70912742614746, + "learning_rate": 1.743830787309048e-05, + "loss": 1.8627, + "step": 427 + }, + { + "epoch": 0.33204034134988364, + "grad_norm": 25.017555236816406, + "learning_rate": 1.7479435957696826e-05, + "loss": 2.9704, + "step": 428 + }, + { + "epoch": 0.33281613653995346, + "grad_norm": 13.582287788391113, + "learning_rate": 1.752056404230317e-05, + "loss": 0.8579, + "step": 429 + }, + { + "epoch": 0.3335919317300233, + "grad_norm": 17.605514526367188, + "learning_rate": 1.7561692126909516e-05, + "loss": 1.5726, + "step": 430 + }, + { + "epoch": 0.3343677269200931, + "grad_norm": 16.1927490234375, + "learning_rate": 1.7602820211515865e-05, + "loss": 1.5125, + "step": 431 + }, + { + "epoch": 0.3351435221101629, + "grad_norm": 10.593696594238281, + "learning_rate": 1.7643948296122207e-05, + "loss": 0.6158, + "step": 432 + }, + { + "epoch": 0.3359193173002327, + "grad_norm": 24.76272201538086, + "learning_rate": 1.7685076380728552e-05, + "loss": 4.0055, + "step": 433 + }, + { + "epoch": 0.33669511249030254, + "grad_norm": 18.812658309936523, + "learning_rate": 1.77262044653349e-05, + "loss": 2.2631, + "step": 434 + }, + { + "epoch": 0.33747090768037236, + "grad_norm": 18.673860549926758, + "learning_rate": 1.7767332549941243e-05, + "loss": 2.3096, + "step": 435 + }, + { + "epoch": 0.33824670287044223, + "grad_norm": 11.739363670349121, + "learning_rate": 1.780846063454759e-05, + "loss": 1.0672, + "step": 436 + }, + { + "epoch": 0.33902249806051205, + "grad_norm": 17.290603637695312, + "learning_rate": 1.7849588719153933e-05, + "loss": 2.2217, + "step": 437 + }, + { + "epoch": 0.33979829325058186, + "grad_norm": 15.590789794921875, + "learning_rate": 1.789071680376028e-05, + "loss": 1.5531, + "step": 438 + }, + { + "epoch": 0.3405740884406517, + "grad_norm": 7.4414167404174805, + "learning_rate": 1.7931844888366627e-05, + "loss": 0.3912, + "step": 439 + }, + { + "epoch": 0.3413498836307215, + "grad_norm": 16.73564910888672, + "learning_rate": 1.797297297297297e-05, + "loss": 2.2856, + "step": 440 + }, + { + "epoch": 0.3413498836307215, + "eval_Qnli-dev_cosine_accuracy": 0.666015625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8762513399124146, + "eval_Qnli-dev_cosine_ap": 0.6925226528996247, + "eval_Qnli-dev_cosine_f1": 0.6677265500794912, + "eval_Qnli-dev_cosine_f1_threshold": 0.7770497798919678, + "eval_Qnli-dev_cosine_precision": 0.5343511450381679, + "eval_Qnli-dev_cosine_recall": 0.8898305084745762, + "eval_Qnli-dev_dot_accuracy": 0.609375, + "eval_Qnli-dev_dot_accuracy_threshold": 556.5031127929688, + "eval_Qnli-dev_dot_ap": 0.5850847365923477, + "eval_Qnli-dev_dot_f1": 0.6505681818181819, + "eval_Qnli-dev_dot_f1_threshold": 406.6336975097656, + "eval_Qnli-dev_dot_precision": 0.4893162393162393, + "eval_Qnli-dev_dot_recall": 0.9703389830508474, + "eval_Qnli-dev_euclidean_accuracy": 0.66796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.730552673339844, + "eval_Qnli-dev_euclidean_ap": 0.6944172611785755, + "eval_Qnli-dev_euclidean_f1": 0.6646248085758041, + "eval_Qnli-dev_euclidean_f1_threshold": 17.14211654663086, + "eval_Qnli-dev_euclidean_precision": 0.5203836930455635, + "eval_Qnli-dev_euclidean_recall": 0.9194915254237288, + "eval_Qnli-dev_manhattan_accuracy": 0.669921875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 214.77349853515625, + "eval_Qnli-dev_manhattan_ap": 0.6931032398015755, + "eval_Qnli-dev_manhattan_f1": 0.6720257234726689, + "eval_Qnli-dev_manhattan_f1_threshold": 289.2326354980469, + "eval_Qnli-dev_manhattan_precision": 0.5414507772020726, + "eval_Qnli-dev_manhattan_recall": 0.885593220338983, + "eval_Qnli-dev_max_accuracy": 0.669921875, + "eval_Qnli-dev_max_accuracy_threshold": 556.5031127929688, + "eval_Qnli-dev_max_ap": 0.6944172611785755, + "eval_Qnli-dev_max_f1": 0.6720257234726689, + "eval_Qnli-dev_max_f1_threshold": 406.6336975097656, + "eval_Qnli-dev_max_precision": 0.5414507772020726, + "eval_Qnli-dev_max_recall": 0.9703389830508474, + "eval_allNLI-dev_cosine_accuracy": 0.705078125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9292815923690796, + "eval_allNLI-dev_cosine_ap": 0.5708892349016189, + "eval_allNLI-dev_cosine_f1": 0.5858123569794049, + "eval_allNLI-dev_cosine_f1_threshold": 0.8687143921852112, + "eval_allNLI-dev_cosine_precision": 0.48484848484848486, + "eval_allNLI-dev_cosine_recall": 0.7398843930635838, + "eval_allNLI-dev_dot_accuracy": 0.673828125, + "eval_allNLI-dev_dot_accuracy_threshold": 573.7271118164062, + "eval_allNLI-dev_dot_ap": 0.4435855844051997, + "eval_allNLI-dev_dot_f1": 0.5302013422818791, + "eval_allNLI-dev_dot_f1_threshold": 430.77691650390625, + "eval_allNLI-dev_dot_precision": 0.3735224586288416, + "eval_allNLI-dev_dot_recall": 0.9132947976878613, + "eval_allNLI-dev_euclidean_accuracy": 0.703125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.651548385620117, + "eval_allNLI-dev_euclidean_ap": 0.5642096320775464, + "eval_allNLI-dev_euclidean_f1": 0.5790754257907543, + "eval_allNLI-dev_euclidean_f1_threshold": 11.96162223815918, + "eval_allNLI-dev_euclidean_precision": 0.5, + "eval_allNLI-dev_euclidean_recall": 0.6878612716763006, + "eval_allNLI-dev_manhattan_accuracy": 0.703125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 180.98876953125, + "eval_allNLI-dev_manhattan_ap": 0.5649666695504983, + "eval_allNLI-dev_manhattan_f1": 0.5995085995085995, + "eval_allNLI-dev_manhattan_f1_threshold": 210.06790161132812, + "eval_allNLI-dev_manhattan_precision": 0.5213675213675214, + "eval_allNLI-dev_manhattan_recall": 0.7052023121387283, + "eval_allNLI-dev_max_accuracy": 0.705078125, + "eval_allNLI-dev_max_accuracy_threshold": 573.7271118164062, + "eval_allNLI-dev_max_ap": 0.5708892349016189, + "eval_allNLI-dev_max_f1": 0.5995085995085995, + "eval_allNLI-dev_max_f1_threshold": 430.77691650390625, + "eval_allNLI-dev_max_precision": 0.5213675213675214, + "eval_allNLI-dev_max_recall": 0.9132947976878613, + "eval_sequential_score": 0.6944172611785755, + "eval_sts-test_pearson_cosine": 0.7522625123254352, + "eval_sts-test_pearson_dot": 0.6403081572846376, + "eval_sts-test_pearson_euclidean": 0.7739743366360256, + "eval_sts-test_pearson_manhattan": 0.7761899710973759, + "eval_sts-test_pearson_max": 0.7761899710973759, + "eval_sts-test_spearman_cosine": 0.7913403590413358, + "eval_sts-test_spearman_dot": 0.6275253683836046, + "eval_sts-test_spearman_euclidean": 0.7763585317527231, + "eval_sts-test_spearman_manhattan": 0.7771581025400033, + "eval_sts-test_spearman_max": 0.7913403590413358, + "eval_vitaminc-pairs_loss": 3.9062092304229736, + "eval_vitaminc-pairs_runtime": 5.4684, + "eval_vitaminc-pairs_samples_per_second": 23.407, + "eval_vitaminc-pairs_steps_per_second": 0.366, + "step": 440 + }, + { + "epoch": 0.3413498836307215, + "eval_negation-triplets_loss": 2.359525680541992, + "eval_negation-triplets_runtime": 1.0733, + "eval_negation-triplets_samples_per_second": 119.255, + "eval_negation-triplets_steps_per_second": 1.863, + "step": 440 + }, + { + "epoch": 0.3413498836307215, + "eval_scitail-pairs-pos_loss": 0.2214130312204361, + "eval_scitail-pairs-pos_runtime": 1.2084, + "eval_scitail-pairs-pos_samples_per_second": 105.927, + "eval_scitail-pairs-pos_steps_per_second": 1.655, + "step": 440 + }, + { + "epoch": 0.3413498836307215, + "eval_scitail-pairs-qa_loss": 0.19177046418190002, + "eval_scitail-pairs-qa_runtime": 0.7703, + "eval_scitail-pairs-qa_samples_per_second": 166.165, + "eval_scitail-pairs-qa_steps_per_second": 2.596, + "step": 440 + }, + { + "epoch": 0.3413498836307215, + "eval_xsum-pairs_loss": 0.8537141680717468, + "eval_xsum-pairs_runtime": 3.9361, + "eval_xsum-pairs_samples_per_second": 32.519, + "eval_xsum-pairs_steps_per_second": 0.508, + "step": 440 + }, + { + "epoch": 0.3413498836307215, + "eval_sciq_pairs_loss": 0.12183572351932526, + "eval_sciq_pairs_runtime": 5.8953, + "eval_sciq_pairs_samples_per_second": 21.712, + "eval_sciq_pairs_steps_per_second": 0.339, + "step": 440 + }, + { + "epoch": 0.3413498836307215, + "eval_qasc_pairs_loss": 1.2399531602859497, + "eval_qasc_pairs_runtime": 0.9066, + "eval_qasc_pairs_samples_per_second": 141.181, + "eval_qasc_pairs_steps_per_second": 2.206, + "step": 440 + }, + { + "epoch": 0.3413498836307215, + "eval_openbookqa_pairs_loss": 2.3348124027252197, + "eval_openbookqa_pairs_runtime": 0.8824, + "eval_openbookqa_pairs_samples_per_second": 145.065, + "eval_openbookqa_pairs_steps_per_second": 2.267, + "step": 440 + }, + { + "epoch": 0.3413498836307215, + "eval_msmarco_pairs_loss": 1.333595633506775, + "eval_msmarco_pairs_runtime": 1.9748, + "eval_msmarco_pairs_samples_per_second": 64.817, + "eval_msmarco_pairs_steps_per_second": 1.013, + "step": 440 + }, + { + "epoch": 0.3413498836307215, + "eval_nq_pairs_loss": 1.7639641761779785, + "eval_nq_pairs_runtime": 3.3313, + "eval_nq_pairs_samples_per_second": 38.423, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 440 + }, + { + "epoch": 0.3413498836307215, + "eval_trivia_pairs_loss": 1.4022107124328613, + "eval_trivia_pairs_runtime": 4.4857, + "eval_trivia_pairs_samples_per_second": 28.535, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 440 + }, + { + "epoch": 0.3413498836307215, + "eval_gooaq_pairs_loss": 1.3557066917419434, + "eval_gooaq_pairs_runtime": 1.4925, + "eval_gooaq_pairs_samples_per_second": 85.759, + "eval_gooaq_pairs_steps_per_second": 1.34, + "step": 440 + }, + { + "epoch": 0.3413498836307215, + "eval_paws-pos_loss": 0.07117103785276413, + "eval_paws-pos_runtime": 1.0283, + "eval_paws-pos_samples_per_second": 124.474, + "eval_paws-pos_steps_per_second": 1.945, + "step": 440 + }, + { + "epoch": 0.3413498836307215, + "eval_global_dataset_loss": 0.7811084985733032, + "eval_global_dataset_runtime": 19.4164, + "eval_global_dataset_samples_per_second": 21.425, + "eval_global_dataset_steps_per_second": 0.361, + "step": 440 + }, + { + "epoch": 0.3421256788207913, + "grad_norm": 9.4876708984375, + "learning_rate": 1.8014101057579318e-05, + "loss": 0.7398, + "step": 441 + }, + { + "epoch": 0.34290147401086113, + "grad_norm": 7.871079444885254, + "learning_rate": 1.8055229142185663e-05, + "loss": 0.4766, + "step": 442 + }, + { + "epoch": 0.34367726920093095, + "grad_norm": 11.031767845153809, + "learning_rate": 1.8096357226792005e-05, + "loss": 1.6592, + "step": 443 + }, + { + "epoch": 0.34445306439100076, + "grad_norm": 7.088497638702393, + "learning_rate": 1.8137485311398354e-05, + "loss": 0.3498, + "step": 444 + }, + { + "epoch": 0.3452288595810706, + "grad_norm": 15.35033893585205, + "learning_rate": 1.81786133960047e-05, + "loss": 1.818, + "step": 445 + }, + { + "epoch": 0.3460046547711404, + "grad_norm": 18.451932907104492, + "learning_rate": 1.8219741480611044e-05, + "loss": 2.332, + "step": 446 + }, + { + "epoch": 0.3467804499612102, + "grad_norm": 8.21351146697998, + "learning_rate": 1.826086956521739e-05, + "loss": 0.6782, + "step": 447 + }, + { + "epoch": 0.3475562451512801, + "grad_norm": 15.57119369506836, + "learning_rate": 1.830199764982374e-05, + "loss": 1.529, + "step": 448 + }, + { + "epoch": 0.3483320403413499, + "grad_norm": 14.127723693847656, + "learning_rate": 1.834312573443008e-05, + "loss": 1.3216, + "step": 449 + }, + { + "epoch": 0.3491078355314197, + "grad_norm": 2.077221155166626, + "learning_rate": 1.8384253819036426e-05, + "loss": 0.138, + "step": 450 + }, + { + "epoch": 0.34988363072148954, + "grad_norm": 11.093355178833008, + "learning_rate": 1.842538190364277e-05, + "loss": 1.1062, + "step": 451 + }, + { + "epoch": 0.35065942591155935, + "grad_norm": 17.88471794128418, + "learning_rate": 1.8466509988249116e-05, + "loss": 1.6739, + "step": 452 + }, + { + "epoch": 0.35143522110162917, + "grad_norm": 12.845499038696289, + "learning_rate": 1.8507638072855465e-05, + "loss": 0.7182, + "step": 453 + }, + { + "epoch": 0.352211016291699, + "grad_norm": 20.143211364746094, + "learning_rate": 1.8548766157461807e-05, + "loss": 2.3198, + "step": 454 + }, + { + "epoch": 0.3529868114817688, + "grad_norm": 20.101293563842773, + "learning_rate": 1.8589894242068152e-05, + "loss": 2.1016, + "step": 455 + }, + { + "epoch": 0.3537626066718386, + "grad_norm": 13.362924575805664, + "learning_rate": 1.86310223266745e-05, + "loss": 0.8252, + "step": 456 + }, + { + "epoch": 0.35453840186190844, + "grad_norm": 12.254059791564941, + "learning_rate": 1.8672150411280843e-05, + "loss": 0.7597, + "step": 457 + }, + { + "epoch": 0.35531419705197825, + "grad_norm": 11.719807624816895, + "learning_rate": 1.871327849588719e-05, + "loss": 0.6912, + "step": 458 + }, + { + "epoch": 0.3560899922420481, + "grad_norm": 18.092191696166992, + "learning_rate": 1.8754406580493537e-05, + "loss": 1.8851, + "step": 459 + }, + { + "epoch": 0.35686578743211794, + "grad_norm": 13.485638618469238, + "learning_rate": 1.879553466509988e-05, + "loss": 1.4324, + "step": 460 + }, + { + "epoch": 0.35686578743211794, + "eval_Qnli-dev_cosine_accuracy": 0.666015625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8486638069152832, + "eval_Qnli-dev_cosine_ap": 0.6779853210254181, + "eval_Qnli-dev_cosine_f1": 0.6555386949924127, + "eval_Qnli-dev_cosine_f1_threshold": 0.7324475049972534, + "eval_Qnli-dev_cosine_precision": 0.5106382978723404, + "eval_Qnli-dev_cosine_recall": 0.9152542372881356, + "eval_Qnli-dev_dot_accuracy": 0.638671875, + "eval_Qnli-dev_dot_accuracy_threshold": 477.6387939453125, + "eval_Qnli-dev_dot_ap": 0.6137232223987072, + "eval_Qnli-dev_dot_f1": 0.6473029045643155, + "eval_Qnli-dev_dot_f1_threshold": 316.2604675292969, + "eval_Qnli-dev_dot_precision": 0.4804928131416838, + "eval_Qnli-dev_dot_recall": 0.9915254237288136, + "eval_Qnli-dev_euclidean_accuracy": 0.658203125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.233789443969727, + "eval_Qnli-dev_euclidean_ap": 0.6752873255687661, + "eval_Qnli-dev_euclidean_f1": 0.665474060822898, + "eval_Qnli-dev_euclidean_f1_threshold": 14.972784042358398, + "eval_Qnli-dev_euclidean_precision": 0.5758513931888545, + "eval_Qnli-dev_euclidean_recall": 0.788135593220339, + "eval_Qnli-dev_manhattan_accuracy": 0.654296875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 210.17254638671875, + "eval_Qnli-dev_manhattan_ap": 0.6766703462138357, + "eval_Qnli-dev_manhattan_f1": 0.660436137071651, + "eval_Qnli-dev_manhattan_f1_threshold": 301.0316467285156, + "eval_Qnli-dev_manhattan_precision": 0.5221674876847291, + "eval_Qnli-dev_manhattan_recall": 0.8983050847457628, + "eval_Qnli-dev_max_accuracy": 0.666015625, + "eval_Qnli-dev_max_accuracy_threshold": 477.6387939453125, + "eval_Qnli-dev_max_ap": 0.6779853210254181, + "eval_Qnli-dev_max_f1": 0.665474060822898, + "eval_Qnli-dev_max_f1_threshold": 316.2604675292969, + "eval_Qnli-dev_max_precision": 0.5758513931888545, + "eval_Qnli-dev_max_recall": 0.9915254237288136, + "eval_allNLI-dev_cosine_accuracy": 0.701171875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9253606200218201, + "eval_allNLI-dev_cosine_ap": 0.5573276509338424, + "eval_allNLI-dev_cosine_f1": 0.5792079207920792, + "eval_allNLI-dev_cosine_f1_threshold": 0.8566697835922241, + "eval_allNLI-dev_cosine_precision": 0.5064935064935064, + "eval_allNLI-dev_cosine_recall": 0.6763005780346821, + "eval_allNLI-dev_dot_accuracy": 0.6796875, + "eval_allNLI-dev_dot_accuracy_threshold": 527.159423828125, + "eval_allNLI-dev_dot_ap": 0.46690480039244775, + "eval_allNLI-dev_dot_f1": 0.5349233390119251, + "eval_allNLI-dev_dot_f1_threshold": 394.58905029296875, + "eval_allNLI-dev_dot_precision": 0.37922705314009664, + "eval_allNLI-dev_dot_recall": 0.9075144508670521, + "eval_allNLI-dev_euclidean_accuracy": 0.701171875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 8.716781616210938, + "eval_allNLI-dev_euclidean_ap": 0.5557980725767, + "eval_allNLI-dev_euclidean_f1": 0.5794392523364487, + "eval_allNLI-dev_euclidean_f1_threshold": 12.963998794555664, + "eval_allNLI-dev_euclidean_precision": 0.48627450980392156, + "eval_allNLI-dev_euclidean_recall": 0.7167630057803468, + "eval_allNLI-dev_manhattan_accuracy": 0.697265625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 161.34320068359375, + "eval_allNLI-dev_manhattan_ap": 0.5567111390562576, + "eval_allNLI-dev_manhattan_f1": 0.5927710843373494, + "eval_allNLI-dev_manhattan_f1_threshold": 225.5025634765625, + "eval_allNLI-dev_manhattan_precision": 0.5082644628099173, + "eval_allNLI-dev_manhattan_recall": 0.7109826589595376, + "eval_allNLI-dev_max_accuracy": 0.701171875, + "eval_allNLI-dev_max_accuracy_threshold": 527.159423828125, + "eval_allNLI-dev_max_ap": 0.5573276509338424, + "eval_allNLI-dev_max_f1": 0.5927710843373494, + "eval_allNLI-dev_max_f1_threshold": 394.58905029296875, + "eval_allNLI-dev_max_precision": 0.5082644628099173, + "eval_allNLI-dev_max_recall": 0.9075144508670521, + "eval_sequential_score": 0.6779853210254181, + "eval_sts-test_pearson_cosine": 0.7677608150695443, + "eval_sts-test_pearson_dot": 0.6657705142732455, + "eval_sts-test_pearson_euclidean": 0.7944178306135051, + "eval_sts-test_pearson_manhattan": 0.7955685051760161, + "eval_sts-test_pearson_max": 0.7955685051760161, + "eval_sts-test_spearman_cosine": 0.80392642755136, + "eval_sts-test_spearman_dot": 0.6505001737180739, + "eval_sts-test_spearman_euclidean": 0.7943252234248809, + "eval_sts-test_spearman_manhattan": 0.7942380927189517, + "eval_sts-test_spearman_max": 0.80392642755136, + "eval_vitaminc-pairs_loss": 4.100461483001709, + "eval_vitaminc-pairs_runtime": 5.4477, + "eval_vitaminc-pairs_samples_per_second": 23.496, + "eval_vitaminc-pairs_steps_per_second": 0.367, + "step": 460 + }, + { + "epoch": 0.35686578743211794, + "eval_negation-triplets_loss": 2.2808547019958496, + "eval_negation-triplets_runtime": 1.0606, + "eval_negation-triplets_samples_per_second": 120.687, + "eval_negation-triplets_steps_per_second": 1.886, + "step": 460 + }, + { + "epoch": 0.35686578743211794, + "eval_scitail-pairs-pos_loss": 0.20019960403442383, + "eval_scitail-pairs-pos_runtime": 1.2152, + "eval_scitail-pairs-pos_samples_per_second": 105.334, + "eval_scitail-pairs-pos_steps_per_second": 1.646, + "step": 460 + }, + { + "epoch": 0.35686578743211794, + "eval_scitail-pairs-qa_loss": 0.14794358611106873, + "eval_scitail-pairs-qa_runtime": 0.7618, + "eval_scitail-pairs-qa_samples_per_second": 168.017, + "eval_scitail-pairs-qa_steps_per_second": 2.625, + "step": 460 + }, + { + "epoch": 0.35686578743211794, + "eval_xsum-pairs_loss": 0.7287623882293701, + "eval_xsum-pairs_runtime": 3.9068, + "eval_xsum-pairs_samples_per_second": 32.764, + "eval_xsum-pairs_steps_per_second": 0.512, + "step": 460 + }, + { + "epoch": 0.35686578743211794, + "eval_sciq_pairs_loss": 0.1277933418750763, + "eval_sciq_pairs_runtime": 5.8889, + "eval_sciq_pairs_samples_per_second": 21.736, + "eval_sciq_pairs_steps_per_second": 0.34, + "step": 460 + }, + { + "epoch": 0.35686578743211794, + "eval_qasc_pairs_loss": 1.146864652633667, + "eval_qasc_pairs_runtime": 0.9111, + "eval_qasc_pairs_samples_per_second": 140.497, + "eval_qasc_pairs_steps_per_second": 2.195, + "step": 460 + }, + { + "epoch": 0.35686578743211794, + "eval_openbookqa_pairs_loss": 2.401211977005005, + "eval_openbookqa_pairs_runtime": 0.8654, + "eval_openbookqa_pairs_samples_per_second": 147.906, + "eval_openbookqa_pairs_steps_per_second": 2.311, + "step": 460 + }, + { + "epoch": 0.35686578743211794, + "eval_msmarco_pairs_loss": 1.1840412616729736, + "eval_msmarco_pairs_runtime": 1.9754, + "eval_msmarco_pairs_samples_per_second": 64.797, + "eval_msmarco_pairs_steps_per_second": 1.012, + "step": 460 + }, + { + "epoch": 0.35686578743211794, + "eval_nq_pairs_loss": 1.8690099716186523, + "eval_nq_pairs_runtime": 3.3352, + "eval_nq_pairs_samples_per_second": 38.379, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 460 + }, + { + "epoch": 0.35686578743211794, + "eval_trivia_pairs_loss": 1.388373851776123, + "eval_trivia_pairs_runtime": 4.4947, + "eval_trivia_pairs_samples_per_second": 28.478, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 460 + }, + { + "epoch": 0.35686578743211794, + "eval_gooaq_pairs_loss": 1.2368732690811157, + "eval_gooaq_pairs_runtime": 1.4696, + "eval_gooaq_pairs_samples_per_second": 87.099, + "eval_gooaq_pairs_steps_per_second": 1.361, + "step": 460 + }, + { + "epoch": 0.35686578743211794, + "eval_paws-pos_loss": 0.06138503551483154, + "eval_paws-pos_runtime": 1.0358, + "eval_paws-pos_samples_per_second": 123.574, + "eval_paws-pos_steps_per_second": 1.931, + "step": 460 + }, + { + "epoch": 0.35686578743211794, + "eval_global_dataset_loss": 0.8328232169151306, + "eval_global_dataset_runtime": 19.402, + "eval_global_dataset_samples_per_second": 21.441, + "eval_global_dataset_steps_per_second": 0.361, + "step": 460 + }, + { + "epoch": 0.35764158262218776, + "grad_norm": 19.271942138671875, + "learning_rate": 1.8836662749706227e-05, + "loss": 3.0981, + "step": 461 + }, + { + "epoch": 0.3584173778122576, + "grad_norm": 16.643905639648438, + "learning_rate": 1.887779083431257e-05, + "loss": 1.5446, + "step": 462 + }, + { + "epoch": 0.3591931730023274, + "grad_norm": 9.805319786071777, + "learning_rate": 1.8918918918918918e-05, + "loss": 0.7448, + "step": 463 + }, + { + "epoch": 0.3599689681923972, + "grad_norm": 12.483269691467285, + "learning_rate": 1.8960047003525263e-05, + "loss": 0.8903, + "step": 464 + }, + { + "epoch": 0.360744763382467, + "grad_norm": 10.567606925964355, + "learning_rate": 1.9001175088131608e-05, + "loss": 0.6648, + "step": 465 + }, + { + "epoch": 0.36152055857253684, + "grad_norm": 12.790678024291992, + "learning_rate": 1.9042303172737954e-05, + "loss": 1.0112, + "step": 466 + }, + { + "epoch": 0.36229635376260666, + "grad_norm": 9.75352954864502, + "learning_rate": 1.90834312573443e-05, + "loss": 0.7606, + "step": 467 + }, + { + "epoch": 0.3630721489526765, + "grad_norm": 9.474419593811035, + "learning_rate": 1.9124559341950644e-05, + "loss": 0.5876, + "step": 468 + }, + { + "epoch": 0.3638479441427463, + "grad_norm": 15.47821044921875, + "learning_rate": 1.916568742655699e-05, + "loss": 1.5486, + "step": 469 + }, + { + "epoch": 0.3646237393328161, + "grad_norm": 9.594441413879395, + "learning_rate": 1.9206815511163338e-05, + "loss": 0.7844, + "step": 470 + }, + { + "epoch": 0.365399534522886, + "grad_norm": 10.21621036529541, + "learning_rate": 1.924794359576968e-05, + "loss": 1.1707, + "step": 471 + }, + { + "epoch": 0.3661753297129558, + "grad_norm": 10.79291820526123, + "learning_rate": 1.9289071680376025e-05, + "loss": 0.7245, + "step": 472 + }, + { + "epoch": 0.3669511249030256, + "grad_norm": 11.436175346374512, + "learning_rate": 1.9330199764982374e-05, + "loss": 0.7898, + "step": 473 + }, + { + "epoch": 0.36772692009309543, + "grad_norm": 10.453071594238281, + "learning_rate": 1.9371327849588716e-05, + "loss": 0.6014, + "step": 474 + }, + { + "epoch": 0.36850271528316525, + "grad_norm": 11.758874893188477, + "learning_rate": 1.9412455934195065e-05, + "loss": 0.635, + "step": 475 + }, + { + "epoch": 0.36927851047323507, + "grad_norm": 1.1648308038711548, + "learning_rate": 1.9453584018801406e-05, + "loss": 0.0885, + "step": 476 + }, + { + "epoch": 0.3700543056633049, + "grad_norm": 7.729386806488037, + "learning_rate": 1.9494712103407755e-05, + "loss": 0.3688, + "step": 477 + }, + { + "epoch": 0.3708301008533747, + "grad_norm": 14.267598152160645, + "learning_rate": 1.95358401880141e-05, + "loss": 1.6557, + "step": 478 + }, + { + "epoch": 0.3716058960434445, + "grad_norm": 11.50204086303711, + "learning_rate": 1.9576968272620442e-05, + "loss": 0.7524, + "step": 479 + }, + { + "epoch": 0.37238169123351433, + "grad_norm": 22.12100601196289, + "learning_rate": 1.961809635722679e-05, + "loss": 1.9615, + "step": 480 + }, + { + "epoch": 0.37238169123351433, + "eval_Qnli-dev_cosine_accuracy": 0.669921875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8463388085365295, + "eval_Qnli-dev_cosine_ap": 0.6828959513877161, + "eval_Qnli-dev_cosine_f1": 0.6677067082683307, + "eval_Qnli-dev_cosine_f1_threshold": 0.7090969085693359, + "eval_Qnli-dev_cosine_precision": 0.528395061728395, + "eval_Qnli-dev_cosine_recall": 0.9067796610169492, + "eval_Qnli-dev_dot_accuracy": 0.623046875, + "eval_Qnli-dev_dot_accuracy_threshold": 450.0521545410156, + "eval_Qnli-dev_dot_ap": 0.5920139709616099, + "eval_Qnli-dev_dot_f1": 0.6536144578313253, + "eval_Qnli-dev_dot_f1_threshold": 354.99444580078125, + "eval_Qnli-dev_dot_precision": 0.5070093457943925, + "eval_Qnli-dev_dot_recall": 0.9194915254237288, + "eval_Qnli-dev_euclidean_accuracy": 0.662109375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.735542297363281, + "eval_Qnli-dev_euclidean_ap": 0.6875483049796004, + "eval_Qnli-dev_euclidean_f1": 0.6706766917293234, + "eval_Qnli-dev_euclidean_f1_threshold": 18.239444732666016, + "eval_Qnli-dev_euclidean_precision": 0.5198135198135199, + "eval_Qnli-dev_euclidean_recall": 0.9449152542372882, + "eval_Qnli-dev_manhattan_accuracy": 0.658203125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 228.6551513671875, + "eval_Qnli-dev_manhattan_ap": 0.6860476994053801, + "eval_Qnli-dev_manhattan_f1": 0.6717557251908397, + "eval_Qnli-dev_manhattan_f1_threshold": 314.2021179199219, + "eval_Qnli-dev_manhattan_precision": 0.5250596658711217, + "eval_Qnli-dev_manhattan_recall": 0.9322033898305084, + "eval_Qnli-dev_max_accuracy": 0.669921875, + "eval_Qnli-dev_max_accuracy_threshold": 450.0521545410156, + "eval_Qnli-dev_max_ap": 0.6875483049796004, + "eval_Qnli-dev_max_f1": 0.6717557251908397, + "eval_Qnli-dev_max_f1_threshold": 354.99444580078125, + "eval_Qnli-dev_max_precision": 0.528395061728395, + "eval_Qnli-dev_max_recall": 0.9449152542372882, + "eval_allNLI-dev_cosine_accuracy": 0.712890625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9102362394332886, + "eval_allNLI-dev_cosine_ap": 0.5745210101303777, + "eval_allNLI-dev_cosine_f1": 0.5879732739420935, + "eval_allNLI-dev_cosine_f1_threshold": 0.8427931070327759, + "eval_allNLI-dev_cosine_precision": 0.4782608695652174, + "eval_allNLI-dev_cosine_recall": 0.7630057803468208, + "eval_allNLI-dev_dot_accuracy": 0.681640625, + "eval_allNLI-dev_dot_accuracy_threshold": 548.634521484375, + "eval_allNLI-dev_dot_ap": 0.46684131701078263, + "eval_allNLI-dev_dot_f1": 0.5319148936170213, + "eval_allNLI-dev_dot_f1_threshold": 427.22930908203125, + "eval_allNLI-dev_dot_precision": 0.4208754208754209, + "eval_allNLI-dev_dot_recall": 0.7225433526011561, + "eval_allNLI-dev_euclidean_accuracy": 0.712890625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.285576820373535, + "eval_allNLI-dev_euclidean_ap": 0.5741588797007808, + "eval_allNLI-dev_euclidean_f1": 0.588235294117647, + "eval_allNLI-dev_euclidean_f1_threshold": 12.576900482177734, + "eval_allNLI-dev_euclidean_precision": 0.49603174603174605, + "eval_allNLI-dev_euclidean_recall": 0.7225433526011561, + "eval_allNLI-dev_manhattan_accuracy": 0.70703125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 180.05662536621094, + "eval_allNLI-dev_manhattan_ap": 0.5735452262870777, + "eval_allNLI-dev_manhattan_f1": 0.6046511627906977, + "eval_allNLI-dev_manhattan_f1_threshold": 210.2524871826172, + "eval_allNLI-dev_manhattan_precision": 0.5467289719626168, + "eval_allNLI-dev_manhattan_recall": 0.6763005780346821, + "eval_allNLI-dev_max_accuracy": 0.712890625, + "eval_allNLI-dev_max_accuracy_threshold": 548.634521484375, + "eval_allNLI-dev_max_ap": 0.5745210101303777, + "eval_allNLI-dev_max_f1": 0.6046511627906977, + "eval_allNLI-dev_max_f1_threshold": 427.22930908203125, + "eval_allNLI-dev_max_precision": 0.5467289719626168, + "eval_allNLI-dev_max_recall": 0.7630057803468208, + "eval_sequential_score": 0.6875483049796004, + "eval_sts-test_pearson_cosine": 0.7534891892135007, + "eval_sts-test_pearson_dot": 0.6479194393449418, + "eval_sts-test_pearson_euclidean": 0.7812492751513812, + "eval_sts-test_pearson_manhattan": 0.7840744590454792, + "eval_sts-test_pearson_max": 0.7840744590454792, + "eval_sts-test_spearman_cosine": 0.7930095363399735, + "eval_sts-test_spearman_dot": 0.6299849872121988, + "eval_sts-test_spearman_euclidean": 0.7833179794167452, + "eval_sts-test_spearman_manhattan": 0.7844261101513191, + "eval_sts-test_spearman_max": 0.7930095363399735, + "eval_vitaminc-pairs_loss": 4.219293594360352, + "eval_vitaminc-pairs_runtime": 5.4525, + "eval_vitaminc-pairs_samples_per_second": 23.475, + "eval_vitaminc-pairs_steps_per_second": 0.367, + "step": 480 + }, + { + "epoch": 0.37238169123351433, + "eval_negation-triplets_loss": 2.2595717906951904, + "eval_negation-triplets_runtime": 1.076, + "eval_negation-triplets_samples_per_second": 118.958, + "eval_negation-triplets_steps_per_second": 1.859, + "step": 480 + }, + { + "epoch": 0.37238169123351433, + "eval_scitail-pairs-pos_loss": 0.198380708694458, + "eval_scitail-pairs-pos_runtime": 1.2515, + "eval_scitail-pairs-pos_samples_per_second": 102.28, + "eval_scitail-pairs-pos_steps_per_second": 1.598, + "step": 480 + }, + { + "epoch": 0.37238169123351433, + "eval_scitail-pairs-qa_loss": 0.09023727476596832, + "eval_scitail-pairs-qa_runtime": 0.7729, + "eval_scitail-pairs-qa_samples_per_second": 165.613, + "eval_scitail-pairs-qa_steps_per_second": 2.588, + "step": 480 + }, + { + "epoch": 0.37238169123351433, + "eval_xsum-pairs_loss": 0.7412304878234863, + "eval_xsum-pairs_runtime": 3.9236, + "eval_xsum-pairs_samples_per_second": 32.623, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 480 + }, + { + "epoch": 0.37238169123351433, + "eval_sciq_pairs_loss": 0.11941801011562347, + "eval_sciq_pairs_runtime": 5.8901, + "eval_sciq_pairs_samples_per_second": 21.731, + "eval_sciq_pairs_steps_per_second": 0.34, + "step": 480 + }, + { + "epoch": 0.37238169123351433, + "eval_qasc_pairs_loss": 1.1291967630386353, + "eval_qasc_pairs_runtime": 0.9112, + "eval_qasc_pairs_samples_per_second": 140.478, + "eval_qasc_pairs_steps_per_second": 2.195, + "step": 480 + }, + { + "epoch": 0.37238169123351433, + "eval_openbookqa_pairs_loss": 2.1877193450927734, + "eval_openbookqa_pairs_runtime": 0.8636, + "eval_openbookqa_pairs_samples_per_second": 148.221, + "eval_openbookqa_pairs_steps_per_second": 2.316, + "step": 480 + }, + { + "epoch": 0.37238169123351433, + "eval_msmarco_pairs_loss": 1.241741418838501, + "eval_msmarco_pairs_runtime": 1.9682, + "eval_msmarco_pairs_samples_per_second": 65.033, + "eval_msmarco_pairs_steps_per_second": 1.016, + "step": 480 + }, + { + "epoch": 0.37238169123351433, + "eval_nq_pairs_loss": 1.911564588546753, + "eval_nq_pairs_runtime": 3.3306, + "eval_nq_pairs_samples_per_second": 38.431, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 480 + }, + { + "epoch": 0.37238169123351433, + "eval_trivia_pairs_loss": 1.336583137512207, + "eval_trivia_pairs_runtime": 4.481, + "eval_trivia_pairs_samples_per_second": 28.565, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 480 + }, + { + "epoch": 0.37238169123351433, + "eval_gooaq_pairs_loss": 1.241878628730774, + "eval_gooaq_pairs_runtime": 1.4874, + "eval_gooaq_pairs_samples_per_second": 86.056, + "eval_gooaq_pairs_steps_per_second": 1.345, + "step": 480 + }, + { + "epoch": 0.37238169123351433, + "eval_paws-pos_loss": 0.05792829394340515, + "eval_paws-pos_runtime": 1.0421, + "eval_paws-pos_samples_per_second": 122.826, + "eval_paws-pos_steps_per_second": 1.919, + "step": 480 + }, + { + "epoch": 0.37238169123351433, + "eval_global_dataset_loss": 0.7699050903320312, + "eval_global_dataset_runtime": 19.3968, + "eval_global_dataset_samples_per_second": 21.447, + "eval_global_dataset_steps_per_second": 0.361, + "step": 480 + }, + { + "epoch": 0.37315748642358415, + "grad_norm": 12.873387336730957, + "learning_rate": 1.9659224441833136e-05, + "loss": 0.8126, + "step": 481 + }, + { + "epoch": 0.373933281613654, + "grad_norm": 14.917364120483398, + "learning_rate": 1.970035252643948e-05, + "loss": 1.9393, + "step": 482 + }, + { + "epoch": 0.37470907680372384, + "grad_norm": 13.657282829284668, + "learning_rate": 1.9741480611045827e-05, + "loss": 0.8192, + "step": 483 + }, + { + "epoch": 0.37548487199379366, + "grad_norm": 15.467995643615723, + "learning_rate": 1.9782608695652172e-05, + "loss": 1.3567, + "step": 484 + }, + { + "epoch": 0.3762606671838635, + "grad_norm": 10.490209579467773, + "learning_rate": 1.9823736780258517e-05, + "loss": 0.9228, + "step": 485 + }, + { + "epoch": 0.3770364623739333, + "grad_norm": 10.652816772460938, + "learning_rate": 1.9864864864864863e-05, + "loss": 0.7702, + "step": 486 + }, + { + "epoch": 0.3778122575640031, + "grad_norm": 11.838706016540527, + "learning_rate": 1.990599294947121e-05, + "loss": 1.4665, + "step": 487 + }, + { + "epoch": 0.3785880527540729, + "grad_norm": 12.171431541442871, + "learning_rate": 1.9947121034077553e-05, + "loss": 0.9226, + "step": 488 + }, + { + "epoch": 0.37936384794414274, + "grad_norm": 12.31843090057373, + "learning_rate": 1.99882491186839e-05, + "loss": 0.8567, + "step": 489 + }, + { + "epoch": 0.38013964313421256, + "grad_norm": 14.622387886047363, + "learning_rate": 2.0029377203290244e-05, + "loss": 1.1429, + "step": 490 + }, + { + "epoch": 0.3809154383242824, + "grad_norm": 6.184370994567871, + "learning_rate": 2.007050528789659e-05, + "loss": 0.2455, + "step": 491 + }, + { + "epoch": 0.3816912335143522, + "grad_norm": 14.788031578063965, + "learning_rate": 2.0111633372502938e-05, + "loss": 1.5031, + "step": 492 + }, + { + "epoch": 0.382467028704422, + "grad_norm": 19.639453887939453, + "learning_rate": 2.015276145710928e-05, + "loss": 1.9204, + "step": 493 + }, + { + "epoch": 0.3832428238944919, + "grad_norm": 18.208860397338867, + "learning_rate": 2.019388954171563e-05, + "loss": 1.829, + "step": 494 + }, + { + "epoch": 0.3840186190845617, + "grad_norm": 20.859495162963867, + "learning_rate": 2.0235017626321974e-05, + "loss": 2.4017, + "step": 495 + }, + { + "epoch": 0.3847944142746315, + "grad_norm": 11.328153610229492, + "learning_rate": 2.0276145710928316e-05, + "loss": 1.5618, + "step": 496 + }, + { + "epoch": 0.38557020946470133, + "grad_norm": 14.648991584777832, + "learning_rate": 2.0317273795534664e-05, + "loss": 1.286, + "step": 497 + }, + { + "epoch": 0.38634600465477115, + "grad_norm": 10.970126152038574, + "learning_rate": 2.035840188014101e-05, + "loss": 0.9989, + "step": 498 + }, + { + "epoch": 0.38712179984484096, + "grad_norm": 11.271415710449219, + "learning_rate": 2.0399529964747355e-05, + "loss": 0.6532, + "step": 499 + }, + { + "epoch": 0.3878975950349108, + "grad_norm": 15.105283737182617, + "learning_rate": 2.04406580493537e-05, + "loss": 1.4656, + "step": 500 + }, + { + "epoch": 0.3878975950349108, + "eval_Qnli-dev_cosine_accuracy": 0.669921875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8747369050979614, + "eval_Qnli-dev_cosine_ap": 0.6865438506499877, + "eval_Qnli-dev_cosine_f1": 0.6708860759493672, + "eval_Qnli-dev_cosine_f1_threshold": 0.7570576667785645, + "eval_Qnli-dev_cosine_precision": 0.5353535353535354, + "eval_Qnli-dev_cosine_recall": 0.8983050847457628, + "eval_Qnli-dev_dot_accuracy": 0.623046875, + "eval_Qnli-dev_dot_accuracy_threshold": 463.1806335449219, + "eval_Qnli-dev_dot_ap": 0.5795813697545705, + "eval_Qnli-dev_dot_f1": 0.6461538461538463, + "eval_Qnli-dev_dot_f1_threshold": 408.7640686035156, + "eval_Qnli-dev_dot_precision": 0.5072463768115942, + "eval_Qnli-dev_dot_recall": 0.8898305084745762, + "eval_Qnli-dev_euclidean_accuracy": 0.662109375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.793764114379883, + "eval_Qnli-dev_euclidean_ap": 0.6906311828355566, + "eval_Qnli-dev_euclidean_f1": 0.6688741721854304, + "eval_Qnli-dev_euclidean_f1_threshold": 15.974777221679688, + "eval_Qnli-dev_euclidean_precision": 0.5489130434782609, + "eval_Qnli-dev_euclidean_recall": 0.8559322033898306, + "eval_Qnli-dev_manhattan_accuracy": 0.66015625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 204.43795776367188, + "eval_Qnli-dev_manhattan_ap": 0.6854604224405945, + "eval_Qnli-dev_manhattan_f1": 0.6721854304635762, + "eval_Qnli-dev_manhattan_f1_threshold": 279.64154052734375, + "eval_Qnli-dev_manhattan_precision": 0.5516304347826086, + "eval_Qnli-dev_manhattan_recall": 0.8601694915254238, + "eval_Qnli-dev_max_accuracy": 0.669921875, + "eval_Qnli-dev_max_accuracy_threshold": 463.1806335449219, + "eval_Qnli-dev_max_ap": 0.6906311828355566, + "eval_Qnli-dev_max_f1": 0.6721854304635762, + "eval_Qnli-dev_max_f1_threshold": 408.7640686035156, + "eval_Qnli-dev_max_precision": 0.5516304347826086, + "eval_Qnli-dev_max_recall": 0.8983050847457628, + "eval_allNLI-dev_cosine_accuracy": 0.712890625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8857035636901855, + "eval_allNLI-dev_cosine_ap": 0.571285194257141, + "eval_allNLI-dev_cosine_f1": 0.5807860262008734, + "eval_allNLI-dev_cosine_f1_threshold": 0.8415230512619019, + "eval_allNLI-dev_cosine_precision": 0.4666666666666667, + "eval_allNLI-dev_cosine_recall": 0.7687861271676301, + "eval_allNLI-dev_dot_accuracy": 0.677734375, + "eval_allNLI-dev_dot_accuracy_threshold": 556.57373046875, + "eval_allNLI-dev_dot_ap": 0.45240099555065105, + "eval_allNLI-dev_dot_f1": 0.5364891518737672, + "eval_allNLI-dev_dot_f1_threshold": 430.94512939453125, + "eval_allNLI-dev_dot_precision": 0.40718562874251496, + "eval_allNLI-dev_dot_recall": 0.7861271676300579, + "eval_allNLI-dev_euclidean_accuracy": 0.708984375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 9.428966522216797, + "eval_allNLI-dev_euclidean_ap": 0.5738231649747512, + "eval_allNLI-dev_euclidean_f1": 0.5852631578947369, + "eval_allNLI-dev_euclidean_f1_threshold": 13.404863357543945, + "eval_allNLI-dev_euclidean_precision": 0.4602649006622517, + "eval_allNLI-dev_euclidean_recall": 0.8034682080924855, + "eval_allNLI-dev_manhattan_accuracy": 0.720703125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 188.47079467773438, + "eval_allNLI-dev_manhattan_ap": 0.5751304692008022, + "eval_allNLI-dev_manhattan_f1": 0.5966587112171838, + "eval_allNLI-dev_manhattan_f1_threshold": 219.94509887695312, + "eval_allNLI-dev_manhattan_precision": 0.508130081300813, + "eval_allNLI-dev_manhattan_recall": 0.7225433526011561, + "eval_allNLI-dev_max_accuracy": 0.720703125, + "eval_allNLI-dev_max_accuracy_threshold": 556.57373046875, + "eval_allNLI-dev_max_ap": 0.5751304692008022, + "eval_allNLI-dev_max_f1": 0.5966587112171838, + "eval_allNLI-dev_max_f1_threshold": 430.94512939453125, + "eval_allNLI-dev_max_precision": 0.508130081300813, + "eval_allNLI-dev_max_recall": 0.8034682080924855, + "eval_sequential_score": 0.6906311828355566, + "eval_sts-test_pearson_cosine": 0.771901760871578, + "eval_sts-test_pearson_dot": 0.671396454662895, + "eval_sts-test_pearson_euclidean": 0.7926178039136434, + "eval_sts-test_pearson_manhattan": 0.7942176115421151, + "eval_sts-test_pearson_max": 0.7942176115421151, + "eval_sts-test_spearman_cosine": 0.8078814992818705, + "eval_sts-test_spearman_dot": 0.6573084963723772, + "eval_sts-test_spearman_euclidean": 0.794281904203854, + "eval_sts-test_spearman_manhattan": 0.7947769426573655, + "eval_sts-test_spearman_max": 0.8078814992818705, + "eval_vitaminc-pairs_loss": 4.061150550842285, + "eval_vitaminc-pairs_runtime": 5.4547, + "eval_vitaminc-pairs_samples_per_second": 23.466, + "eval_vitaminc-pairs_steps_per_second": 0.367, + "step": 500 + }, + { + "epoch": 0.3878975950349108, + "eval_negation-triplets_loss": 2.2792365550994873, + "eval_negation-triplets_runtime": 1.0799, + "eval_negation-triplets_samples_per_second": 118.531, + "eval_negation-triplets_steps_per_second": 1.852, + "step": 500 + }, + { + "epoch": 0.3878975950349108, + "eval_scitail-pairs-pos_loss": 0.2175576537847519, + "eval_scitail-pairs-pos_runtime": 1.2619, + "eval_scitail-pairs-pos_samples_per_second": 101.434, + "eval_scitail-pairs-pos_steps_per_second": 1.585, + "step": 500 + }, + { + "epoch": 0.3878975950349108, + "eval_scitail-pairs-qa_loss": 0.07303012162446976, + "eval_scitail-pairs-qa_runtime": 0.7684, + "eval_scitail-pairs-qa_samples_per_second": 166.591, + "eval_scitail-pairs-qa_steps_per_second": 2.603, + "step": 500 + }, + { + "epoch": 0.3878975950349108, + "eval_xsum-pairs_loss": 0.6853438019752502, + "eval_xsum-pairs_runtime": 3.9237, + "eval_xsum-pairs_samples_per_second": 32.622, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 500 + }, + { + "epoch": 0.3878975950349108, + "eval_sciq_pairs_loss": 0.10957910120487213, + "eval_sciq_pairs_runtime": 5.9387, + "eval_sciq_pairs_samples_per_second": 21.553, + "eval_sciq_pairs_steps_per_second": 0.337, + "step": 500 + }, + { + "epoch": 0.3878975950349108, + "eval_qasc_pairs_loss": 1.0403637886047363, + "eval_qasc_pairs_runtime": 0.9266, + "eval_qasc_pairs_samples_per_second": 138.145, + "eval_qasc_pairs_steps_per_second": 2.159, + "step": 500 + }, + { + "epoch": 0.3878975950349108, + "eval_openbookqa_pairs_loss": 2.1081655025482178, + "eval_openbookqa_pairs_runtime": 0.8637, + "eval_openbookqa_pairs_samples_per_second": 148.197, + "eval_openbookqa_pairs_steps_per_second": 2.316, + "step": 500 + }, + { + "epoch": 0.3878975950349108, + "eval_msmarco_pairs_loss": 1.0579735040664673, + "eval_msmarco_pairs_runtime": 1.9798, + "eval_msmarco_pairs_samples_per_second": 64.653, + "eval_msmarco_pairs_steps_per_second": 1.01, + "step": 500 + }, + { + "epoch": 0.3878975950349108, + "eval_nq_pairs_loss": 1.645302414894104, + "eval_nq_pairs_runtime": 3.3362, + "eval_nq_pairs_samples_per_second": 38.368, + "eval_nq_pairs_steps_per_second": 0.599, + "step": 500 + }, + { + "epoch": 0.3878975950349108, + "eval_trivia_pairs_loss": 1.2547576427459717, + "eval_trivia_pairs_runtime": 4.4885, + "eval_trivia_pairs_samples_per_second": 28.517, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 500 + }, + { + "epoch": 0.3878975950349108, + "eval_gooaq_pairs_loss": 1.167790412902832, + "eval_gooaq_pairs_runtime": 1.4668, + "eval_gooaq_pairs_samples_per_second": 87.264, + "eval_gooaq_pairs_steps_per_second": 1.364, + "step": 500 + }, + { + "epoch": 0.3878975950349108, + "eval_paws-pos_loss": 0.059369876980781555, + "eval_paws-pos_runtime": 1.0333, + "eval_paws-pos_samples_per_second": 123.881, + "eval_paws-pos_steps_per_second": 1.936, + "step": 500 + }, + { + "epoch": 0.3878975950349108, + "eval_global_dataset_loss": 0.6943269968032837, + "eval_global_dataset_runtime": 19.4141, + "eval_global_dataset_samples_per_second": 21.428, + "eval_global_dataset_steps_per_second": 0.361, + "step": 500 + }, + { + "epoch": 0.3886733902249806, + "grad_norm": 12.631108283996582, + "learning_rate": 2.0481786133960042e-05, + "loss": 1.6093, + "step": 501 + }, + { + "epoch": 0.3894491854150504, + "grad_norm": 14.593748092651367, + "learning_rate": 2.052291421856639e-05, + "loss": 1.6896, + "step": 502 + }, + { + "epoch": 0.39022498060512023, + "grad_norm": 13.730030059814453, + "learning_rate": 2.0564042303172736e-05, + "loss": 1.3149, + "step": 503 + }, + { + "epoch": 0.39100077579519005, + "grad_norm": 11.460733413696289, + "learning_rate": 2.060517038777908e-05, + "loss": 1.0461, + "step": 504 + }, + { + "epoch": 0.39177657098525986, + "grad_norm": 11.096057891845703, + "learning_rate": 2.0646298472385427e-05, + "loss": 1.2417, + "step": 505 + }, + { + "epoch": 0.39255236617532974, + "grad_norm": 10.991378784179688, + "learning_rate": 2.0687426556991772e-05, + "loss": 0.7121, + "step": 506 + }, + { + "epoch": 0.39332816136539955, + "grad_norm": 9.571194648742676, + "learning_rate": 2.0728554641598117e-05, + "loss": 0.8205, + "step": 507 + }, + { + "epoch": 0.39410395655546937, + "grad_norm": 6.761598587036133, + "learning_rate": 2.0769682726204463e-05, + "loss": 0.2886, + "step": 508 + }, + { + "epoch": 0.3948797517455392, + "grad_norm": 10.539600372314453, + "learning_rate": 2.081081081081081e-05, + "loss": 0.6667, + "step": 509 + }, + { + "epoch": 0.395655546935609, + "grad_norm": 18.16293716430664, + "learning_rate": 2.0851938895417153e-05, + "loss": 1.7981, + "step": 510 + }, + { + "epoch": 0.3964313421256788, + "grad_norm": 8.688652038574219, + "learning_rate": 2.0893066980023502e-05, + "loss": 0.4364, + "step": 511 + }, + { + "epoch": 0.39720713731574864, + "grad_norm": 11.366159439086914, + "learning_rate": 2.0934195064629847e-05, + "loss": 0.7245, + "step": 512 + }, + { + "epoch": 0.39798293250581845, + "grad_norm": 9.486770629882812, + "learning_rate": 2.097532314923619e-05, + "loss": 0.6328, + "step": 513 + }, + { + "epoch": 0.39875872769588827, + "grad_norm": 14.476900100708008, + "learning_rate": 2.1016451233842538e-05, + "loss": 1.3796, + "step": 514 + }, + { + "epoch": 0.3995345228859581, + "grad_norm": 14.469146728515625, + "learning_rate": 2.105757931844888e-05, + "loss": 1.2157, + "step": 515 + }, + { + "epoch": 0.4003103180760279, + "grad_norm": 17.520402908325195, + "learning_rate": 2.1098707403055228e-05, + "loss": 1.4128, + "step": 516 + }, + { + "epoch": 0.4010861132660978, + "grad_norm": 14.814453125, + "learning_rate": 2.1139835487661573e-05, + "loss": 2.0686, + "step": 517 + }, + { + "epoch": 0.4018619084561676, + "grad_norm": 9.749579429626465, + "learning_rate": 2.1180963572267915e-05, + "loss": 1.115, + "step": 518 + }, + { + "epoch": 0.4026377036462374, + "grad_norm": 11.475977897644043, + "learning_rate": 2.1222091656874264e-05, + "loss": 0.8495, + "step": 519 + }, + { + "epoch": 0.4034134988363072, + "grad_norm": 10.206400871276855, + "learning_rate": 2.126321974148061e-05, + "loss": 0.5301, + "step": 520 + }, + { + "epoch": 0.4034134988363072, + "eval_Qnli-dev_cosine_accuracy": 0.6640625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8630666136741638, + "eval_Qnli-dev_cosine_ap": 0.6864801399197513, + "eval_Qnli-dev_cosine_f1": 0.6676923076923076, + "eval_Qnli-dev_cosine_f1_threshold": 0.7470320463180542, + "eval_Qnli-dev_cosine_precision": 0.5241545893719807, + "eval_Qnli-dev_cosine_recall": 0.9194915254237288, + "eval_Qnli-dev_dot_accuracy": 0.626953125, + "eval_Qnli-dev_dot_accuracy_threshold": 509.13177490234375, + "eval_Qnli-dev_dot_ap": 0.5961767874874291, + "eval_Qnli-dev_dot_f1": 0.6540145985401459, + "eval_Qnli-dev_dot_f1_threshold": 382.5948486328125, + "eval_Qnli-dev_dot_precision": 0.49888641425389757, + "eval_Qnli-dev_dot_recall": 0.9491525423728814, + "eval_Qnli-dev_euclidean_accuracy": 0.6640625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.71471118927002, + "eval_Qnli-dev_euclidean_ap": 0.6869527631961445, + "eval_Qnli-dev_euclidean_f1": 0.6711185308848081, + "eval_Qnli-dev_euclidean_f1_threshold": 15.946113586425781, + "eval_Qnli-dev_euclidean_precision": 0.5537190082644629, + "eval_Qnli-dev_euclidean_recall": 0.8516949152542372, + "eval_Qnli-dev_manhattan_accuracy": 0.671875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 208.53021240234375, + "eval_Qnli-dev_manhattan_ap": 0.6899294790824355, + "eval_Qnli-dev_manhattan_f1": 0.6754098360655738, + "eval_Qnli-dev_manhattan_f1_threshold": 283.5461730957031, + "eval_Qnli-dev_manhattan_precision": 0.5508021390374331, + "eval_Qnli-dev_manhattan_recall": 0.8728813559322034, + "eval_Qnli-dev_max_accuracy": 0.671875, + "eval_Qnli-dev_max_accuracy_threshold": 509.13177490234375, + "eval_Qnli-dev_max_ap": 0.6899294790824355, + "eval_Qnli-dev_max_f1": 0.6754098360655738, + "eval_Qnli-dev_max_f1_threshold": 382.5948486328125, + "eval_Qnli-dev_max_precision": 0.5537190082644629, + "eval_Qnli-dev_max_recall": 0.9491525423728814, + "eval_allNLI-dev_cosine_accuracy": 0.71484375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9217908382415771, + "eval_allNLI-dev_cosine_ap": 0.5865821070643723, + "eval_allNLI-dev_cosine_f1": 0.5882352941176471, + "eval_allNLI-dev_cosine_f1_threshold": 0.8343489170074463, + "eval_allNLI-dev_cosine_precision": 0.483271375464684, + "eval_allNLI-dev_cosine_recall": 0.7514450867052023, + "eval_allNLI-dev_dot_accuracy": 0.677734375, + "eval_allNLI-dev_dot_accuracy_threshold": 541.32568359375, + "eval_allNLI-dev_dot_ap": 0.47010002301609466, + "eval_allNLI-dev_dot_f1": 0.5317725752508361, + "eval_allNLI-dev_dot_f1_threshold": 379.1038818359375, + "eval_allNLI-dev_dot_precision": 0.37411764705882355, + "eval_allNLI-dev_dot_recall": 0.9190751445086706, + "eval_allNLI-dev_euclidean_accuracy": 0.716796875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.514617919921875, + "eval_allNLI-dev_euclidean_ap": 0.5824948849494698, + "eval_allNLI-dev_euclidean_f1": 0.5829787234042553, + "eval_allNLI-dev_euclidean_f1_threshold": 14.028602600097656, + "eval_allNLI-dev_euclidean_precision": 0.4612794612794613, + "eval_allNLI-dev_euclidean_recall": 0.791907514450867, + "eval_allNLI-dev_manhattan_accuracy": 0.716796875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 186.78890991210938, + "eval_allNLI-dev_manhattan_ap": 0.5825071015866208, + "eval_allNLI-dev_manhattan_f1": 0.5952380952380952, + "eval_allNLI-dev_manhattan_f1_threshold": 229.18360900878906, + "eval_allNLI-dev_manhattan_precision": 0.5060728744939271, + "eval_allNLI-dev_manhattan_recall": 0.7225433526011561, + "eval_allNLI-dev_max_accuracy": 0.716796875, + "eval_allNLI-dev_max_accuracy_threshold": 541.32568359375, + "eval_allNLI-dev_max_ap": 0.5865821070643723, + "eval_allNLI-dev_max_f1": 0.5952380952380952, + "eval_allNLI-dev_max_f1_threshold": 379.1038818359375, + "eval_allNLI-dev_max_precision": 0.5060728744939271, + "eval_allNLI-dev_max_recall": 0.9190751445086706, + "eval_sequential_score": 0.6899294790824355, + "eval_sts-test_pearson_cosine": 0.7943926827825525, + "eval_sts-test_pearson_dot": 0.6888392553853662, + "eval_sts-test_pearson_euclidean": 0.8146723367382795, + "eval_sts-test_pearson_manhattan": 0.8140557940999833, + "eval_sts-test_pearson_max": 0.8146723367382795, + "eval_sts-test_spearman_cosine": 0.8229409700089974, + "eval_sts-test_spearman_dot": 0.6768808633032584, + "eval_sts-test_spearman_euclidean": 0.8112659029271178, + "eval_sts-test_spearman_manhattan": 0.8109135762461833, + "eval_sts-test_spearman_max": 0.8229409700089974, + "eval_vitaminc-pairs_loss": 4.049212455749512, + "eval_vitaminc-pairs_runtime": 5.5044, + "eval_vitaminc-pairs_samples_per_second": 23.254, + "eval_vitaminc-pairs_steps_per_second": 0.363, + "step": 520 + }, + { + "epoch": 0.4034134988363072, + "eval_negation-triplets_loss": 2.2466907501220703, + "eval_negation-triplets_runtime": 1.1021, + "eval_negation-triplets_samples_per_second": 116.146, + "eval_negation-triplets_steps_per_second": 1.815, + "step": 520 + }, + { + "epoch": 0.4034134988363072, + "eval_scitail-pairs-pos_loss": 0.2528344690799713, + "eval_scitail-pairs-pos_runtime": 1.2466, + "eval_scitail-pairs-pos_samples_per_second": 102.677, + "eval_scitail-pairs-pos_steps_per_second": 1.604, + "step": 520 + }, + { + "epoch": 0.4034134988363072, + "eval_scitail-pairs-qa_loss": 0.08246471732854843, + "eval_scitail-pairs-qa_runtime": 0.8541, + "eval_scitail-pairs-qa_samples_per_second": 149.862, + "eval_scitail-pairs-qa_steps_per_second": 2.342, + "step": 520 + }, + { + "epoch": 0.4034134988363072, + "eval_xsum-pairs_loss": 0.6523213386535645, + "eval_xsum-pairs_runtime": 3.9686, + "eval_xsum-pairs_samples_per_second": 32.253, + "eval_xsum-pairs_steps_per_second": 0.504, + "step": 520 + }, + { + "epoch": 0.4034134988363072, + "eval_sciq_pairs_loss": 0.10934119671583176, + "eval_sciq_pairs_runtime": 6.0262, + "eval_sciq_pairs_samples_per_second": 21.241, + "eval_sciq_pairs_steps_per_second": 0.332, + "step": 520 + }, + { + "epoch": 0.4034134988363072, + "eval_qasc_pairs_loss": 1.0355889797210693, + "eval_qasc_pairs_runtime": 0.9133, + "eval_qasc_pairs_samples_per_second": 140.148, + "eval_qasc_pairs_steps_per_second": 2.19, + "step": 520 + }, + { + "epoch": 0.4034134988363072, + "eval_openbookqa_pairs_loss": 2.0900936126708984, + "eval_openbookqa_pairs_runtime": 0.8706, + "eval_openbookqa_pairs_samples_per_second": 147.029, + "eval_openbookqa_pairs_steps_per_second": 2.297, + "step": 520 + }, + { + "epoch": 0.4034134988363072, + "eval_msmarco_pairs_loss": 0.9958593249320984, + "eval_msmarco_pairs_runtime": 1.9878, + "eval_msmarco_pairs_samples_per_second": 64.394, + "eval_msmarco_pairs_steps_per_second": 1.006, + "step": 520 + }, + { + "epoch": 0.4034134988363072, + "eval_nq_pairs_loss": 1.6257421970367432, + "eval_nq_pairs_runtime": 3.3565, + "eval_nq_pairs_samples_per_second": 38.135, + "eval_nq_pairs_steps_per_second": 0.596, + "step": 520 + }, + { + "epoch": 0.4034134988363072, + "eval_trivia_pairs_loss": 1.2904225587844849, + "eval_trivia_pairs_runtime": 4.5418, + "eval_trivia_pairs_samples_per_second": 28.183, + "eval_trivia_pairs_steps_per_second": 0.44, + "step": 520 + }, + { + "epoch": 0.4034134988363072, + "eval_gooaq_pairs_loss": 1.158769130706787, + "eval_gooaq_pairs_runtime": 1.5102, + "eval_gooaq_pairs_samples_per_second": 84.756, + "eval_gooaq_pairs_steps_per_second": 1.324, + "step": 520 + }, + { + "epoch": 0.4034134988363072, + "eval_paws-pos_loss": 0.056799858808517456, + "eval_paws-pos_runtime": 1.0837, + "eval_paws-pos_samples_per_second": 118.116, + "eval_paws-pos_steps_per_second": 1.846, + "step": 520 + }, + { + "epoch": 0.4034134988363072, + "eval_global_dataset_loss": 0.703998327255249, + "eval_global_dataset_runtime": 19.4918, + "eval_global_dataset_samples_per_second": 21.342, + "eval_global_dataset_steps_per_second": 0.359, + "step": 520 + }, + { + "epoch": 0.40418929402637704, + "grad_norm": 17.330259323120117, + "learning_rate": 2.1304347826086955e-05, + "loss": 1.534, + "step": 521 + }, + { + "epoch": 0.40496508921644686, + "grad_norm": 17.98264503479004, + "learning_rate": 2.13454759106933e-05, + "loss": 2.8681, + "step": 522 + }, + { + "epoch": 0.4057408844065167, + "grad_norm": 12.535171508789062, + "learning_rate": 2.1386603995299645e-05, + "loss": 1.4974, + "step": 523 + }, + { + "epoch": 0.4065166795965865, + "grad_norm": 15.35007381439209, + "learning_rate": 2.142773207990599e-05, + "loss": 1.3792, + "step": 524 + }, + { + "epoch": 0.4072924747866563, + "grad_norm": 10.744861602783203, + "learning_rate": 2.1468860164512336e-05, + "loss": 0.8891, + "step": 525 + }, + { + "epoch": 0.40806826997672613, + "grad_norm": 9.942862510681152, + "learning_rate": 2.1509988249118684e-05, + "loss": 0.6328, + "step": 526 + }, + { + "epoch": 0.40884406516679594, + "grad_norm": 15.72616958618164, + "learning_rate": 2.1551116333725026e-05, + "loss": 1.8621, + "step": 527 + }, + { + "epoch": 0.40961986035686576, + "grad_norm": 7.739931106567383, + "learning_rate": 2.1592244418331375e-05, + "loss": 0.4133, + "step": 528 + }, + { + "epoch": 0.41039565554693563, + "grad_norm": 9.970248222351074, + "learning_rate": 2.1633372502937717e-05, + "loss": 0.7418, + "step": 529 + }, + { + "epoch": 0.41117145073700545, + "grad_norm": 9.07480525970459, + "learning_rate": 2.1674500587544062e-05, + "loss": 0.5625, + "step": 530 + }, + { + "epoch": 0.41194724592707527, + "grad_norm": 8.35373306274414, + "learning_rate": 2.171562867215041e-05, + "loss": 0.5545, + "step": 531 + }, + { + "epoch": 0.4127230411171451, + "grad_norm": 10.86092472076416, + "learning_rate": 2.1756756756756753e-05, + "loss": 1.1808, + "step": 532 + }, + { + "epoch": 0.4134988363072149, + "grad_norm": 14.283552169799805, + "learning_rate": 2.17978848413631e-05, + "loss": 1.8035, + "step": 533 + }, + { + "epoch": 0.4142746314972847, + "grad_norm": 4.8627424240112305, + "learning_rate": 2.1839012925969447e-05, + "loss": 0.1492, + "step": 534 + }, + { + "epoch": 0.41505042668735453, + "grad_norm": 10.93820858001709, + "learning_rate": 2.188014101057579e-05, + "loss": 1.2314, + "step": 535 + }, + { + "epoch": 0.41582622187742435, + "grad_norm": 14.4031982421875, + "learning_rate": 2.1921269095182137e-05, + "loss": 1.5375, + "step": 536 + }, + { + "epoch": 0.41660201706749417, + "grad_norm": 23.42047119140625, + "learning_rate": 2.1962397179788483e-05, + "loss": 2.9917, + "step": 537 + }, + { + "epoch": 0.417377812257564, + "grad_norm": 11.167613983154297, + "learning_rate": 2.2003525264394828e-05, + "loss": 0.6153, + "step": 538 + }, + { + "epoch": 0.4181536074476338, + "grad_norm": 22.974531173706055, + "learning_rate": 2.2044653349001173e-05, + "loss": 2.3485, + "step": 539 + }, + { + "epoch": 0.4189294026377037, + "grad_norm": 13.186490058898926, + "learning_rate": 2.2085781433607515e-05, + "loss": 0.7165, + "step": 540 + }, + { + "epoch": 0.4189294026377037, + "eval_Qnli-dev_cosine_accuracy": 0.677734375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8607775568962097, + "eval_Qnli-dev_cosine_ap": 0.6876732451586249, + "eval_Qnli-dev_cosine_f1": 0.674496644295302, + "eval_Qnli-dev_cosine_f1_threshold": 0.7569329738616943, + "eval_Qnli-dev_cosine_precision": 0.5583333333333333, + "eval_Qnli-dev_cosine_recall": 0.8516949152542372, + "eval_Qnli-dev_dot_accuracy": 0.6328125, + "eval_Qnli-dev_dot_accuracy_threshold": 435.5442810058594, + "eval_Qnli-dev_dot_ap": 0.6207899421535183, + "eval_Qnli-dev_dot_f1": 0.6634460547504026, + "eval_Qnli-dev_dot_f1_threshold": 364.42340087890625, + "eval_Qnli-dev_dot_precision": 0.535064935064935, + "eval_Qnli-dev_dot_recall": 0.8728813559322034, + "eval_Qnli-dev_euclidean_accuracy": 0.66015625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.227877616882324, + "eval_Qnli-dev_euclidean_ap": 0.6882683194952817, + "eval_Qnli-dev_euclidean_f1": 0.6781411359724612, + "eval_Qnli-dev_euclidean_f1_threshold": 15.63144302368164, + "eval_Qnli-dev_euclidean_precision": 0.5710144927536231, + "eval_Qnli-dev_euclidean_recall": 0.8347457627118644, + "eval_Qnli-dev_manhattan_accuracy": 0.66796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 207.87457275390625, + "eval_Qnli-dev_manhattan_ap": 0.686343836883915, + "eval_Qnli-dev_manhattan_f1": 0.6767169179229481, + "eval_Qnli-dev_manhattan_f1_threshold": 283.05743408203125, + "eval_Qnli-dev_manhattan_precision": 0.5595567867036011, + "eval_Qnli-dev_manhattan_recall": 0.8559322033898306, + "eval_Qnli-dev_max_accuracy": 0.677734375, + "eval_Qnli-dev_max_accuracy_threshold": 435.5442810058594, + "eval_Qnli-dev_max_ap": 0.6882683194952817, + "eval_Qnli-dev_max_f1": 0.6781411359724612, + "eval_Qnli-dev_max_f1_threshold": 364.42340087890625, + "eval_Qnli-dev_max_precision": 0.5710144927536231, + "eval_Qnli-dev_max_recall": 0.8728813559322034, + "eval_allNLI-dev_cosine_accuracy": 0.70703125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9104803800582886, + "eval_allNLI-dev_cosine_ap": 0.5805694368067533, + "eval_allNLI-dev_cosine_f1": 0.5803921568627451, + "eval_allNLI-dev_cosine_f1_threshold": 0.7738338708877563, + "eval_allNLI-dev_cosine_precision": 0.4391691394658754, + "eval_allNLI-dev_cosine_recall": 0.8554913294797688, + "eval_allNLI-dev_dot_accuracy": 0.6796875, + "eval_allNLI-dev_dot_accuracy_threshold": 460.63214111328125, + "eval_allNLI-dev_dot_ap": 0.4777128944678729, + "eval_allNLI-dev_dot_f1": 0.5577981651376146, + "eval_allNLI-dev_dot_f1_threshold": 351.3927917480469, + "eval_allNLI-dev_dot_precision": 0.40860215053763443, + "eval_allNLI-dev_dot_recall": 0.8786127167630058, + "eval_allNLI-dev_euclidean_accuracy": 0.716796875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.379776954650879, + "eval_allNLI-dev_euclidean_ap": 0.5815048328459478, + "eval_allNLI-dev_euclidean_f1": 0.5860655737704918, + "eval_allNLI-dev_euclidean_f1_threshold": 14.440160751342773, + "eval_allNLI-dev_euclidean_precision": 0.45396825396825397, + "eval_allNLI-dev_euclidean_recall": 0.8265895953757225, + "eval_allNLI-dev_manhattan_accuracy": 0.720703125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 189.49766540527344, + "eval_allNLI-dev_manhattan_ap": 0.5851021469692314, + "eval_allNLI-dev_manhattan_f1": 0.5879518072289156, + "eval_allNLI-dev_manhattan_f1_threshold": 235.60931396484375, + "eval_allNLI-dev_manhattan_precision": 0.5041322314049587, + "eval_allNLI-dev_manhattan_recall": 0.7052023121387283, + "eval_allNLI-dev_max_accuracy": 0.720703125, + "eval_allNLI-dev_max_accuracy_threshold": 460.63214111328125, + "eval_allNLI-dev_max_ap": 0.5851021469692314, + "eval_allNLI-dev_max_f1": 0.5879518072289156, + "eval_allNLI-dev_max_f1_threshold": 351.3927917480469, + "eval_allNLI-dev_max_precision": 0.5041322314049587, + "eval_allNLI-dev_max_recall": 0.8786127167630058, + "eval_sequential_score": 0.6882683194952817, + "eval_sts-test_pearson_cosine": 0.7891107355413437, + "eval_sts-test_pearson_dot": 0.7386287226644515, + "eval_sts-test_pearson_euclidean": 0.8139455031957386, + "eval_sts-test_pearson_manhattan": 0.8126098848427838, + "eval_sts-test_pearson_max": 0.8139455031957386, + "eval_sts-test_spearman_cosine": 0.8216120101639881, + "eval_sts-test_spearman_dot": 0.7303319041629301, + "eval_sts-test_spearman_euclidean": 0.8135028394790285, + "eval_sts-test_spearman_manhattan": 0.8127091758582519, + "eval_sts-test_spearman_max": 0.8216120101639881, + "eval_vitaminc-pairs_loss": 4.483358383178711, + "eval_vitaminc-pairs_runtime": 5.5001, + "eval_vitaminc-pairs_samples_per_second": 23.272, + "eval_vitaminc-pairs_steps_per_second": 0.364, + "step": 540 + }, + { + "epoch": 0.4189294026377037, + "eval_negation-triplets_loss": 2.18699312210083, + "eval_negation-triplets_runtime": 1.0783, + "eval_negation-triplets_samples_per_second": 118.7, + "eval_negation-triplets_steps_per_second": 1.855, + "step": 540 + }, + { + "epoch": 0.4189294026377037, + "eval_scitail-pairs-pos_loss": 0.3008614778518677, + "eval_scitail-pairs-pos_runtime": 1.2724, + "eval_scitail-pairs-pos_samples_per_second": 100.597, + "eval_scitail-pairs-pos_steps_per_second": 1.572, + "step": 540 + }, + { + "epoch": 0.4189294026377037, + "eval_scitail-pairs-qa_loss": 0.07090522348880768, + "eval_scitail-pairs-qa_runtime": 0.7905, + "eval_scitail-pairs-qa_samples_per_second": 161.927, + "eval_scitail-pairs-qa_steps_per_second": 2.53, + "step": 540 + }, + { + "epoch": 0.4189294026377037, + "eval_xsum-pairs_loss": 0.6539483666419983, + "eval_xsum-pairs_runtime": 3.9209, + "eval_xsum-pairs_samples_per_second": 32.646, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 540 + }, + { + "epoch": 0.4189294026377037, + "eval_sciq_pairs_loss": 0.10181967914104462, + "eval_sciq_pairs_runtime": 5.9007, + "eval_sciq_pairs_samples_per_second": 21.692, + "eval_sciq_pairs_steps_per_second": 0.339, + "step": 540 + }, + { + "epoch": 0.4189294026377037, + "eval_qasc_pairs_loss": 1.0931895971298218, + "eval_qasc_pairs_runtime": 0.9228, + "eval_qasc_pairs_samples_per_second": 138.714, + "eval_qasc_pairs_steps_per_second": 2.167, + "step": 540 + }, + { + "epoch": 0.4189294026377037, + "eval_openbookqa_pairs_loss": 2.048208713531494, + "eval_openbookqa_pairs_runtime": 0.8647, + "eval_openbookqa_pairs_samples_per_second": 148.025, + "eval_openbookqa_pairs_steps_per_second": 2.313, + "step": 540 + }, + { + "epoch": 0.4189294026377037, + "eval_msmarco_pairs_loss": 0.9137493968009949, + "eval_msmarco_pairs_runtime": 1.9669, + "eval_msmarco_pairs_samples_per_second": 65.078, + "eval_msmarco_pairs_steps_per_second": 1.017, + "step": 540 + }, + { + "epoch": 0.4189294026377037, + "eval_nq_pairs_loss": 1.7704282999038696, + "eval_nq_pairs_runtime": 3.3334, + "eval_nq_pairs_samples_per_second": 38.399, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 540 + }, + { + "epoch": 0.4189294026377037, + "eval_trivia_pairs_loss": 1.327826976776123, + "eval_trivia_pairs_runtime": 4.475, + "eval_trivia_pairs_samples_per_second": 28.603, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 540 + }, + { + "epoch": 0.4189294026377037, + "eval_gooaq_pairs_loss": 1.157819390296936, + "eval_gooaq_pairs_runtime": 1.4597, + "eval_gooaq_pairs_samples_per_second": 87.692, + "eval_gooaq_pairs_steps_per_second": 1.37, + "step": 540 + }, + { + "epoch": 0.4189294026377037, + "eval_paws-pos_loss": 0.051995135843753815, + "eval_paws-pos_runtime": 1.0319, + "eval_paws-pos_samples_per_second": 124.049, + "eval_paws-pos_steps_per_second": 1.938, + "step": 540 + }, + { + "epoch": 0.4189294026377037, + "eval_global_dataset_loss": 0.716935932636261, + "eval_global_dataset_runtime": 19.4318, + "eval_global_dataset_samples_per_second": 21.408, + "eval_global_dataset_steps_per_second": 0.36, + "step": 540 + }, + { + "epoch": 0.4197051978277735, + "grad_norm": 12.334939956665039, + "learning_rate": 2.2126909518213864e-05, + "loss": 0.5742, + "step": 541 + }, + { + "epoch": 0.4204809930178433, + "grad_norm": 1.654783010482788, + "learning_rate": 2.216803760282021e-05, + "loss": 0.0717, + "step": 542 + }, + { + "epoch": 0.4212567882079131, + "grad_norm": 7.320519924163818, + "learning_rate": 2.2209165687426558e-05, + "loss": 0.2671, + "step": 543 + }, + { + "epoch": 0.42203258339798294, + "grad_norm": 10.110030174255371, + "learning_rate": 2.22502937720329e-05, + "loss": 0.4795, + "step": 544 + }, + { + "epoch": 0.42280837858805276, + "grad_norm": 17.791595458984375, + "learning_rate": 2.229142185663925e-05, + "loss": 1.2338, + "step": 545 + }, + { + "epoch": 0.4235841737781226, + "grad_norm": 10.264248847961426, + "learning_rate": 2.233254994124559e-05, + "loss": 0.6141, + "step": 546 + }, + { + "epoch": 0.4243599689681924, + "grad_norm": 11.364494323730469, + "learning_rate": 2.2373678025851936e-05, + "loss": 0.6861, + "step": 547 + }, + { + "epoch": 0.4251357641582622, + "grad_norm": 14.532057762145996, + "learning_rate": 2.2414806110458284e-05, + "loss": 1.1294, + "step": 548 + }, + { + "epoch": 0.425911559348332, + "grad_norm": 16.047151565551758, + "learning_rate": 2.2455934195064626e-05, + "loss": 1.5741, + "step": 549 + }, + { + "epoch": 0.42668735453840184, + "grad_norm": 13.938572883605957, + "learning_rate": 2.2497062279670975e-05, + "loss": 1.8814, + "step": 550 + }, + { + "epoch": 0.42746314972847166, + "grad_norm": 12.493745803833008, + "learning_rate": 2.253819036427732e-05, + "loss": 1.4718, + "step": 551 + }, + { + "epoch": 0.42823894491854153, + "grad_norm": 12.112238883972168, + "learning_rate": 2.2579318448883662e-05, + "loss": 0.8628, + "step": 552 + }, + { + "epoch": 0.42901474010861135, + "grad_norm": 16.684898376464844, + "learning_rate": 2.262044653349001e-05, + "loss": 1.7411, + "step": 553 + }, + { + "epoch": 0.42979053529868116, + "grad_norm": 16.532133102416992, + "learning_rate": 2.2661574618096353e-05, + "loss": 1.6768, + "step": 554 + }, + { + "epoch": 0.430566330488751, + "grad_norm": 9.877392768859863, + "learning_rate": 2.27027027027027e-05, + "loss": 0.6208, + "step": 555 + }, + { + "epoch": 0.4313421256788208, + "grad_norm": 8.792377471923828, + "learning_rate": 2.2743830787309047e-05, + "loss": 0.7062, + "step": 556 + }, + { + "epoch": 0.4321179208688906, + "grad_norm": 10.039380073547363, + "learning_rate": 2.2784958871915395e-05, + "loss": 1.0287, + "step": 557 + }, + { + "epoch": 0.43289371605896043, + "grad_norm": 10.16191291809082, + "learning_rate": 2.2826086956521737e-05, + "loss": 0.641, + "step": 558 + }, + { + "epoch": 0.43366951124903025, + "grad_norm": 7.070035934448242, + "learning_rate": 2.2867215041128082e-05, + "loss": 0.407, + "step": 559 + }, + { + "epoch": 0.43444530643910007, + "grad_norm": 12.430944442749023, + "learning_rate": 2.2908343125734428e-05, + "loss": 1.6435, + "step": 560 + }, + { + "epoch": 0.43444530643910007, + "eval_Qnli-dev_cosine_accuracy": 0.6796875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8646827936172485, + "eval_Qnli-dev_cosine_ap": 0.7047500859185081, + "eval_Qnli-dev_cosine_f1": 0.6735751295336788, + "eval_Qnli-dev_cosine_f1_threshold": 0.7815592288970947, + "eval_Qnli-dev_cosine_precision": 0.5685131195335277, + "eval_Qnli-dev_cosine_recall": 0.826271186440678, + "eval_Qnli-dev_dot_accuracy": 0.638671875, + "eval_Qnli-dev_dot_accuracy_threshold": 440.2266540527344, + "eval_Qnli-dev_dot_ap": 0.6266901766954593, + "eval_Qnli-dev_dot_f1": 0.6521106259097526, + "eval_Qnli-dev_dot_f1_threshold": 354.78875732421875, + "eval_Qnli-dev_dot_precision": 0.49667405764966743, + "eval_Qnli-dev_dot_recall": 0.9491525423728814, + "eval_Qnli-dev_euclidean_accuracy": 0.671875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.156564712524414, + "eval_Qnli-dev_euclidean_ap": 0.7035172952301721, + "eval_Qnli-dev_euclidean_f1": 0.6764168190127972, + "eval_Qnli-dev_euclidean_f1_threshold": 14.6650390625, + "eval_Qnli-dev_euclidean_precision": 0.594855305466238, + "eval_Qnli-dev_euclidean_recall": 0.7838983050847458, + "eval_Qnli-dev_manhattan_accuracy": 0.6796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 208.739501953125, + "eval_Qnli-dev_manhattan_ap": 0.7008656551075609, + "eval_Qnli-dev_manhattan_f1": 0.682142857142857, + "eval_Qnli-dev_manhattan_f1_threshold": 264.25848388671875, + "eval_Qnli-dev_manhattan_precision": 0.5895061728395061, + "eval_Qnli-dev_manhattan_recall": 0.809322033898305, + "eval_Qnli-dev_max_accuracy": 0.6796875, + "eval_Qnli-dev_max_accuracy_threshold": 440.2266540527344, + "eval_Qnli-dev_max_ap": 0.7047500859185081, + "eval_Qnli-dev_max_f1": 0.682142857142857, + "eval_Qnli-dev_max_f1_threshold": 354.78875732421875, + "eval_Qnli-dev_max_precision": 0.594855305466238, + "eval_Qnli-dev_max_recall": 0.9491525423728814, + "eval_allNLI-dev_cosine_accuracy": 0.71484375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9083981513977051, + "eval_allNLI-dev_cosine_ap": 0.5860855983218396, + "eval_allNLI-dev_cosine_f1": 0.5901639344262295, + "eval_allNLI-dev_cosine_f1_threshold": 0.7923876047134399, + "eval_allNLI-dev_cosine_precision": 0.45714285714285713, + "eval_allNLI-dev_cosine_recall": 0.8323699421965318, + "eval_allNLI-dev_dot_accuracy": 0.681640625, + "eval_allNLI-dev_dot_accuracy_threshold": 508.9306945800781, + "eval_allNLI-dev_dot_ap": 0.4813780236759137, + "eval_allNLI-dev_dot_f1": 0.5576923076923077, + "eval_allNLI-dev_dot_f1_threshold": 381.4449462890625, + "eval_allNLI-dev_dot_precision": 0.41786743515850144, + "eval_allNLI-dev_dot_recall": 0.838150289017341, + "eval_allNLI-dev_euclidean_accuracy": 0.712890625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 9.198894500732422, + "eval_allNLI-dev_euclidean_ap": 0.5883566904425912, + "eval_allNLI-dev_euclidean_f1": 0.5922746781115881, + "eval_allNLI-dev_euclidean_f1_threshold": 14.1674165725708, + "eval_allNLI-dev_euclidean_precision": 0.4709897610921502, + "eval_allNLI-dev_euclidean_recall": 0.7976878612716763, + "eval_allNLI-dev_manhattan_accuracy": 0.716796875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 199.45140075683594, + "eval_allNLI-dev_manhattan_ap": 0.5950713823618554, + "eval_allNLI-dev_manhattan_f1": 0.6008771929824561, + "eval_allNLI-dev_manhattan_f1_threshold": 249.6264190673828, + "eval_allNLI-dev_manhattan_precision": 0.4840989399293286, + "eval_allNLI-dev_manhattan_recall": 0.791907514450867, + "eval_allNLI-dev_max_accuracy": 0.716796875, + "eval_allNLI-dev_max_accuracy_threshold": 508.9306945800781, + "eval_allNLI-dev_max_ap": 0.5950713823618554, + "eval_allNLI-dev_max_f1": 0.6008771929824561, + "eval_allNLI-dev_max_f1_threshold": 381.4449462890625, + "eval_allNLI-dev_max_precision": 0.4840989399293286, + "eval_allNLI-dev_max_recall": 0.838150289017341, + "eval_sequential_score": 0.7047500859185081, + "eval_sts-test_pearson_cosine": 0.7967003868457398, + "eval_sts-test_pearson_dot": 0.7108868024491241, + "eval_sts-test_pearson_euclidean": 0.8250765977440847, + "eval_sts-test_pearson_manhattan": 0.8248898142245689, + "eval_sts-test_pearson_max": 0.8250765977440847, + "eval_sts-test_spearman_cosine": 0.8297944914373061, + "eval_sts-test_spearman_dot": 0.6941270596670145, + "eval_sts-test_spearman_euclidean": 0.8229759207441439, + "eval_sts-test_spearman_manhattan": 0.8223873074030629, + "eval_sts-test_spearman_max": 0.8297944914373061, + "eval_vitaminc-pairs_loss": 3.898796558380127, + "eval_vitaminc-pairs_runtime": 5.4964, + "eval_vitaminc-pairs_samples_per_second": 23.288, + "eval_vitaminc-pairs_steps_per_second": 0.364, + "step": 560 + }, + { + "epoch": 0.43444530643910007, + "eval_negation-triplets_loss": 2.1589903831481934, + "eval_negation-triplets_runtime": 1.0811, + "eval_negation-triplets_samples_per_second": 118.399, + "eval_negation-triplets_steps_per_second": 1.85, + "step": 560 + }, + { + "epoch": 0.43444530643910007, + "eval_scitail-pairs-pos_loss": 0.2700319290161133, + "eval_scitail-pairs-pos_runtime": 1.2565, + "eval_scitail-pairs-pos_samples_per_second": 101.869, + "eval_scitail-pairs-pos_steps_per_second": 1.592, + "step": 560 + }, + { + "epoch": 0.43444530643910007, + "eval_scitail-pairs-qa_loss": 0.08734163641929626, + "eval_scitail-pairs-qa_runtime": 0.7785, + "eval_scitail-pairs-qa_samples_per_second": 164.425, + "eval_scitail-pairs-qa_steps_per_second": 2.569, + "step": 560 + }, + { + "epoch": 0.43444530643910007, + "eval_xsum-pairs_loss": 0.63297438621521, + "eval_xsum-pairs_runtime": 3.9247, + "eval_xsum-pairs_samples_per_second": 32.614, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 560 + }, + { + "epoch": 0.43444530643910007, + "eval_sciq_pairs_loss": 0.09908274561166763, + "eval_sciq_pairs_runtime": 5.9426, + "eval_sciq_pairs_samples_per_second": 21.539, + "eval_sciq_pairs_steps_per_second": 0.337, + "step": 560 + }, + { + "epoch": 0.43444530643910007, + "eval_qasc_pairs_loss": 1.0264776945114136, + "eval_qasc_pairs_runtime": 0.9175, + "eval_qasc_pairs_samples_per_second": 139.507, + "eval_qasc_pairs_steps_per_second": 2.18, + "step": 560 + }, + { + "epoch": 0.43444530643910007, + "eval_openbookqa_pairs_loss": 1.9953017234802246, + "eval_openbookqa_pairs_runtime": 0.8879, + "eval_openbookqa_pairs_samples_per_second": 144.155, + "eval_openbookqa_pairs_steps_per_second": 2.252, + "step": 560 + }, + { + "epoch": 0.43444530643910007, + "eval_msmarco_pairs_loss": 0.7671529650688171, + "eval_msmarco_pairs_runtime": 1.9882, + "eval_msmarco_pairs_samples_per_second": 64.38, + "eval_msmarco_pairs_steps_per_second": 1.006, + "step": 560 + }, + { + "epoch": 0.43444530643910007, + "eval_nq_pairs_loss": 1.5152760744094849, + "eval_nq_pairs_runtime": 3.3274, + "eval_nq_pairs_samples_per_second": 38.468, + "eval_nq_pairs_steps_per_second": 0.601, + "step": 560 + }, + { + "epoch": 0.43444530643910007, + "eval_trivia_pairs_loss": 1.2602167129516602, + "eval_trivia_pairs_runtime": 4.4739, + "eval_trivia_pairs_samples_per_second": 28.61, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 560 + }, + { + "epoch": 0.43444530643910007, + "eval_gooaq_pairs_loss": 1.0592139959335327, + "eval_gooaq_pairs_runtime": 1.4694, + "eval_gooaq_pairs_samples_per_second": 87.108, + "eval_gooaq_pairs_steps_per_second": 1.361, + "step": 560 + }, + { + "epoch": 0.43444530643910007, + "eval_paws-pos_loss": 0.05214548483490944, + "eval_paws-pos_runtime": 1.0328, + "eval_paws-pos_samples_per_second": 123.929, + "eval_paws-pos_steps_per_second": 1.936, + "step": 560 + }, + { + "epoch": 0.43444530643910007, + "eval_global_dataset_loss": 0.6379899978637695, + "eval_global_dataset_runtime": 19.3994, + "eval_global_dataset_samples_per_second": 21.444, + "eval_global_dataset_steps_per_second": 0.361, + "step": 560 + }, + { + "epoch": 0.4352211016291699, + "grad_norm": 8.060587882995605, + "learning_rate": 2.2949471210340773e-05, + "loss": 0.5159, + "step": 561 + }, + { + "epoch": 0.4359968968192397, + "grad_norm": 12.575251579284668, + "learning_rate": 2.299059929494712e-05, + "loss": 0.7526, + "step": 562 + }, + { + "epoch": 0.4367726920093095, + "grad_norm": 1.629343867301941, + "learning_rate": 2.3031727379553464e-05, + "loss": 0.1587, + "step": 563 + }, + { + "epoch": 0.4375484871993794, + "grad_norm": 11.41546630859375, + "learning_rate": 2.307285546415981e-05, + "loss": 0.5435, + "step": 564 + }, + { + "epoch": 0.4383242823894492, + "grad_norm": 1.7364354133605957, + "learning_rate": 2.3113983548766158e-05, + "loss": 0.0538, + "step": 565 + }, + { + "epoch": 0.439100077579519, + "grad_norm": 12.36942195892334, + "learning_rate": 2.31551116333725e-05, + "loss": 0.8029, + "step": 566 + }, + { + "epoch": 0.43987587276958884, + "grad_norm": 10.771207809448242, + "learning_rate": 2.3196239717978848e-05, + "loss": 0.5922, + "step": 567 + }, + { + "epoch": 0.44065166795965865, + "grad_norm": 11.95109748840332, + "learning_rate": 2.323736780258519e-05, + "loss": 0.758, + "step": 568 + }, + { + "epoch": 0.44142746314972847, + "grad_norm": 12.328636169433594, + "learning_rate": 2.3278495887191535e-05, + "loss": 1.0187, + "step": 569 + }, + { + "epoch": 0.4422032583397983, + "grad_norm": 19.241289138793945, + "learning_rate": 2.3319623971797884e-05, + "loss": 1.6983, + "step": 570 + }, + { + "epoch": 0.4429790535298681, + "grad_norm": 5.179631233215332, + "learning_rate": 2.3360752056404226e-05, + "loss": 0.1848, + "step": 571 + }, + { + "epoch": 0.4437548487199379, + "grad_norm": 5.7642974853515625, + "learning_rate": 2.3401880141010575e-05, + "loss": 0.2219, + "step": 572 + }, + { + "epoch": 0.44453064391000774, + "grad_norm": 14.506200790405273, + "learning_rate": 2.344300822561692e-05, + "loss": 1.3927, + "step": 573 + }, + { + "epoch": 0.44530643910007756, + "grad_norm": 19.637861251831055, + "learning_rate": 2.3484136310223265e-05, + "loss": 1.8411, + "step": 574 + }, + { + "epoch": 0.44608223429014743, + "grad_norm": 11.264130592346191, + "learning_rate": 2.352526439482961e-05, + "loss": 1.2575, + "step": 575 + }, + { + "epoch": 0.44685802948021724, + "grad_norm": 9.943875312805176, + "learning_rate": 2.3566392479435956e-05, + "loss": 0.4613, + "step": 576 + }, + { + "epoch": 0.44763382467028706, + "grad_norm": 18.227487564086914, + "learning_rate": 2.36075205640423e-05, + "loss": 1.4468, + "step": 577 + }, + { + "epoch": 0.4484096198603569, + "grad_norm": 7.544916152954102, + "learning_rate": 2.3648648648648646e-05, + "loss": 0.2955, + "step": 578 + }, + { + "epoch": 0.4491854150504267, + "grad_norm": 20.145084381103516, + "learning_rate": 2.3689776733254995e-05, + "loss": 1.9513, + "step": 579 + }, + { + "epoch": 0.4499612102404965, + "grad_norm": 10.472765922546387, + "learning_rate": 2.3730904817861337e-05, + "loss": 0.6613, + "step": 580 + }, + { + "epoch": 0.4499612102404965, + "eval_Qnli-dev_cosine_accuracy": 0.671875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8580237627029419, + "eval_Qnli-dev_cosine_ap": 0.6906106027616309, + "eval_Qnli-dev_cosine_f1": 0.6645056726094004, + "eval_Qnli-dev_cosine_f1_threshold": 0.7568535208702087, + "eval_Qnli-dev_cosine_precision": 0.5380577427821522, + "eval_Qnli-dev_cosine_recall": 0.8686440677966102, + "eval_Qnli-dev_dot_accuracy": 0.6171875, + "eval_Qnli-dev_dot_accuracy_threshold": 424.6160888671875, + "eval_Qnli-dev_dot_ap": 0.5877253516691959, + "eval_Qnli-dev_dot_f1": 0.6535552193645991, + "eval_Qnli-dev_dot_f1_threshold": 366.8108825683594, + "eval_Qnli-dev_dot_precision": 0.508235294117647, + "eval_Qnli-dev_dot_recall": 0.9152542372881356, + "eval_Qnli-dev_euclidean_accuracy": 0.681640625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.068756103515625, + "eval_Qnli-dev_euclidean_ap": 0.6926701107983333, + "eval_Qnli-dev_euclidean_f1": 0.6644407345575959, + "eval_Qnli-dev_euclidean_f1_threshold": 15.61347770690918, + "eval_Qnli-dev_euclidean_precision": 0.5482093663911846, + "eval_Qnli-dev_euclidean_recall": 0.8432203389830508, + "eval_Qnli-dev_manhattan_accuracy": 0.671875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 234.57846069335938, + "eval_Qnli-dev_manhattan_ap": 0.6898931246400537, + "eval_Qnli-dev_manhattan_f1": 0.6666666666666666, + "eval_Qnli-dev_manhattan_f1_threshold": 300.1756286621094, + "eval_Qnli-dev_manhattan_precision": 0.5215311004784688, + "eval_Qnli-dev_manhattan_recall": 0.923728813559322, + "eval_Qnli-dev_max_accuracy": 0.681640625, + "eval_Qnli-dev_max_accuracy_threshold": 424.6160888671875, + "eval_Qnli-dev_max_ap": 0.6926701107983333, + "eval_Qnli-dev_max_f1": 0.6666666666666666, + "eval_Qnli-dev_max_f1_threshold": 366.8108825683594, + "eval_Qnli-dev_max_precision": 0.5482093663911846, + "eval_Qnli-dev_max_recall": 0.923728813559322, + "eval_allNLI-dev_cosine_accuracy": 0.70703125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.910841166973114, + "eval_allNLI-dev_cosine_ap": 0.5792769511418232, + "eval_allNLI-dev_cosine_f1": 0.6000000000000001, + "eval_allNLI-dev_cosine_f1_threshold": 0.8421999216079712, + "eval_allNLI-dev_cosine_precision": 0.5286343612334802, + "eval_allNLI-dev_cosine_recall": 0.6936416184971098, + "eval_allNLI-dev_dot_accuracy": 0.669921875, + "eval_allNLI-dev_dot_accuracy_threshold": 516.0306396484375, + "eval_allNLI-dev_dot_ap": 0.4544345636312002, + "eval_allNLI-dev_dot_f1": 0.5593869731800767, + "eval_allNLI-dev_dot_f1_threshold": 381.7574157714844, + "eval_allNLI-dev_dot_precision": 0.4183381088825215, + "eval_allNLI-dev_dot_recall": 0.8439306358381503, + "eval_allNLI-dev_euclidean_accuracy": 0.708984375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 9.458547592163086, + "eval_allNLI-dev_euclidean_ap": 0.5804084999214881, + "eval_allNLI-dev_euclidean_f1": 0.6040609137055837, + "eval_allNLI-dev_euclidean_f1_threshold": 12.48227310180664, + "eval_allNLI-dev_euclidean_precision": 0.5384615384615384, + "eval_allNLI-dev_euclidean_recall": 0.6878612716763006, + "eval_allNLI-dev_manhattan_accuracy": 0.71484375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 192.69842529296875, + "eval_allNLI-dev_manhattan_ap": 0.5873005763609229, + "eval_allNLI-dev_manhattan_f1": 0.6115288220551378, + "eval_allNLI-dev_manhattan_f1_threshold": 226.03591918945312, + "eval_allNLI-dev_manhattan_precision": 0.5398230088495575, + "eval_allNLI-dev_manhattan_recall": 0.7052023121387283, + "eval_allNLI-dev_max_accuracy": 0.71484375, + "eval_allNLI-dev_max_accuracy_threshold": 516.0306396484375, + "eval_allNLI-dev_max_ap": 0.5873005763609229, + "eval_allNLI-dev_max_f1": 0.6115288220551378, + "eval_allNLI-dev_max_f1_threshold": 381.7574157714844, + "eval_allNLI-dev_max_precision": 0.5398230088495575, + "eval_allNLI-dev_max_recall": 0.8439306358381503, + "eval_sequential_score": 0.6926701107983333, + "eval_sts-test_pearson_cosine": 0.7830476745769297, + "eval_sts-test_pearson_dot": 0.7134737787708518, + "eval_sts-test_pearson_euclidean": 0.8084054218256234, + "eval_sts-test_pearson_manhattan": 0.8085754477519613, + "eval_sts-test_pearson_max": 0.8085754477519613, + "eval_sts-test_spearman_cosine": 0.8163193496172189, + "eval_sts-test_spearman_dot": 0.7038809998486226, + "eval_sts-test_spearman_euclidean": 0.8066933085406675, + "eval_sts-test_spearman_manhattan": 0.8065405724855386, + "eval_sts-test_spearman_max": 0.8163193496172189, + "eval_vitaminc-pairs_loss": 3.749843120574951, + "eval_vitaminc-pairs_runtime": 5.5021, + "eval_vitaminc-pairs_samples_per_second": 23.264, + "eval_vitaminc-pairs_steps_per_second": 0.363, + "step": 580 + }, + { + "epoch": 0.4499612102404965, + "eval_negation-triplets_loss": 2.174208879470825, + "eval_negation-triplets_runtime": 1.0775, + "eval_negation-triplets_samples_per_second": 118.796, + "eval_negation-triplets_steps_per_second": 1.856, + "step": 580 + }, + { + "epoch": 0.4499612102404965, + "eval_scitail-pairs-pos_loss": 0.20740848779678345, + "eval_scitail-pairs-pos_runtime": 1.2653, + "eval_scitail-pairs-pos_samples_per_second": 101.162, + "eval_scitail-pairs-pos_steps_per_second": 1.581, + "step": 580 + }, + { + "epoch": 0.4499612102404965, + "eval_scitail-pairs-qa_loss": 0.041522521525621414, + "eval_scitail-pairs-qa_runtime": 0.7805, + "eval_scitail-pairs-qa_samples_per_second": 163.994, + "eval_scitail-pairs-qa_steps_per_second": 2.562, + "step": 580 + }, + { + "epoch": 0.4499612102404965, + "eval_xsum-pairs_loss": 0.6239547729492188, + "eval_xsum-pairs_runtime": 3.9039, + "eval_xsum-pairs_samples_per_second": 32.788, + "eval_xsum-pairs_steps_per_second": 0.512, + "step": 580 + }, + { + "epoch": 0.4499612102404965, + "eval_sciq_pairs_loss": 0.08843886107206345, + "eval_sciq_pairs_runtime": 5.9325, + "eval_sciq_pairs_samples_per_second": 21.576, + "eval_sciq_pairs_steps_per_second": 0.337, + "step": 580 + }, + { + "epoch": 0.4499612102404965, + "eval_qasc_pairs_loss": 0.8510118722915649, + "eval_qasc_pairs_runtime": 0.9184, + "eval_qasc_pairs_samples_per_second": 139.369, + "eval_qasc_pairs_steps_per_second": 2.178, + "step": 580 + }, + { + "epoch": 0.4499612102404965, + "eval_openbookqa_pairs_loss": 2.0407955646514893, + "eval_openbookqa_pairs_runtime": 0.8719, + "eval_openbookqa_pairs_samples_per_second": 146.801, + "eval_openbookqa_pairs_steps_per_second": 2.294, + "step": 580 + }, + { + "epoch": 0.4499612102404965, + "eval_msmarco_pairs_loss": 0.7645685076713562, + "eval_msmarco_pairs_runtime": 1.9805, + "eval_msmarco_pairs_samples_per_second": 64.631, + "eval_msmarco_pairs_steps_per_second": 1.01, + "step": 580 + }, + { + "epoch": 0.4499612102404965, + "eval_nq_pairs_loss": 1.5348799228668213, + "eval_nq_pairs_runtime": 3.3191, + "eval_nq_pairs_samples_per_second": 38.564, + "eval_nq_pairs_steps_per_second": 0.603, + "step": 580 + }, + { + "epoch": 0.4499612102404965, + "eval_trivia_pairs_loss": 1.1610352993011475, + "eval_trivia_pairs_runtime": 4.4651, + "eval_trivia_pairs_samples_per_second": 28.667, + "eval_trivia_pairs_steps_per_second": 0.448, + "step": 580 + }, + { + "epoch": 0.4499612102404965, + "eval_gooaq_pairs_loss": 1.169563889503479, + "eval_gooaq_pairs_runtime": 1.474, + "eval_gooaq_pairs_samples_per_second": 86.836, + "eval_gooaq_pairs_steps_per_second": 1.357, + "step": 580 + }, + { + "epoch": 0.4499612102404965, + "eval_paws-pos_loss": 0.04819422587752342, + "eval_paws-pos_runtime": 1.0244, + "eval_paws-pos_samples_per_second": 124.955, + "eval_paws-pos_steps_per_second": 1.952, + "step": 580 + }, + { + "epoch": 0.4499612102404965, + "eval_global_dataset_loss": 0.6053475737571716, + "eval_global_dataset_runtime": 19.4254, + "eval_global_dataset_samples_per_second": 21.415, + "eval_global_dataset_steps_per_second": 0.36, + "step": 580 + }, + { + "epoch": 0.45073700543056633, + "grad_norm": 11.696528434753418, + "learning_rate": 2.3772032902467682e-05, + "loss": 0.7514, + "step": 581 + }, + { + "epoch": 0.45151280062063615, + "grad_norm": 16.50416374206543, + "learning_rate": 2.381316098707403e-05, + "loss": 1.5422, + "step": 582 + }, + { + "epoch": 0.45228859581070596, + "grad_norm": 4.802879333496094, + "learning_rate": 2.3854289071680373e-05, + "loss": 0.1387, + "step": 583 + }, + { + "epoch": 0.4530643910007758, + "grad_norm": 11.374921798706055, + "learning_rate": 2.389541715628672e-05, + "loss": 0.8216, + "step": 584 + }, + { + "epoch": 0.4538401861908456, + "grad_norm": 9.237058639526367, + "learning_rate": 2.3936545240893063e-05, + "loss": 0.4173, + "step": 585 + }, + { + "epoch": 0.4546159813809154, + "grad_norm": 15.692505836486816, + "learning_rate": 2.397767332549941e-05, + "loss": 1.8684, + "step": 586 + }, + { + "epoch": 0.4553917765709853, + "grad_norm": 9.588768005371094, + "learning_rate": 2.4018801410105757e-05, + "loss": 0.6013, + "step": 587 + }, + { + "epoch": 0.4561675717610551, + "grad_norm": 9.542437553405762, + "learning_rate": 2.40599294947121e-05, + "loss": 0.6788, + "step": 588 + }, + { + "epoch": 0.4569433669511249, + "grad_norm": 17.467397689819336, + "learning_rate": 2.4101057579318448e-05, + "loss": 1.2349, + "step": 589 + }, + { + "epoch": 0.45771916214119474, + "grad_norm": 23.1237850189209, + "learning_rate": 2.4142185663924793e-05, + "loss": 3.4482, + "step": 590 + }, + { + "epoch": 0.45849495733126455, + "grad_norm": 14.146553039550781, + "learning_rate": 2.418331374853114e-05, + "loss": 1.3932, + "step": 591 + }, + { + "epoch": 0.45927075252133437, + "grad_norm": 5.95766544342041, + "learning_rate": 2.4224441833137484e-05, + "loss": 0.2497, + "step": 592 + }, + { + "epoch": 0.4600465477114042, + "grad_norm": 18.849685668945312, + "learning_rate": 2.4265569917743826e-05, + "loss": 1.9145, + "step": 593 + }, + { + "epoch": 0.460822342901474, + "grad_norm": 10.007401466369629, + "learning_rate": 2.4306698002350174e-05, + "loss": 0.8326, + "step": 594 + }, + { + "epoch": 0.4615981380915438, + "grad_norm": 9.367300987243652, + "learning_rate": 2.434782608695652e-05, + "loss": 0.6774, + "step": 595 + }, + { + "epoch": 0.46237393328161364, + "grad_norm": 10.141812324523926, + "learning_rate": 2.4388954171562868e-05, + "loss": 1.1696, + "step": 596 + }, + { + "epoch": 0.46314972847168345, + "grad_norm": 7.813787937164307, + "learning_rate": 2.443008225616921e-05, + "loss": 0.4832, + "step": 597 + }, + { + "epoch": 0.46392552366175327, + "grad_norm": 11.965947151184082, + "learning_rate": 2.4471210340775555e-05, + "loss": 1.7278, + "step": 598 + }, + { + "epoch": 0.46470131885182314, + "grad_norm": 21.20833396911621, + "learning_rate": 2.45123384253819e-05, + "loss": 3.4533, + "step": 599 + }, + { + "epoch": 0.46547711404189296, + "grad_norm": 6.598328113555908, + "learning_rate": 2.4553466509988246e-05, + "loss": 0.3008, + "step": 600 + }, + { + "epoch": 0.46547711404189296, + "eval_Qnli-dev_cosine_accuracy": 0.669921875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8546062707901001, + "eval_Qnli-dev_cosine_ap": 0.6845613555308852, + "eval_Qnli-dev_cosine_f1": 0.6733668341708542, + "eval_Qnli-dev_cosine_f1_threshold": 0.7931475639343262, + "eval_Qnli-dev_cosine_precision": 0.556786703601108, + "eval_Qnli-dev_cosine_recall": 0.8516949152542372, + "eval_Qnli-dev_dot_accuracy": 0.623046875, + "eval_Qnli-dev_dot_accuracy_threshold": 475.34100341796875, + "eval_Qnli-dev_dot_ap": 0.589675055189002, + "eval_Qnli-dev_dot_f1": 0.6545454545454547, + "eval_Qnli-dev_dot_f1_threshold": 418.89013671875, + "eval_Qnli-dev_dot_precision": 0.5365853658536586, + "eval_Qnli-dev_dot_recall": 0.8389830508474576, + "eval_Qnli-dev_euclidean_accuracy": 0.669921875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.711584091186523, + "eval_Qnli-dev_euclidean_ap": 0.68823807349383, + "eval_Qnli-dev_euclidean_f1": 0.6677165354330709, + "eval_Qnli-dev_euclidean_f1_threshold": 16.21005630493164, + "eval_Qnli-dev_euclidean_precision": 0.531328320802005, + "eval_Qnli-dev_euclidean_recall": 0.8983050847457628, + "eval_Qnli-dev_manhattan_accuracy": 0.66015625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 219.9610137939453, + "eval_Qnli-dev_manhattan_ap": 0.6910623883856871, + "eval_Qnli-dev_manhattan_f1": 0.672811059907834, + "eval_Qnli-dev_manhattan_f1_threshold": 291.64935302734375, + "eval_Qnli-dev_manhattan_precision": 0.5277108433734939, + "eval_Qnli-dev_manhattan_recall": 0.9279661016949152, + "eval_Qnli-dev_max_accuracy": 0.669921875, + "eval_Qnli-dev_max_accuracy_threshold": 475.34100341796875, + "eval_Qnli-dev_max_ap": 0.6910623883856871, + "eval_Qnli-dev_max_f1": 0.6733668341708542, + "eval_Qnli-dev_max_f1_threshold": 418.89013671875, + "eval_Qnli-dev_max_precision": 0.556786703601108, + "eval_Qnli-dev_max_recall": 0.9279661016949152, + "eval_allNLI-dev_cosine_accuracy": 0.70703125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9173166751861572, + "eval_allNLI-dev_cosine_ap": 0.5882976992957594, + "eval_allNLI-dev_cosine_f1": 0.6072289156626506, + "eval_allNLI-dev_cosine_f1_threshold": 0.8463546633720398, + "eval_allNLI-dev_cosine_precision": 0.5206611570247934, + "eval_allNLI-dev_cosine_recall": 0.7283236994219653, + "eval_allNLI-dev_dot_accuracy": 0.67578125, + "eval_allNLI-dev_dot_accuracy_threshold": 573.2156982421875, + "eval_allNLI-dev_dot_ap": 0.4861132554738923, + "eval_allNLI-dev_dot_f1": 0.5644171779141105, + "eval_allNLI-dev_dot_f1_threshold": 412.581298828125, + "eval_allNLI-dev_dot_precision": 0.43670886075949367, + "eval_allNLI-dev_dot_recall": 0.7976878612716763, + "eval_allNLI-dev_euclidean_accuracy": 0.70703125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 9.563434600830078, + "eval_allNLI-dev_euclidean_ap": 0.5892715220992573, + "eval_allNLI-dev_euclidean_f1": 0.6071428571428571, + "eval_allNLI-dev_euclidean_f1_threshold": 13.366493225097656, + "eval_allNLI-dev_euclidean_precision": 0.49454545454545457, + "eval_allNLI-dev_euclidean_recall": 0.7861271676300579, + "eval_allNLI-dev_manhattan_accuracy": 0.712890625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 211.37059020996094, + "eval_allNLI-dev_manhattan_ap": 0.5886668227955144, + "eval_allNLI-dev_manhattan_f1": 0.6112266112266113, + "eval_allNLI-dev_manhattan_f1_threshold": 250.82669067382812, + "eval_allNLI-dev_manhattan_precision": 0.4772727272727273, + "eval_allNLI-dev_manhattan_recall": 0.8497109826589595, + "eval_allNLI-dev_max_accuracy": 0.712890625, + "eval_allNLI-dev_max_accuracy_threshold": 573.2156982421875, + "eval_allNLI-dev_max_ap": 0.5892715220992573, + "eval_allNLI-dev_max_f1": 0.6112266112266113, + "eval_allNLI-dev_max_f1_threshold": 412.581298828125, + "eval_allNLI-dev_max_precision": 0.5206611570247934, + "eval_allNLI-dev_max_recall": 0.8497109826589595, + "eval_sequential_score": 0.6910623883856871, + "eval_sts-test_pearson_cosine": 0.7963356679643628, + "eval_sts-test_pearson_dot": 0.7140726972439533, + "eval_sts-test_pearson_euclidean": 0.8266048092120172, + "eval_sts-test_pearson_manhattan": 0.8247628412570818, + "eval_sts-test_pearson_max": 0.8266048092120172, + "eval_sts-test_spearman_cosine": 0.8302385134528314, + "eval_sts-test_spearman_dot": 0.7051283965086464, + "eval_sts-test_spearman_euclidean": 0.8232894033301907, + "eval_sts-test_spearman_manhattan": 0.8211249190704728, + "eval_sts-test_spearman_max": 0.8302385134528314, + "eval_vitaminc-pairs_loss": 3.8651621341705322, + "eval_vitaminc-pairs_runtime": 5.4814, + "eval_vitaminc-pairs_samples_per_second": 23.352, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 600 + }, + { + "epoch": 0.46547711404189296, + "eval_negation-triplets_loss": 2.1524651050567627, + "eval_negation-triplets_runtime": 1.0681, + "eval_negation-triplets_samples_per_second": 119.841, + "eval_negation-triplets_steps_per_second": 1.873, + "step": 600 + }, + { + "epoch": 0.46547711404189296, + "eval_scitail-pairs-pos_loss": 0.21253004670143127, + "eval_scitail-pairs-pos_runtime": 1.3023, + "eval_scitail-pairs-pos_samples_per_second": 98.285, + "eval_scitail-pairs-pos_steps_per_second": 1.536, + "step": 600 + }, + { + "epoch": 0.46547711404189296, + "eval_scitail-pairs-qa_loss": 0.06775914877653122, + "eval_scitail-pairs-qa_runtime": 0.7769, + "eval_scitail-pairs-qa_samples_per_second": 164.753, + "eval_scitail-pairs-qa_steps_per_second": 2.574, + "step": 600 + }, + { + "epoch": 0.46547711404189296, + "eval_xsum-pairs_loss": 0.6388774514198303, + "eval_xsum-pairs_runtime": 3.9218, + "eval_xsum-pairs_samples_per_second": 32.638, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 600 + }, + { + "epoch": 0.46547711404189296, + "eval_sciq_pairs_loss": 0.09762287139892578, + "eval_sciq_pairs_runtime": 5.899, + "eval_sciq_pairs_samples_per_second": 21.699, + "eval_sciq_pairs_steps_per_second": 0.339, + "step": 600 + }, + { + "epoch": 0.46547711404189296, + "eval_qasc_pairs_loss": 0.8573689460754395, + "eval_qasc_pairs_runtime": 0.9019, + "eval_qasc_pairs_samples_per_second": 141.92, + "eval_qasc_pairs_steps_per_second": 2.218, + "step": 600 + }, + { + "epoch": 0.46547711404189296, + "eval_openbookqa_pairs_loss": 1.850547194480896, + "eval_openbookqa_pairs_runtime": 0.8564, + "eval_openbookqa_pairs_samples_per_second": 149.47, + "eval_openbookqa_pairs_steps_per_second": 2.335, + "step": 600 + }, + { + "epoch": 0.46547711404189296, + "eval_msmarco_pairs_loss": 0.747097909450531, + "eval_msmarco_pairs_runtime": 1.9665, + "eval_msmarco_pairs_samples_per_second": 65.089, + "eval_msmarco_pairs_steps_per_second": 1.017, + "step": 600 + }, + { + "epoch": 0.46547711404189296, + "eval_nq_pairs_loss": 1.386681318283081, + "eval_nq_pairs_runtime": 3.3195, + "eval_nq_pairs_samples_per_second": 38.56, + "eval_nq_pairs_steps_per_second": 0.603, + "step": 600 + }, + { + "epoch": 0.46547711404189296, + "eval_trivia_pairs_loss": 1.2255401611328125, + "eval_trivia_pairs_runtime": 4.4764, + "eval_trivia_pairs_samples_per_second": 28.595, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 600 + }, + { + "epoch": 0.46547711404189296, + "eval_gooaq_pairs_loss": 1.087661623954773, + "eval_gooaq_pairs_runtime": 1.4674, + "eval_gooaq_pairs_samples_per_second": 87.229, + "eval_gooaq_pairs_steps_per_second": 1.363, + "step": 600 + }, + { + "epoch": 0.46547711404189296, + "eval_paws-pos_loss": 0.05389420688152313, + "eval_paws-pos_runtime": 1.0371, + "eval_paws-pos_samples_per_second": 123.421, + "eval_paws-pos_steps_per_second": 1.928, + "step": 600 + }, + { + "epoch": 0.46547711404189296, + "eval_global_dataset_loss": 0.6292778849601746, + "eval_global_dataset_runtime": 19.4321, + "eval_global_dataset_samples_per_second": 21.408, + "eval_global_dataset_steps_per_second": 0.36, + "step": 600 + }, + { + "epoch": 0.4662529092319628, + "grad_norm": 13.785943031311035, + "learning_rate": 2.4594594594594595e-05, + "loss": 1.4827, + "step": 601 + }, + { + "epoch": 0.4670287044220326, + "grad_norm": 9.836690902709961, + "learning_rate": 2.4635722679200937e-05, + "loss": 0.632, + "step": 602 + }, + { + "epoch": 0.4678044996121024, + "grad_norm": 12.741625785827637, + "learning_rate": 2.4676850763807282e-05, + "loss": 1.0095, + "step": 603 + }, + { + "epoch": 0.4685802948021722, + "grad_norm": 9.55395221710205, + "learning_rate": 2.471797884841363e-05, + "loss": 0.5745, + "step": 604 + }, + { + "epoch": 0.46935608999224204, + "grad_norm": 9.53487777709961, + "learning_rate": 2.4759106933019972e-05, + "loss": 0.6422, + "step": 605 + }, + { + "epoch": 0.47013188518231186, + "grad_norm": 11.145259857177734, + "learning_rate": 2.480023501762632e-05, + "loss": 0.7526, + "step": 606 + }, + { + "epoch": 0.4709076803723817, + "grad_norm": 15.308662414550781, + "learning_rate": 2.4841363102232666e-05, + "loss": 1.5273, + "step": 607 + }, + { + "epoch": 0.4716834755624515, + "grad_norm": 1.5294910669326782, + "learning_rate": 2.4882491186839012e-05, + "loss": 0.0704, + "step": 608 + }, + { + "epoch": 0.4724592707525213, + "grad_norm": 10.075404167175293, + "learning_rate": 2.4923619271445357e-05, + "loss": 0.6261, + "step": 609 + }, + { + "epoch": 0.4732350659425912, + "grad_norm": 6.055626392364502, + "learning_rate": 2.49647473560517e-05, + "loss": 0.2799, + "step": 610 + }, + { + "epoch": 0.474010861132661, + "grad_norm": 12.325764656066895, + "learning_rate": 2.5005875440658048e-05, + "loss": 1.5529, + "step": 611 + }, + { + "epoch": 0.4747866563227308, + "grad_norm": 9.178507804870605, + "learning_rate": 2.5047003525264393e-05, + "loss": 0.5666, + "step": 612 + }, + { + "epoch": 0.47556245151280063, + "grad_norm": 8.433951377868652, + "learning_rate": 2.5088131609870738e-05, + "loss": 0.4426, + "step": 613 + }, + { + "epoch": 0.47633824670287045, + "grad_norm": 10.24896240234375, + "learning_rate": 2.5129259694477083e-05, + "loss": 0.9742, + "step": 614 + }, + { + "epoch": 0.47711404189294027, + "grad_norm": 9.197361946105957, + "learning_rate": 2.517038777908343e-05, + "loss": 0.4615, + "step": 615 + }, + { + "epoch": 0.4778898370830101, + "grad_norm": 9.704472541809082, + "learning_rate": 2.5211515863689774e-05, + "loss": 0.4863, + "step": 616 + }, + { + "epoch": 0.4786656322730799, + "grad_norm": 18.687057495117188, + "learning_rate": 2.525264394829612e-05, + "loss": 1.4415, + "step": 617 + }, + { + "epoch": 0.4794414274631497, + "grad_norm": 13.103899002075195, + "learning_rate": 2.5293772032902468e-05, + "loss": 0.866, + "step": 618 + }, + { + "epoch": 0.48021722265321953, + "grad_norm": 13.415831565856934, + "learning_rate": 2.533490011750881e-05, + "loss": 1.7064, + "step": 619 + }, + { + "epoch": 0.48099301784328935, + "grad_norm": 12.0526762008667, + "learning_rate": 2.5376028202115155e-05, + "loss": 1.5001, + "step": 620 + }, + { + "epoch": 0.48099301784328935, + "eval_Qnli-dev_cosine_accuracy": 0.662109375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8541783094406128, + "eval_Qnli-dev_cosine_ap": 0.6909571670552859, + "eval_Qnli-dev_cosine_f1": 0.6708860759493672, + "eval_Qnli-dev_cosine_f1_threshold": 0.7411938905715942, + "eval_Qnli-dev_cosine_precision": 0.5353535353535354, + "eval_Qnli-dev_cosine_recall": 0.8983050847457628, + "eval_Qnli-dev_dot_accuracy": 0.623046875, + "eval_Qnli-dev_dot_accuracy_threshold": 438.66668701171875, + "eval_Qnli-dev_dot_ap": 0.5838936342619079, + "eval_Qnli-dev_dot_f1": 0.660377358490566, + "eval_Qnli-dev_dot_f1_threshold": 371.63031005859375, + "eval_Qnli-dev_dot_precision": 0.525, + "eval_Qnli-dev_dot_recall": 0.8898305084745762, + "eval_Qnli-dev_euclidean_accuracy": 0.66015625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.478965759277344, + "eval_Qnli-dev_euclidean_ap": 0.6923881432370431, + "eval_Qnli-dev_euclidean_f1": 0.6702702702702703, + "eval_Qnli-dev_euclidean_f1_threshold": 14.774309158325195, + "eval_Qnli-dev_euclidean_precision": 0.5830721003134797, + "eval_Qnli-dev_euclidean_recall": 0.788135593220339, + "eval_Qnli-dev_manhattan_accuracy": 0.66015625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 224.37429809570312, + "eval_Qnli-dev_manhattan_ap": 0.6961468323934411, + "eval_Qnli-dev_manhattan_f1": 0.6759581881533101, + "eval_Qnli-dev_manhattan_f1_threshold": 269.2730712890625, + "eval_Qnli-dev_manhattan_precision": 0.5739644970414202, + "eval_Qnli-dev_manhattan_recall": 0.8220338983050848, + "eval_Qnli-dev_max_accuracy": 0.662109375, + "eval_Qnli-dev_max_accuracy_threshold": 438.66668701171875, + "eval_Qnli-dev_max_ap": 0.6961468323934411, + "eval_Qnli-dev_max_f1": 0.6759581881533101, + "eval_Qnli-dev_max_f1_threshold": 371.63031005859375, + "eval_Qnli-dev_max_precision": 0.5830721003134797, + "eval_Qnli-dev_max_recall": 0.8983050847457628, + "eval_allNLI-dev_cosine_accuracy": 0.712890625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8691669702529907, + "eval_allNLI-dev_cosine_ap": 0.5879530656854549, + "eval_allNLI-dev_cosine_f1": 0.6106194690265486, + "eval_allNLI-dev_cosine_f1_threshold": 0.8064013123512268, + "eval_allNLI-dev_cosine_precision": 0.4946236559139785, + "eval_allNLI-dev_cosine_recall": 0.7976878612716763, + "eval_allNLI-dev_dot_accuracy": 0.677734375, + "eval_allNLI-dev_dot_accuracy_threshold": 462.31378173828125, + "eval_allNLI-dev_dot_ap": 0.47263393740678317, + "eval_allNLI-dev_dot_f1": 0.5720164609053497, + "eval_allNLI-dev_dot_f1_threshold": 382.060546875, + "eval_allNLI-dev_dot_precision": 0.4440894568690096, + "eval_allNLI-dev_dot_recall": 0.8034682080924855, + "eval_allNLI-dev_euclidean_accuracy": 0.712890625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.587467193603516, + "eval_allNLI-dev_euclidean_ap": 0.5864231085900702, + "eval_allNLI-dev_euclidean_f1": 0.6073752711496746, + "eval_allNLI-dev_euclidean_f1_threshold": 14.026927947998047, + "eval_allNLI-dev_euclidean_precision": 0.4861111111111111, + "eval_allNLI-dev_euclidean_recall": 0.8092485549132948, + "eval_allNLI-dev_manhattan_accuracy": 0.716796875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 211.80812072753906, + "eval_allNLI-dev_manhattan_ap": 0.5904776044396406, + "eval_allNLI-dev_manhattan_f1": 0.6153846153846154, + "eval_allNLI-dev_manhattan_f1_threshold": 238.34698486328125, + "eval_allNLI-dev_manhattan_precision": 0.5267489711934157, + "eval_allNLI-dev_manhattan_recall": 0.7398843930635838, + "eval_allNLI-dev_max_accuracy": 0.716796875, + "eval_allNLI-dev_max_accuracy_threshold": 462.31378173828125, + "eval_allNLI-dev_max_ap": 0.5904776044396406, + "eval_allNLI-dev_max_f1": 0.6153846153846154, + "eval_allNLI-dev_max_f1_threshold": 382.060546875, + "eval_allNLI-dev_max_precision": 0.5267489711934157, + "eval_allNLI-dev_max_recall": 0.8092485549132948, + "eval_sequential_score": 0.6961468323934411, + "eval_sts-test_pearson_cosine": 0.807324798824198, + "eval_sts-test_pearson_dot": 0.7772455289329745, + "eval_sts-test_pearson_euclidean": 0.8298443537503775, + "eval_sts-test_pearson_manhattan": 0.8262557698820469, + "eval_sts-test_pearson_max": 0.8298443537503775, + "eval_sts-test_spearman_cosine": 0.8349379776653264, + "eval_sts-test_spearman_dot": 0.7772892684871415, + "eval_sts-test_spearman_euclidean": 0.8241774921125029, + "eval_sts-test_spearman_manhattan": 0.8208295607452898, + "eval_sts-test_spearman_max": 0.8349379776653264, + "eval_vitaminc-pairs_loss": 4.181798458099365, + "eval_vitaminc-pairs_runtime": 5.4888, + "eval_vitaminc-pairs_samples_per_second": 23.32, + "eval_vitaminc-pairs_steps_per_second": 0.364, + "step": 620 + }, + { + "epoch": 0.48099301784328935, + "eval_negation-triplets_loss": 2.0779168605804443, + "eval_negation-triplets_runtime": 1.0714, + "eval_negation-triplets_samples_per_second": 119.469, + "eval_negation-triplets_steps_per_second": 1.867, + "step": 620 + }, + { + "epoch": 0.48099301784328935, + "eval_scitail-pairs-pos_loss": 0.24158324301242828, + "eval_scitail-pairs-pos_runtime": 1.2667, + "eval_scitail-pairs-pos_samples_per_second": 101.053, + "eval_scitail-pairs-pos_steps_per_second": 1.579, + "step": 620 + }, + { + "epoch": 0.48099301784328935, + "eval_scitail-pairs-qa_loss": 0.04018586128950119, + "eval_scitail-pairs-qa_runtime": 0.7708, + "eval_scitail-pairs-qa_samples_per_second": 166.056, + "eval_scitail-pairs-qa_steps_per_second": 2.595, + "step": 620 + }, + { + "epoch": 0.48099301784328935, + "eval_xsum-pairs_loss": 0.6078460216522217, + "eval_xsum-pairs_runtime": 3.9134, + "eval_xsum-pairs_samples_per_second": 32.708, + "eval_xsum-pairs_steps_per_second": 0.511, + "step": 620 + }, + { + "epoch": 0.48099301784328935, + "eval_sciq_pairs_loss": 0.08504968881607056, + "eval_sciq_pairs_runtime": 5.9307, + "eval_sciq_pairs_samples_per_second": 21.583, + "eval_sciq_pairs_steps_per_second": 0.337, + "step": 620 + }, + { + "epoch": 0.48099301784328935, + "eval_qasc_pairs_loss": 0.8349754214286804, + "eval_qasc_pairs_runtime": 0.9008, + "eval_qasc_pairs_samples_per_second": 142.097, + "eval_qasc_pairs_steps_per_second": 2.22, + "step": 620 + }, + { + "epoch": 0.48099301784328935, + "eval_openbookqa_pairs_loss": 1.880592703819275, + "eval_openbookqa_pairs_runtime": 0.8689, + "eval_openbookqa_pairs_samples_per_second": 147.316, + "eval_openbookqa_pairs_steps_per_second": 2.302, + "step": 620 + }, + { + "epoch": 0.48099301784328935, + "eval_msmarco_pairs_loss": 0.7329286932945251, + "eval_msmarco_pairs_runtime": 1.9747, + "eval_msmarco_pairs_samples_per_second": 64.821, + "eval_msmarco_pairs_steps_per_second": 1.013, + "step": 620 + }, + { + "epoch": 0.48099301784328935, + "eval_nq_pairs_loss": 1.375463843345642, + "eval_nq_pairs_runtime": 3.338, + "eval_nq_pairs_samples_per_second": 38.346, + "eval_nq_pairs_steps_per_second": 0.599, + "step": 620 + }, + { + "epoch": 0.48099301784328935, + "eval_trivia_pairs_loss": 1.2347986698150635, + "eval_trivia_pairs_runtime": 4.4899, + "eval_trivia_pairs_samples_per_second": 28.508, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 620 + }, + { + "epoch": 0.48099301784328935, + "eval_gooaq_pairs_loss": 1.143622875213623, + "eval_gooaq_pairs_runtime": 1.4714, + "eval_gooaq_pairs_samples_per_second": 86.991, + "eval_gooaq_pairs_steps_per_second": 1.359, + "step": 620 + }, + { + "epoch": 0.48099301784328935, + "eval_paws-pos_loss": 0.045590754598379135, + "eval_paws-pos_runtime": 1.0358, + "eval_paws-pos_samples_per_second": 123.573, + "eval_paws-pos_steps_per_second": 1.931, + "step": 620 + }, + { + "epoch": 0.48099301784328935, + "eval_global_dataset_loss": 0.5951407551765442, + "eval_global_dataset_runtime": 19.4168, + "eval_global_dataset_samples_per_second": 21.425, + "eval_global_dataset_steps_per_second": 0.361, + "step": 620 + }, + { + "epoch": 0.48176881303335917, + "grad_norm": 3.3584859371185303, + "learning_rate": 2.5417156286721504e-05, + "loss": 0.0943, + "step": 621 + }, + { + "epoch": 0.48254460822342904, + "grad_norm": 13.811856269836426, + "learning_rate": 2.5458284371327846e-05, + "loss": 0.5077, + "step": 622 + }, + { + "epoch": 0.48332040341349886, + "grad_norm": 9.970905303955078, + "learning_rate": 2.5499412455934194e-05, + "loss": 0.6746, + "step": 623 + }, + { + "epoch": 0.4840961986035687, + "grad_norm": 10.551204681396484, + "learning_rate": 2.5540540540540536e-05, + "loss": 0.6953, + "step": 624 + }, + { + "epoch": 0.4848719937936385, + "grad_norm": 3.4784350395202637, + "learning_rate": 2.5581668625146885e-05, + "loss": 0.0762, + "step": 625 + }, + { + "epoch": 0.4856477889837083, + "grad_norm": 8.014740943908691, + "learning_rate": 2.562279670975323e-05, + "loss": 0.4902, + "step": 626 + }, + { + "epoch": 0.4864235841737781, + "grad_norm": 8.231643676757812, + "learning_rate": 2.5663924794359572e-05, + "loss": 0.4359, + "step": 627 + }, + { + "epoch": 0.48719937936384794, + "grad_norm": 12.098679542541504, + "learning_rate": 2.570505287896592e-05, + "loss": 1.5008, + "step": 628 + }, + { + "epoch": 0.48797517455391776, + "grad_norm": 11.154753684997559, + "learning_rate": 2.5746180963572266e-05, + "loss": 0.901, + "step": 629 + }, + { + "epoch": 0.4887509697439876, + "grad_norm": 8.817254066467285, + "learning_rate": 2.578730904817861e-05, + "loss": 0.3591, + "step": 630 + }, + { + "epoch": 0.4895267649340574, + "grad_norm": 14.75892162322998, + "learning_rate": 2.5828437132784957e-05, + "loss": 1.3486, + "step": 631 + }, + { + "epoch": 0.4903025601241272, + "grad_norm": 10.355876922607422, + "learning_rate": 2.5869565217391302e-05, + "loss": 0.6322, + "step": 632 + }, + { + "epoch": 0.4910783553141971, + "grad_norm": 10.594958305358887, + "learning_rate": 2.5910693301997647e-05, + "loss": 0.6911, + "step": 633 + }, + { + "epoch": 0.4918541505042669, + "grad_norm": 9.579641342163086, + "learning_rate": 2.5951821386603993e-05, + "loss": 0.4946, + "step": 634 + }, + { + "epoch": 0.4926299456943367, + "grad_norm": 14.586888313293457, + "learning_rate": 2.599294947121034e-05, + "loss": 1.3486, + "step": 635 + }, + { + "epoch": 0.49340574088440653, + "grad_norm": 5.319215774536133, + "learning_rate": 2.6034077555816683e-05, + "loss": 0.1691, + "step": 636 + }, + { + "epoch": 0.49418153607447635, + "grad_norm": 3.732201099395752, + "learning_rate": 2.607520564042303e-05, + "loss": 0.1111, + "step": 637 + }, + { + "epoch": 0.49495733126454616, + "grad_norm": 2.3062543869018555, + "learning_rate": 2.6116333725029374e-05, + "loss": 0.0328, + "step": 638 + }, + { + "epoch": 0.495733126454616, + "grad_norm": 10.06052017211914, + "learning_rate": 2.615746180963572e-05, + "loss": 0.4712, + "step": 639 + }, + { + "epoch": 0.4965089216446858, + "grad_norm": 19.183080673217773, + "learning_rate": 2.6198589894242068e-05, + "loss": 1.479, + "step": 640 + }, + { + "epoch": 0.4965089216446858, + "eval_Qnli-dev_cosine_accuracy": 0.658203125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8601222038269043, + "eval_Qnli-dev_cosine_ap": 0.686720032091346, + "eval_Qnli-dev_cosine_f1": 0.6749226006191951, + "eval_Qnli-dev_cosine_f1_threshold": 0.6975479125976562, + "eval_Qnli-dev_cosine_precision": 0.5317073170731708, + "eval_Qnli-dev_cosine_recall": 0.923728813559322, + "eval_Qnli-dev_dot_accuracy": 0.626953125, + "eval_Qnli-dev_dot_accuracy_threshold": 407.0140380859375, + "eval_Qnli-dev_dot_ap": 0.6013138350063858, + "eval_Qnli-dev_dot_f1": 0.6617862371888725, + "eval_Qnli-dev_dot_f1_threshold": 313.9571838378906, + "eval_Qnli-dev_dot_precision": 0.5055928411633109, + "eval_Qnli-dev_dot_recall": 0.9576271186440678, + "eval_Qnli-dev_euclidean_accuracy": 0.6484375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.301708221435547, + "eval_Qnli-dev_euclidean_ap": 0.6848164247929951, + "eval_Qnli-dev_euclidean_f1": 0.6751592356687899, + "eval_Qnli-dev_euclidean_f1_threshold": 17.13784408569336, + "eval_Qnli-dev_euclidean_precision": 0.5408163265306123, + "eval_Qnli-dev_euclidean_recall": 0.8983050847457628, + "eval_Qnli-dev_manhattan_accuracy": 0.65625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 233.0150909423828, + "eval_Qnli-dev_manhattan_ap": 0.6881294548528416, + "eval_Qnli-dev_manhattan_f1": 0.68125, + "eval_Qnli-dev_manhattan_f1_threshold": 307.3375244140625, + "eval_Qnli-dev_manhattan_precision": 0.5396039603960396, + "eval_Qnli-dev_manhattan_recall": 0.923728813559322, + "eval_Qnli-dev_max_accuracy": 0.658203125, + "eval_Qnli-dev_max_accuracy_threshold": 407.0140380859375, + "eval_Qnli-dev_max_ap": 0.6881294548528416, + "eval_Qnli-dev_max_f1": 0.68125, + "eval_Qnli-dev_max_f1_threshold": 313.9571838378906, + "eval_Qnli-dev_max_precision": 0.5408163265306123, + "eval_Qnli-dev_max_recall": 0.9576271186440678, + "eval_allNLI-dev_cosine_accuracy": 0.712890625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.923399806022644, + "eval_allNLI-dev_cosine_ap": 0.5970543483716874, + "eval_allNLI-dev_cosine_f1": 0.6134020618556701, + "eval_allNLI-dev_cosine_f1_threshold": 0.8509269952774048, + "eval_allNLI-dev_cosine_precision": 0.5534883720930233, + "eval_allNLI-dev_cosine_recall": 0.6878612716763006, + "eval_allNLI-dev_dot_accuracy": 0.677734375, + "eval_allNLI-dev_dot_accuracy_threshold": 447.8410949707031, + "eval_allNLI-dev_dot_ap": 0.4805633326048575, + "eval_allNLI-dev_dot_f1": 0.5677179962894249, + "eval_allNLI-dev_dot_f1_threshold": 357.12249755859375, + "eval_allNLI-dev_dot_precision": 0.4180327868852459, + "eval_allNLI-dev_dot_recall": 0.884393063583815, + "eval_allNLI-dev_euclidean_accuracy": 0.712890625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 8.865427017211914, + "eval_allNLI-dev_euclidean_ap": 0.5969117820320243, + "eval_allNLI-dev_euclidean_f1": 0.6181818181818183, + "eval_allNLI-dev_euclidean_f1_threshold": 12.873517036437988, + "eval_allNLI-dev_euclidean_precision": 0.5093632958801498, + "eval_allNLI-dev_euclidean_recall": 0.7861271676300579, + "eval_allNLI-dev_manhattan_accuracy": 0.712890625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 195.62420654296875, + "eval_allNLI-dev_manhattan_ap": 0.6016030414147083, + "eval_allNLI-dev_manhattan_f1": 0.6238095238095238, + "eval_allNLI-dev_manhattan_f1_threshold": 224.73388671875, + "eval_allNLI-dev_manhattan_precision": 0.5303643724696356, + "eval_allNLI-dev_manhattan_recall": 0.7572254335260116, + "eval_allNLI-dev_max_accuracy": 0.712890625, + "eval_allNLI-dev_max_accuracy_threshold": 447.8410949707031, + "eval_allNLI-dev_max_ap": 0.6016030414147083, + "eval_allNLI-dev_max_f1": 0.6238095238095238, + "eval_allNLI-dev_max_f1_threshold": 357.12249755859375, + "eval_allNLI-dev_max_precision": 0.5534883720930233, + "eval_allNLI-dev_max_recall": 0.884393063583815, + "eval_sequential_score": 0.6881294548528416, + "eval_sts-test_pearson_cosine": 0.7934872507202126, + "eval_sts-test_pearson_dot": 0.7596638361218665, + "eval_sts-test_pearson_euclidean": 0.8149453649979157, + "eval_sts-test_pearson_manhattan": 0.8142733776143798, + "eval_sts-test_pearson_max": 0.8149453649979157, + "eval_sts-test_spearman_cosine": 0.8221940714563749, + "eval_sts-test_spearman_dot": 0.7498613098828755, + "eval_sts-test_spearman_euclidean": 0.8123915314048957, + "eval_sts-test_spearman_manhattan": 0.8101784472753069, + "eval_sts-test_spearman_max": 0.8221940714563749, + "eval_vitaminc-pairs_loss": 3.602985143661499, + "eval_vitaminc-pairs_runtime": 5.5062, + "eval_vitaminc-pairs_samples_per_second": 23.247, + "eval_vitaminc-pairs_steps_per_second": 0.363, + "step": 640 + }, + { + "epoch": 0.4965089216446858, + "eval_negation-triplets_loss": 2.081930160522461, + "eval_negation-triplets_runtime": 1.0932, + "eval_negation-triplets_samples_per_second": 117.092, + "eval_negation-triplets_steps_per_second": 1.83, + "step": 640 + }, + { + "epoch": 0.4965089216446858, + "eval_scitail-pairs-pos_loss": 0.24661773443222046, + "eval_scitail-pairs-pos_runtime": 1.2906, + "eval_scitail-pairs-pos_samples_per_second": 99.178, + "eval_scitail-pairs-pos_steps_per_second": 1.55, + "step": 640 + }, + { + "epoch": 0.4965089216446858, + "eval_scitail-pairs-qa_loss": 0.035427238792181015, + "eval_scitail-pairs-qa_runtime": 0.7845, + "eval_scitail-pairs-qa_samples_per_second": 163.161, + "eval_scitail-pairs-qa_steps_per_second": 2.549, + "step": 640 + }, + { + "epoch": 0.4965089216446858, + "eval_xsum-pairs_loss": 0.6712877154350281, + "eval_xsum-pairs_runtime": 3.9261, + "eval_xsum-pairs_samples_per_second": 32.602, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 640 + }, + { + "epoch": 0.4965089216446858, + "eval_sciq_pairs_loss": 0.09236491471529007, + "eval_sciq_pairs_runtime": 5.9782, + "eval_sciq_pairs_samples_per_second": 21.411, + "eval_sciq_pairs_steps_per_second": 0.335, + "step": 640 + }, + { + "epoch": 0.4965089216446858, + "eval_qasc_pairs_loss": 0.7825262546539307, + "eval_qasc_pairs_runtime": 0.918, + "eval_qasc_pairs_samples_per_second": 139.434, + "eval_qasc_pairs_steps_per_second": 2.179, + "step": 640 + }, + { + "epoch": 0.4965089216446858, + "eval_openbookqa_pairs_loss": 1.8152996301651, + "eval_openbookqa_pairs_runtime": 0.8717, + "eval_openbookqa_pairs_samples_per_second": 146.837, + "eval_openbookqa_pairs_steps_per_second": 2.294, + "step": 640 + }, + { + "epoch": 0.4965089216446858, + "eval_msmarco_pairs_loss": 0.8499898314476013, + "eval_msmarco_pairs_runtime": 1.9764, + "eval_msmarco_pairs_samples_per_second": 64.764, + "eval_msmarco_pairs_steps_per_second": 1.012, + "step": 640 + }, + { + "epoch": 0.4965089216446858, + "eval_nq_pairs_loss": 1.3879870176315308, + "eval_nq_pairs_runtime": 3.3372, + "eval_nq_pairs_samples_per_second": 38.355, + "eval_nq_pairs_steps_per_second": 0.599, + "step": 640 + }, + { + "epoch": 0.4965089216446858, + "eval_trivia_pairs_loss": 1.2515995502471924, + "eval_trivia_pairs_runtime": 4.502, + "eval_trivia_pairs_samples_per_second": 28.432, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 640 + }, + { + "epoch": 0.4965089216446858, + "eval_gooaq_pairs_loss": 1.1852116584777832, + "eval_gooaq_pairs_runtime": 1.4703, + "eval_gooaq_pairs_samples_per_second": 87.055, + "eval_gooaq_pairs_steps_per_second": 1.36, + "step": 640 + }, + { + "epoch": 0.4965089216446858, + "eval_paws-pos_loss": 0.04851341247558594, + "eval_paws-pos_runtime": 1.0391, + "eval_paws-pos_samples_per_second": 123.189, + "eval_paws-pos_steps_per_second": 1.925, + "step": 640 + }, + { + "epoch": 0.4965089216446858, + "eval_global_dataset_loss": 0.5946388244628906, + "eval_global_dataset_runtime": 19.4336, + "eval_global_dataset_samples_per_second": 21.406, + "eval_global_dataset_steps_per_second": 0.36, + "step": 640 + }, + { + "epoch": 0.4972847168347556, + "grad_norm": 21.70700454711914, + "learning_rate": 2.623971797884841e-05, + "loss": 1.7148, + "step": 641 + }, + { + "epoch": 0.49806051202482543, + "grad_norm": 13.550596237182617, + "learning_rate": 2.628084606345476e-05, + "loss": 0.9538, + "step": 642 + }, + { + "epoch": 0.49883630721489525, + "grad_norm": 12.98841667175293, + "learning_rate": 2.6321974148061104e-05, + "loss": 0.6802, + "step": 643 + }, + { + "epoch": 0.49961210240496506, + "grad_norm": 14.727252960205078, + "learning_rate": 2.6363102232667446e-05, + "loss": 1.379, + "step": 644 + }, + { + "epoch": 0.5003878975950349, + "grad_norm": 15.235370635986328, + "learning_rate": 2.6404230317273794e-05, + "loss": 1.2811, + "step": 645 + }, + { + "epoch": 0.5011636927851048, + "grad_norm": 9.874800682067871, + "learning_rate": 2.644535840188014e-05, + "loss": 0.6225, + "step": 646 + }, + { + "epoch": 0.5019394879751745, + "grad_norm": 8.094162940979004, + "learning_rate": 2.6486486486486485e-05, + "loss": 0.4069, + "step": 647 + }, + { + "epoch": 0.5027152831652444, + "grad_norm": 10.230360984802246, + "learning_rate": 2.652761457109283e-05, + "loss": 0.6908, + "step": 648 + }, + { + "epoch": 0.5034910783553141, + "grad_norm": 9.848830223083496, + "learning_rate": 2.6568742655699172e-05, + "loss": 1.2194, + "step": 649 + }, + { + "epoch": 0.504266873545384, + "grad_norm": 7.501684665679932, + "learning_rate": 2.660987074030552e-05, + "loss": 0.4051, + "step": 650 + }, + { + "epoch": 0.5050426687354539, + "grad_norm": 9.760265350341797, + "learning_rate": 2.6650998824911866e-05, + "loss": 0.7243, + "step": 651 + }, + { + "epoch": 0.5058184639255237, + "grad_norm": 9.546927452087402, + "learning_rate": 2.669212690951821e-05, + "loss": 0.5951, + "step": 652 + }, + { + "epoch": 0.5065942591155935, + "grad_norm": 10.041341781616211, + "learning_rate": 2.6733254994124557e-05, + "loss": 0.697, + "step": 653 + }, + { + "epoch": 0.5073700543056633, + "grad_norm": 12.542648315429688, + "learning_rate": 2.6774383078730905e-05, + "loss": 1.1804, + "step": 654 + }, + { + "epoch": 0.5081458494957332, + "grad_norm": 16.35436248779297, + "learning_rate": 2.6815511163337247e-05, + "loss": 1.5398, + "step": 655 + }, + { + "epoch": 0.5089216446858029, + "grad_norm": 10.068395614624023, + "learning_rate": 2.6856639247943592e-05, + "loss": 0.6309, + "step": 656 + }, + { + "epoch": 0.5096974398758728, + "grad_norm": 10.921636581420898, + "learning_rate": 2.689776733254994e-05, + "loss": 1.1841, + "step": 657 + }, + { + "epoch": 0.5104732350659426, + "grad_norm": 8.969158172607422, + "learning_rate": 2.6938895417156283e-05, + "loss": 0.5512, + "step": 658 + }, + { + "epoch": 0.5112490302560124, + "grad_norm": 8.403101921081543, + "learning_rate": 2.698002350176263e-05, + "loss": 0.4967, + "step": 659 + }, + { + "epoch": 0.5120248254460822, + "grad_norm": 8.566783905029297, + "learning_rate": 2.7021151586368977e-05, + "loss": 0.6073, + "step": 660 + }, + { + "epoch": 0.5120248254460822, + "eval_Qnli-dev_cosine_accuracy": 0.654296875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8996038436889648, + "eval_Qnli-dev_cosine_ap": 0.7004482553812913, + "eval_Qnli-dev_cosine_f1": 0.6656394453004623, + "eval_Qnli-dev_cosine_f1_threshold": 0.7712167501449585, + "eval_Qnli-dev_cosine_precision": 0.5230024213075061, + "eval_Qnli-dev_cosine_recall": 0.9152542372881356, + "eval_Qnli-dev_dot_accuracy": 0.607421875, + "eval_Qnli-dev_dot_accuracy_threshold": 473.8558349609375, + "eval_Qnli-dev_dot_ap": 0.5821478992671898, + "eval_Qnli-dev_dot_f1": 0.654867256637168, + "eval_Qnli-dev_dot_f1_threshold": 376.8409423828125, + "eval_Qnli-dev_dot_precision": 0.502262443438914, + "eval_Qnli-dev_dot_recall": 0.940677966101695, + "eval_Qnli-dev_euclidean_accuracy": 0.662109375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.721826553344727, + "eval_Qnli-dev_euclidean_ap": 0.7013404158321883, + "eval_Qnli-dev_euclidean_f1": 0.6700336700336701, + "eval_Qnli-dev_euclidean_f1_threshold": 14.47653865814209, + "eval_Qnli-dev_euclidean_precision": 0.5558659217877095, + "eval_Qnli-dev_euclidean_recall": 0.8432203389830508, + "eval_Qnli-dev_manhattan_accuracy": 0.662109375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 205.10516357421875, + "eval_Qnli-dev_manhattan_ap": 0.6999425706618274, + "eval_Qnli-dev_manhattan_f1": 0.6758620689655173, + "eval_Qnli-dev_manhattan_f1_threshold": 250.75860595703125, + "eval_Qnli-dev_manhattan_precision": 0.5697674418604651, + "eval_Qnli-dev_manhattan_recall": 0.8305084745762712, + "eval_Qnli-dev_max_accuracy": 0.662109375, + "eval_Qnli-dev_max_accuracy_threshold": 473.8558349609375, + "eval_Qnli-dev_max_ap": 0.7013404158321883, + "eval_Qnli-dev_max_f1": 0.6758620689655173, + "eval_Qnli-dev_max_f1_threshold": 376.8409423828125, + "eval_Qnli-dev_max_precision": 0.5697674418604651, + "eval_Qnli-dev_max_recall": 0.940677966101695, + "eval_allNLI-dev_cosine_accuracy": 0.708984375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.923484206199646, + "eval_allNLI-dev_cosine_ap": 0.5939446852176556, + "eval_allNLI-dev_cosine_f1": 0.62004662004662, + "eval_allNLI-dev_cosine_f1_threshold": 0.8559108972549438, + "eval_allNLI-dev_cosine_precision": 0.51953125, + "eval_allNLI-dev_cosine_recall": 0.7687861271676301, + "eval_allNLI-dev_dot_accuracy": 0.671875, + "eval_allNLI-dev_dot_accuracy_threshold": 485.31005859375, + "eval_allNLI-dev_dot_ap": 0.4760478892867671, + "eval_allNLI-dev_dot_f1": 0.5488029465930018, + "eval_allNLI-dev_dot_f1_threshold": 391.462646484375, + "eval_allNLI-dev_dot_precision": 0.4027027027027027, + "eval_allNLI-dev_dot_recall": 0.861271676300578, + "eval_allNLI-dev_euclidean_accuracy": 0.716796875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 8.66196060180664, + "eval_allNLI-dev_euclidean_ap": 0.5937230814089611, + "eval_allNLI-dev_euclidean_f1": 0.62582056892779, + "eval_allNLI-dev_euclidean_f1_threshold": 12.413965225219727, + "eval_allNLI-dev_euclidean_precision": 0.5035211267605634, + "eval_allNLI-dev_euclidean_recall": 0.8265895953757225, + "eval_allNLI-dev_manhattan_accuracy": 0.712890625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 162.57366943359375, + "eval_allNLI-dev_manhattan_ap": 0.5944493712705848, + "eval_allNLI-dev_manhattan_f1": 0.6161137440758294, + "eval_allNLI-dev_manhattan_f1_threshold": 212.20864868164062, + "eval_allNLI-dev_manhattan_precision": 0.5220883534136547, + "eval_allNLI-dev_manhattan_recall": 0.7514450867052023, + "eval_allNLI-dev_max_accuracy": 0.716796875, + "eval_allNLI-dev_max_accuracy_threshold": 485.31005859375, + "eval_allNLI-dev_max_ap": 0.5944493712705848, + "eval_allNLI-dev_max_f1": 0.62582056892779, + "eval_allNLI-dev_max_f1_threshold": 391.462646484375, + "eval_allNLI-dev_max_precision": 0.5220883534136547, + "eval_allNLI-dev_max_recall": 0.861271676300578, + "eval_sequential_score": 0.7013404158321883, + "eval_sts-test_pearson_cosine": 0.8024477677767311, + "eval_sts-test_pearson_dot": 0.7207474095128635, + "eval_sts-test_pearson_euclidean": 0.8240013125588865, + "eval_sts-test_pearson_manhattan": 0.8202441757516123, + "eval_sts-test_pearson_max": 0.8240013125588865, + "eval_sts-test_spearman_cosine": 0.8286941474222141, + "eval_sts-test_spearman_dot": 0.7141839037537189, + "eval_sts-test_spearman_euclidean": 0.8182952968126948, + "eval_sts-test_spearman_manhattan": 0.8143809039787892, + "eval_sts-test_spearman_max": 0.8286941474222141, + "eval_vitaminc-pairs_loss": 3.6424007415771484, + "eval_vitaminc-pairs_runtime": 5.5225, + "eval_vitaminc-pairs_samples_per_second": 23.178, + "eval_vitaminc-pairs_steps_per_second": 0.362, + "step": 660 + }, + { + "epoch": 0.5120248254460822, + "eval_negation-triplets_loss": 1.972699761390686, + "eval_negation-triplets_runtime": 1.0905, + "eval_negation-triplets_samples_per_second": 117.374, + "eval_negation-triplets_steps_per_second": 1.834, + "step": 660 + }, + { + "epoch": 0.5120248254460822, + "eval_scitail-pairs-pos_loss": 0.2093043029308319, + "eval_scitail-pairs-pos_runtime": 1.3112, + "eval_scitail-pairs-pos_samples_per_second": 97.617, + "eval_scitail-pairs-pos_steps_per_second": 1.525, + "step": 660 + }, + { + "epoch": 0.5120248254460822, + "eval_scitail-pairs-qa_loss": 0.042313043028116226, + "eval_scitail-pairs-qa_runtime": 0.7806, + "eval_scitail-pairs-qa_samples_per_second": 163.97, + "eval_scitail-pairs-qa_steps_per_second": 2.562, + "step": 660 + }, + { + "epoch": 0.5120248254460822, + "eval_xsum-pairs_loss": 0.6652930378913879, + "eval_xsum-pairs_runtime": 3.9336, + "eval_xsum-pairs_samples_per_second": 32.54, + "eval_xsum-pairs_steps_per_second": 0.508, + "step": 660 + }, + { + "epoch": 0.5120248254460822, + "eval_sciq_pairs_loss": 0.07537802308797836, + "eval_sciq_pairs_runtime": 5.9709, + "eval_sciq_pairs_samples_per_second": 21.437, + "eval_sciq_pairs_steps_per_second": 0.335, + "step": 660 + }, + { + "epoch": 0.5120248254460822, + "eval_qasc_pairs_loss": 0.7105440497398376, + "eval_qasc_pairs_runtime": 0.92, + "eval_qasc_pairs_samples_per_second": 139.131, + "eval_qasc_pairs_steps_per_second": 2.174, + "step": 660 + }, + { + "epoch": 0.5120248254460822, + "eval_openbookqa_pairs_loss": 1.8124297857284546, + "eval_openbookqa_pairs_runtime": 0.8772, + "eval_openbookqa_pairs_samples_per_second": 145.922, + "eval_openbookqa_pairs_steps_per_second": 2.28, + "step": 660 + }, + { + "epoch": 0.5120248254460822, + "eval_msmarco_pairs_loss": 0.7674368023872375, + "eval_msmarco_pairs_runtime": 1.9834, + "eval_msmarco_pairs_samples_per_second": 64.537, + "eval_msmarco_pairs_steps_per_second": 1.008, + "step": 660 + }, + { + "epoch": 0.5120248254460822, + "eval_nq_pairs_loss": 1.2386846542358398, + "eval_nq_pairs_runtime": 3.3394, + "eval_nq_pairs_samples_per_second": 38.331, + "eval_nq_pairs_steps_per_second": 0.599, + "step": 660 + }, + { + "epoch": 0.5120248254460822, + "eval_trivia_pairs_loss": 1.059404730796814, + "eval_trivia_pairs_runtime": 4.4937, + "eval_trivia_pairs_samples_per_second": 28.484, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 660 + }, + { + "epoch": 0.5120248254460822, + "eval_gooaq_pairs_loss": 1.0584646463394165, + "eval_gooaq_pairs_runtime": 1.4687, + "eval_gooaq_pairs_samples_per_second": 87.15, + "eval_gooaq_pairs_steps_per_second": 1.362, + "step": 660 + }, + { + "epoch": 0.5120248254460822, + "eval_paws-pos_loss": 0.05129885673522949, + "eval_paws-pos_runtime": 1.0343, + "eval_paws-pos_samples_per_second": 123.755, + "eval_paws-pos_steps_per_second": 1.934, + "step": 660 + }, + { + "epoch": 0.5120248254460822, + "eval_global_dataset_loss": 0.575580894947052, + "eval_global_dataset_runtime": 19.4356, + "eval_global_dataset_samples_per_second": 21.404, + "eval_global_dataset_steps_per_second": 0.36, + "step": 660 + }, + { + "epoch": 0.5128006206361521, + "grad_norm": 4.421882629394531, + "learning_rate": 2.706227967097532e-05, + "loss": 0.1732, + "step": 661 + }, + { + "epoch": 0.5135764158262219, + "grad_norm": 8.8997163772583, + "learning_rate": 2.7103407755581668e-05, + "loss": 1.046, + "step": 662 + }, + { + "epoch": 0.5143522110162917, + "grad_norm": 9.779997825622559, + "learning_rate": 2.714453584018801e-05, + "loss": 0.5477, + "step": 663 + }, + { + "epoch": 0.5151280062063616, + "grad_norm": 12.259178161621094, + "learning_rate": 2.7185663924794358e-05, + "loss": 1.4658, + "step": 664 + }, + { + "epoch": 0.5159038013964313, + "grad_norm": 9.47537899017334, + "learning_rate": 2.7226792009400703e-05, + "loss": 0.5393, + "step": 665 + }, + { + "epoch": 0.5166795965865012, + "grad_norm": 10.445185661315918, + "learning_rate": 2.7267920094007045e-05, + "loss": 0.6117, + "step": 666 + }, + { + "epoch": 0.517455391776571, + "grad_norm": 17.931814193725586, + "learning_rate": 2.7309048178613394e-05, + "loss": 1.598, + "step": 667 + }, + { + "epoch": 0.5182311869666408, + "grad_norm": 14.246345520019531, + "learning_rate": 2.735017626321974e-05, + "loss": 1.1418, + "step": 668 + }, + { + "epoch": 0.5190069821567106, + "grad_norm": 17.05196762084961, + "learning_rate": 2.7391304347826085e-05, + "loss": 1.4109, + "step": 669 + }, + { + "epoch": 0.5197827773467805, + "grad_norm": 10.566195487976074, + "learning_rate": 2.743243243243243e-05, + "loss": 0.7768, + "step": 670 + }, + { + "epoch": 0.5205585725368502, + "grad_norm": 12.296074867248535, + "learning_rate": 2.747356051703878e-05, + "loss": 1.0928, + "step": 671 + }, + { + "epoch": 0.5213343677269201, + "grad_norm": 15.5636625289917, + "learning_rate": 2.751468860164512e-05, + "loss": 1.3598, + "step": 672 + }, + { + "epoch": 0.5221101629169899, + "grad_norm": 13.134448051452637, + "learning_rate": 2.7555816686251466e-05, + "loss": 1.4472, + "step": 673 + }, + { + "epoch": 0.5228859581070597, + "grad_norm": 15.9171781539917, + "learning_rate": 2.7596944770857814e-05, + "loss": 1.6369, + "step": 674 + }, + { + "epoch": 0.5236617532971296, + "grad_norm": 13.004121780395508, + "learning_rate": 2.7638072855464156e-05, + "loss": 1.1756, + "step": 675 + }, + { + "epoch": 0.5244375484871994, + "grad_norm": 10.514543533325195, + "learning_rate": 2.7679200940070505e-05, + "loss": 0.741, + "step": 676 + }, + { + "epoch": 0.5252133436772692, + "grad_norm": 9.157439231872559, + "learning_rate": 2.7720329024676847e-05, + "loss": 1.0661, + "step": 677 + }, + { + "epoch": 0.525989138867339, + "grad_norm": 13.167852401733398, + "learning_rate": 2.7761457109283192e-05, + "loss": 1.3677, + "step": 678 + }, + { + "epoch": 0.5267649340574089, + "grad_norm": 13.578255653381348, + "learning_rate": 2.780258519388954e-05, + "loss": 1.5355, + "step": 679 + }, + { + "epoch": 0.5275407292474786, + "grad_norm": 12.97792911529541, + "learning_rate": 2.7843713278495883e-05, + "loss": 1.4658, + "step": 680 + }, + { + "epoch": 0.5275407292474786, + "eval_Qnli-dev_cosine_accuracy": 0.654296875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8644264340400696, + "eval_Qnli-dev_cosine_ap": 0.6922050596627256, + "eval_Qnli-dev_cosine_f1": 0.6696562032884903, + "eval_Qnli-dev_cosine_f1_threshold": 0.7277312874794006, + "eval_Qnli-dev_cosine_precision": 0.5173210161662818, + "eval_Qnli-dev_cosine_recall": 0.9491525423728814, + "eval_Qnli-dev_dot_accuracy": 0.61328125, + "eval_Qnli-dev_dot_accuracy_threshold": 463.3890075683594, + "eval_Qnli-dev_dot_ap": 0.6026016894387726, + "eval_Qnli-dev_dot_f1": 0.6595419847328244, + "eval_Qnli-dev_dot_f1_threshold": 386.73138427734375, + "eval_Qnli-dev_dot_precision": 0.5155131264916468, + "eval_Qnli-dev_dot_recall": 0.9152542372881356, + "eval_Qnli-dev_euclidean_accuracy": 0.65625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.804697036743164, + "eval_Qnli-dev_euclidean_ap": 0.6866391477878779, + "eval_Qnli-dev_euclidean_f1": 0.6736526946107785, + "eval_Qnli-dev_euclidean_f1_threshold": 17.03914451599121, + "eval_Qnli-dev_euclidean_precision": 0.5208333333333334, + "eval_Qnli-dev_euclidean_recall": 0.9533898305084746, + "eval_Qnli-dev_manhattan_accuracy": 0.658203125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 198.21234130859375, + "eval_Qnli-dev_manhattan_ap": 0.6895014812351475, + "eval_Qnli-dev_manhattan_f1": 0.6776859504132232, + "eval_Qnli-dev_manhattan_f1_threshold": 268.044677734375, + "eval_Qnli-dev_manhattan_precision": 0.5555555555555556, + "eval_Qnli-dev_manhattan_recall": 0.8686440677966102, + "eval_Qnli-dev_max_accuracy": 0.658203125, + "eval_Qnli-dev_max_accuracy_threshold": 463.3890075683594, + "eval_Qnli-dev_max_ap": 0.6922050596627256, + "eval_Qnli-dev_max_f1": 0.6776859504132232, + "eval_Qnli-dev_max_f1_threshold": 386.73138427734375, + "eval_Qnli-dev_max_precision": 0.5555555555555556, + "eval_Qnli-dev_max_recall": 0.9533898305084746, + "eval_allNLI-dev_cosine_accuracy": 0.712890625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.934099555015564, + "eval_allNLI-dev_cosine_ap": 0.5873651502429142, + "eval_allNLI-dev_cosine_f1": 0.6115288220551378, + "eval_allNLI-dev_cosine_f1_threshold": 0.8647106885910034, + "eval_allNLI-dev_cosine_precision": 0.5398230088495575, + "eval_allNLI-dev_cosine_recall": 0.7052023121387283, + "eval_allNLI-dev_dot_accuracy": 0.673828125, + "eval_allNLI-dev_dot_accuracy_threshold": 525.916748046875, + "eval_allNLI-dev_dot_ap": 0.4668792849736388, + "eval_allNLI-dev_dot_f1": 0.5579710144927535, + "eval_allNLI-dev_dot_f1_threshold": 390.6253356933594, + "eval_allNLI-dev_dot_precision": 0.40633245382585753, + "eval_allNLI-dev_dot_recall": 0.8901734104046243, + "eval_allNLI-dev_euclidean_accuracy": 0.71484375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 8.74948501586914, + "eval_allNLI-dev_euclidean_ap": 0.5864169354164954, + "eval_allNLI-dev_euclidean_f1": 0.6095238095238095, + "eval_allNLI-dev_euclidean_f1_threshold": 12.080808639526367, + "eval_allNLI-dev_euclidean_precision": 0.5182186234817814, + "eval_allNLI-dev_euclidean_recall": 0.7398843930635838, + "eval_allNLI-dev_manhattan_accuracy": 0.708984375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 151.1354522705078, + "eval_allNLI-dev_manhattan_ap": 0.5867145227357142, + "eval_allNLI-dev_manhattan_f1": 0.6075949367088609, + "eval_allNLI-dev_manhattan_f1_threshold": 208.61737060546875, + "eval_allNLI-dev_manhattan_precision": 0.5405405405405406, + "eval_allNLI-dev_manhattan_recall": 0.6936416184971098, + "eval_allNLI-dev_max_accuracy": 0.71484375, + "eval_allNLI-dev_max_accuracy_threshold": 525.916748046875, + "eval_allNLI-dev_max_ap": 0.5873651502429142, + "eval_allNLI-dev_max_f1": 0.6115288220551378, + "eval_allNLI-dev_max_f1_threshold": 390.6253356933594, + "eval_allNLI-dev_max_precision": 0.5405405405405406, + "eval_allNLI-dev_max_recall": 0.8901734104046243, + "eval_sequential_score": 0.6922050596627256, + "eval_sts-test_pearson_cosine": 0.8021800342859011, + "eval_sts-test_pearson_dot": 0.7361186235573641, + "eval_sts-test_pearson_euclidean": 0.8288547012828461, + "eval_sts-test_pearson_manhattan": 0.8282797630468803, + "eval_sts-test_pearson_max": 0.8288547012828461, + "eval_sts-test_spearman_cosine": 0.839665500918714, + "eval_sts-test_spearman_dot": 0.7279097421401206, + "eval_sts-test_spearman_euclidean": 0.8282389823426023, + "eval_sts-test_spearman_manhattan": 0.8274848340856346, + "eval_sts-test_spearman_max": 0.839665500918714, + "eval_vitaminc-pairs_loss": 3.6738266944885254, + "eval_vitaminc-pairs_runtime": 5.4865, + "eval_vitaminc-pairs_samples_per_second": 23.33, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 680 + }, + { + "epoch": 0.5275407292474786, + "eval_negation-triplets_loss": 1.974615454673767, + "eval_negation-triplets_runtime": 1.08, + "eval_negation-triplets_samples_per_second": 118.521, + "eval_negation-triplets_steps_per_second": 1.852, + "step": 680 + }, + { + "epoch": 0.5275407292474786, + "eval_scitail-pairs-pos_loss": 0.205085888504982, + "eval_scitail-pairs-pos_runtime": 1.2766, + "eval_scitail-pairs-pos_samples_per_second": 100.266, + "eval_scitail-pairs-pos_steps_per_second": 1.567, + "step": 680 + }, + { + "epoch": 0.5275407292474786, + "eval_scitail-pairs-qa_loss": 0.04349052533507347, + "eval_scitail-pairs-qa_runtime": 0.7803, + "eval_scitail-pairs-qa_samples_per_second": 164.045, + "eval_scitail-pairs-qa_steps_per_second": 2.563, + "step": 680 + }, + { + "epoch": 0.5275407292474786, + "eval_xsum-pairs_loss": 0.5800377726554871, + "eval_xsum-pairs_runtime": 3.9272, + "eval_xsum-pairs_samples_per_second": 32.594, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 680 + }, + { + "epoch": 0.5275407292474786, + "eval_sciq_pairs_loss": 0.08015203475952148, + "eval_sciq_pairs_runtime": 5.9538, + "eval_sciq_pairs_samples_per_second": 21.499, + "eval_sciq_pairs_steps_per_second": 0.336, + "step": 680 + }, + { + "epoch": 0.5275407292474786, + "eval_qasc_pairs_loss": 0.6561543345451355, + "eval_qasc_pairs_runtime": 0.9173, + "eval_qasc_pairs_samples_per_second": 139.534, + "eval_qasc_pairs_steps_per_second": 2.18, + "step": 680 + }, + { + "epoch": 0.5275407292474786, + "eval_openbookqa_pairs_loss": 1.7960286140441895, + "eval_openbookqa_pairs_runtime": 0.8707, + "eval_openbookqa_pairs_samples_per_second": 147.013, + "eval_openbookqa_pairs_steps_per_second": 2.297, + "step": 680 + }, + { + "epoch": 0.5275407292474786, + "eval_msmarco_pairs_loss": 0.680317223072052, + "eval_msmarco_pairs_runtime": 1.9735, + "eval_msmarco_pairs_samples_per_second": 64.859, + "eval_msmarco_pairs_steps_per_second": 1.013, + "step": 680 + }, + { + "epoch": 0.5275407292474786, + "eval_nq_pairs_loss": 1.3359739780426025, + "eval_nq_pairs_runtime": 3.3344, + "eval_nq_pairs_samples_per_second": 38.388, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 680 + }, + { + "epoch": 0.5275407292474786, + "eval_trivia_pairs_loss": 0.9878504872322083, + "eval_trivia_pairs_runtime": 4.4789, + "eval_trivia_pairs_samples_per_second": 28.578, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 680 + }, + { + "epoch": 0.5275407292474786, + "eval_gooaq_pairs_loss": 0.9781962633132935, + "eval_gooaq_pairs_runtime": 1.4847, + "eval_gooaq_pairs_samples_per_second": 86.213, + "eval_gooaq_pairs_steps_per_second": 1.347, + "step": 680 + }, + { + "epoch": 0.5275407292474786, + "eval_paws-pos_loss": 0.05222996696829796, + "eval_paws-pos_runtime": 1.0371, + "eval_paws-pos_samples_per_second": 123.416, + "eval_paws-pos_steps_per_second": 1.928, + "step": 680 + }, + { + "epoch": 0.5275407292474786, + "eval_global_dataset_loss": 0.5629461407661438, + "eval_global_dataset_runtime": 19.3857, + "eval_global_dataset_samples_per_second": 21.459, + "eval_global_dataset_steps_per_second": 0.361, + "step": 680 + }, + { + "epoch": 0.5283165244375485, + "grad_norm": 13.323928833007812, + "learning_rate": 2.788484136310223e-05, + "loss": 1.2859, + "step": 681 + }, + { + "epoch": 0.5290923196276183, + "grad_norm": 12.878803253173828, + "learning_rate": 2.7925969447708577e-05, + "loss": 1.2639, + "step": 682 + }, + { + "epoch": 0.5298681148176881, + "grad_norm": 3.876774787902832, + "learning_rate": 2.796709753231492e-05, + "loss": 0.1404, + "step": 683 + }, + { + "epoch": 0.5306439100077579, + "grad_norm": 3.3705224990844727, + "learning_rate": 2.8008225616921267e-05, + "loss": 0.0899, + "step": 684 + }, + { + "epoch": 0.5314197051978278, + "grad_norm": 14.240885734558105, + "learning_rate": 2.8049353701527613e-05, + "loss": 1.3069, + "step": 685 + }, + { + "epoch": 0.5321955003878976, + "grad_norm": 16.7686824798584, + "learning_rate": 2.8090481786133958e-05, + "loss": 1.3905, + "step": 686 + }, + { + "epoch": 0.5329712955779674, + "grad_norm": 8.619479179382324, + "learning_rate": 2.8131609870740303e-05, + "loss": 0.4895, + "step": 687 + }, + { + "epoch": 0.5337470907680373, + "grad_norm": 12.283214569091797, + "learning_rate": 2.8172737955346652e-05, + "loss": 1.4455, + "step": 688 + }, + { + "epoch": 0.534522885958107, + "grad_norm": 11.175566673278809, + "learning_rate": 2.8213866039952994e-05, + "loss": 0.5758, + "step": 689 + }, + { + "epoch": 0.5352986811481769, + "grad_norm": 9.76007080078125, + "learning_rate": 2.825499412455934e-05, + "loss": 0.7879, + "step": 690 + }, + { + "epoch": 0.5360744763382467, + "grad_norm": 13.363899230957031, + "learning_rate": 2.8296122209165684e-05, + "loss": 0.6647, + "step": 691 + }, + { + "epoch": 0.5368502715283165, + "grad_norm": 20.8702449798584, + "learning_rate": 2.833725029377203e-05, + "loss": 2.8455, + "step": 692 + }, + { + "epoch": 0.5376260667183863, + "grad_norm": 16.099815368652344, + "learning_rate": 2.8378378378378378e-05, + "loss": 1.1986, + "step": 693 + }, + { + "epoch": 0.5384018619084562, + "grad_norm": 6.164529323577881, + "learning_rate": 2.841950646298472e-05, + "loss": 0.1884, + "step": 694 + }, + { + "epoch": 0.5391776570985259, + "grad_norm": 9.595008850097656, + "learning_rate": 2.8460634547591065e-05, + "loss": 0.611, + "step": 695 + }, + { + "epoch": 0.5399534522885958, + "grad_norm": 13.926246643066406, + "learning_rate": 2.8501762632197414e-05, + "loss": 1.3208, + "step": 696 + }, + { + "epoch": 0.5407292474786657, + "grad_norm": 7.313656330108643, + "learning_rate": 2.8542890716803756e-05, + "loss": 0.3679, + "step": 697 + }, + { + "epoch": 0.5415050426687354, + "grad_norm": 12.229180335998535, + "learning_rate": 2.8584018801410105e-05, + "loss": 0.9536, + "step": 698 + }, + { + "epoch": 0.5422808378588053, + "grad_norm": 13.998590469360352, + "learning_rate": 2.862514688601645e-05, + "loss": 1.3225, + "step": 699 + }, + { + "epoch": 0.5430566330488751, + "grad_norm": 7.1051926612854, + "learning_rate": 2.8666274970622792e-05, + "loss": 0.3916, + "step": 700 + }, + { + "epoch": 0.5430566330488751, + "eval_Qnli-dev_cosine_accuracy": 0.66015625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8566777110099792, + "eval_Qnli-dev_cosine_ap": 0.6928397191777511, + "eval_Qnli-dev_cosine_f1": 0.6676737160120847, + "eval_Qnli-dev_cosine_f1_threshold": 0.7259476184844971, + "eval_Qnli-dev_cosine_precision": 0.5187793427230047, + "eval_Qnli-dev_cosine_recall": 0.9364406779661016, + "eval_Qnli-dev_dot_accuracy": 0.640625, + "eval_Qnli-dev_dot_accuracy_threshold": 472.6525573730469, + "eval_Qnli-dev_dot_ap": 0.6044743448434562, + "eval_Qnli-dev_dot_f1": 0.6723259762308998, + "eval_Qnli-dev_dot_f1_threshold": 417.5654602050781, + "eval_Qnli-dev_dot_precision": 0.5609065155807366, + "eval_Qnli-dev_dot_recall": 0.8389830508474576, + "eval_Qnli-dev_euclidean_accuracy": 0.658203125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.445531845092773, + "eval_Qnli-dev_euclidean_ap": 0.6879518054174698, + "eval_Qnli-dev_euclidean_f1": 0.6636636636636637, + "eval_Qnli-dev_euclidean_f1_threshold": 17.553129196166992, + "eval_Qnli-dev_euclidean_precision": 0.513953488372093, + "eval_Qnli-dev_euclidean_recall": 0.9364406779661016, + "eval_Qnli-dev_manhattan_accuracy": 0.658203125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 216.95904541015625, + "eval_Qnli-dev_manhattan_ap": 0.6921323777683974, + "eval_Qnli-dev_manhattan_f1": 0.670807453416149, + "eval_Qnli-dev_manhattan_f1_threshold": 295.45440673828125, + "eval_Qnli-dev_manhattan_precision": 0.5294117647058824, + "eval_Qnli-dev_manhattan_recall": 0.9152542372881356, + "eval_Qnli-dev_max_accuracy": 0.66015625, + "eval_Qnli-dev_max_accuracy_threshold": 472.6525573730469, + "eval_Qnli-dev_max_ap": 0.6928397191777511, + "eval_Qnli-dev_max_f1": 0.6723259762308998, + "eval_Qnli-dev_max_f1_threshold": 417.5654602050781, + "eval_Qnli-dev_max_precision": 0.5609065155807366, + "eval_Qnli-dev_max_recall": 0.9364406779661016, + "eval_allNLI-dev_cosine_accuracy": 0.71484375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.913508415222168, + "eval_allNLI-dev_cosine_ap": 0.5911741601194938, + "eval_allNLI-dev_cosine_f1": 0.6125290023201856, + "eval_allNLI-dev_cosine_f1_threshold": 0.8410093188285828, + "eval_allNLI-dev_cosine_precision": 0.5116279069767442, + "eval_allNLI-dev_cosine_recall": 0.7630057803468208, + "eval_allNLI-dev_dot_accuracy": 0.673828125, + "eval_allNLI-dev_dot_accuracy_threshold": 539.990478515625, + "eval_allNLI-dev_dot_ap": 0.48083911646945526, + "eval_allNLI-dev_dot_f1": 0.5535353535353535, + "eval_allNLI-dev_dot_f1_threshold": 412.3243408203125, + "eval_allNLI-dev_dot_precision": 0.4254658385093168, + "eval_allNLI-dev_dot_recall": 0.791907514450867, + "eval_allNLI-dev_euclidean_accuracy": 0.720703125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 9.402048110961914, + "eval_allNLI-dev_euclidean_ap": 0.5934062939780034, + "eval_allNLI-dev_euclidean_f1": 0.6143790849673202, + "eval_allNLI-dev_euclidean_f1_threshold": 13.571496963500977, + "eval_allNLI-dev_euclidean_precision": 0.493006993006993, + "eval_allNLI-dev_euclidean_recall": 0.815028901734104, + "eval_allNLI-dev_manhattan_accuracy": 0.71484375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 176.13900756835938, + "eval_allNLI-dev_manhattan_ap": 0.5924668029161843, + "eval_allNLI-dev_manhattan_f1": 0.611353711790393, + "eval_allNLI-dev_manhattan_f1_threshold": 245.57916259765625, + "eval_allNLI-dev_manhattan_precision": 0.49122807017543857, + "eval_allNLI-dev_manhattan_recall": 0.8092485549132948, + "eval_allNLI-dev_max_accuracy": 0.720703125, + "eval_allNLI-dev_max_accuracy_threshold": 539.990478515625, + "eval_allNLI-dev_max_ap": 0.5934062939780034, + "eval_allNLI-dev_max_f1": 0.6143790849673202, + "eval_allNLI-dev_max_f1_threshold": 412.3243408203125, + "eval_allNLI-dev_max_precision": 0.5116279069767442, + "eval_allNLI-dev_max_recall": 0.815028901734104, + "eval_sequential_score": 0.6928397191777511, + "eval_sts-test_pearson_cosine": 0.819462683188855, + "eval_sts-test_pearson_dot": 0.7482268675684371, + "eval_sts-test_pearson_euclidean": 0.841975338798404, + "eval_sts-test_pearson_manhattan": 0.8394824776397919, + "eval_sts-test_pearson_max": 0.841975338798404, + "eval_sts-test_spearman_cosine": 0.8489902765035784, + "eval_sts-test_spearman_dot": 0.7450188206389735, + "eval_sts-test_spearman_euclidean": 0.8375093403936148, + "eval_sts-test_spearman_manhattan": 0.8368243775867581, + "eval_sts-test_spearman_max": 0.8489902765035784, + "eval_vitaminc-pairs_loss": 3.752849817276001, + "eval_vitaminc-pairs_runtime": 5.4704, + "eval_vitaminc-pairs_samples_per_second": 23.399, + "eval_vitaminc-pairs_steps_per_second": 0.366, + "step": 700 + }, + { + "epoch": 0.5430566330488751, + "eval_negation-triplets_loss": 1.9750179052352905, + "eval_negation-triplets_runtime": 1.0754, + "eval_negation-triplets_samples_per_second": 119.022, + "eval_negation-triplets_steps_per_second": 1.86, + "step": 700 + }, + { + "epoch": 0.5430566330488751, + "eval_scitail-pairs-pos_loss": 0.21943919360637665, + "eval_scitail-pairs-pos_runtime": 1.273, + "eval_scitail-pairs-pos_samples_per_second": 100.551, + "eval_scitail-pairs-pos_steps_per_second": 1.571, + "step": 700 + }, + { + "epoch": 0.5430566330488751, + "eval_scitail-pairs-qa_loss": 0.035450175404548645, + "eval_scitail-pairs-qa_runtime": 0.7798, + "eval_scitail-pairs-qa_samples_per_second": 164.139, + "eval_scitail-pairs-qa_steps_per_second": 2.565, + "step": 700 + }, + { + "epoch": 0.5430566330488751, + "eval_xsum-pairs_loss": 0.5657890439033508, + "eval_xsum-pairs_runtime": 3.9265, + "eval_xsum-pairs_samples_per_second": 32.599, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 700 + }, + { + "epoch": 0.5430566330488751, + "eval_sciq_pairs_loss": 0.09307930618524551, + "eval_sciq_pairs_runtime": 5.864, + "eval_sciq_pairs_samples_per_second": 21.828, + "eval_sciq_pairs_steps_per_second": 0.341, + "step": 700 + }, + { + "epoch": 0.5430566330488751, + "eval_qasc_pairs_loss": 0.6284787654876709, + "eval_qasc_pairs_runtime": 0.9012, + "eval_qasc_pairs_samples_per_second": 142.038, + "eval_qasc_pairs_steps_per_second": 2.219, + "step": 700 + }, + { + "epoch": 0.5430566330488751, + "eval_openbookqa_pairs_loss": 1.6339654922485352, + "eval_openbookqa_pairs_runtime": 0.8546, + "eval_openbookqa_pairs_samples_per_second": 149.777, + "eval_openbookqa_pairs_steps_per_second": 2.34, + "step": 700 + }, + { + "epoch": 0.5430566330488751, + "eval_msmarco_pairs_loss": 0.7008572220802307, + "eval_msmarco_pairs_runtime": 1.972, + "eval_msmarco_pairs_samples_per_second": 64.908, + "eval_msmarco_pairs_steps_per_second": 1.014, + "step": 700 + }, + { + "epoch": 0.5430566330488751, + "eval_nq_pairs_loss": 1.3421883583068848, + "eval_nq_pairs_runtime": 3.3324, + "eval_nq_pairs_samples_per_second": 38.411, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 700 + }, + { + "epoch": 0.5430566330488751, + "eval_trivia_pairs_loss": 1.2664722204208374, + "eval_trivia_pairs_runtime": 4.4989, + "eval_trivia_pairs_samples_per_second": 28.452, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 700 + }, + { + "epoch": 0.5430566330488751, + "eval_gooaq_pairs_loss": 1.0903469324111938, + "eval_gooaq_pairs_runtime": 1.473, + "eval_gooaq_pairs_samples_per_second": 86.898, + "eval_gooaq_pairs_steps_per_second": 1.358, + "step": 700 + }, + { + "epoch": 0.5430566330488751, + "eval_paws-pos_loss": 0.04687386006116867, + "eval_paws-pos_runtime": 1.0378, + "eval_paws-pos_samples_per_second": 123.336, + "eval_paws-pos_steps_per_second": 1.927, + "step": 700 + }, + { + "epoch": 0.5430566330488751, + "eval_global_dataset_loss": 0.5735803842544556, + "eval_global_dataset_runtime": 19.4309, + "eval_global_dataset_samples_per_second": 21.409, + "eval_global_dataset_steps_per_second": 0.36, + "step": 700 + }, + { + "epoch": 0.543832428238945, + "grad_norm": 18.120939254760742, + "learning_rate": 2.870740305522914e-05, + "loss": 1.7663, + "step": 701 + }, + { + "epoch": 0.5446082234290147, + "grad_norm": 16.150516510009766, + "learning_rate": 2.8748531139835482e-05, + "loss": 1.2925, + "step": 702 + }, + { + "epoch": 0.5453840186190846, + "grad_norm": 7.960175037384033, + "learning_rate": 2.878965922444183e-05, + "loss": 0.4823, + "step": 703 + }, + { + "epoch": 0.5461598138091543, + "grad_norm": 10.59144115447998, + "learning_rate": 2.8830787309048176e-05, + "loss": 1.0832, + "step": 704 + }, + { + "epoch": 0.5469356089992242, + "grad_norm": 11.555217742919922, + "learning_rate": 2.8871915393654522e-05, + "loss": 1.2817, + "step": 705 + }, + { + "epoch": 0.547711404189294, + "grad_norm": 9.403205871582031, + "learning_rate": 2.8913043478260867e-05, + "loss": 0.6168, + "step": 706 + }, + { + "epoch": 0.5484871993793639, + "grad_norm": 6.8033647537231445, + "learning_rate": 2.8954171562867212e-05, + "loss": 0.2665, + "step": 707 + }, + { + "epoch": 0.5492629945694336, + "grad_norm": 8.434525489807129, + "learning_rate": 2.8995299647473558e-05, + "loss": 0.6778, + "step": 708 + }, + { + "epoch": 0.5500387897595035, + "grad_norm": 11.152477264404297, + "learning_rate": 2.9036427732079903e-05, + "loss": 1.2399, + "step": 709 + }, + { + "epoch": 0.5508145849495734, + "grad_norm": 13.578471183776855, + "learning_rate": 2.907755581668625e-05, + "loss": 1.2126, + "step": 710 + }, + { + "epoch": 0.5515903801396431, + "grad_norm": 15.13119125366211, + "learning_rate": 2.9118683901292593e-05, + "loss": 1.3319, + "step": 711 + }, + { + "epoch": 0.552366175329713, + "grad_norm": 9.639263153076172, + "learning_rate": 2.915981198589894e-05, + "loss": 0.5835, + "step": 712 + }, + { + "epoch": 0.5531419705197828, + "grad_norm": 14.466522216796875, + "learning_rate": 2.9200940070505287e-05, + "loss": 1.188, + "step": 713 + }, + { + "epoch": 0.5539177657098526, + "grad_norm": 10.483856201171875, + "learning_rate": 2.924206815511163e-05, + "loss": 0.5465, + "step": 714 + }, + { + "epoch": 0.5546935608999224, + "grad_norm": 19.497995376586914, + "learning_rate": 2.9283196239717978e-05, + "loss": 2.9539, + "step": 715 + }, + { + "epoch": 0.5554693560899923, + "grad_norm": 7.568907260894775, + "learning_rate": 2.932432432432432e-05, + "loss": 0.477, + "step": 716 + }, + { + "epoch": 0.556245151280062, + "grad_norm": 9.21462345123291, + "learning_rate": 2.9365452408930665e-05, + "loss": 0.5837, + "step": 717 + }, + { + "epoch": 0.5570209464701319, + "grad_norm": 13.529400825500488, + "learning_rate": 2.9406580493537014e-05, + "loss": 1.3591, + "step": 718 + }, + { + "epoch": 0.5577967416602017, + "grad_norm": 8.857770919799805, + "learning_rate": 2.9447708578143356e-05, + "loss": 0.5642, + "step": 719 + }, + { + "epoch": 0.5585725368502715, + "grad_norm": 8.515542984008789, + "learning_rate": 2.9488836662749704e-05, + "loss": 0.5866, + "step": 720 + }, + { + "epoch": 0.5585725368502715, + "eval_Qnli-dev_cosine_accuracy": 0.673828125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8940407037734985, + "eval_Qnli-dev_cosine_ap": 0.7205450298424871, + "eval_Qnli-dev_cosine_f1": 0.6752827140549272, + "eval_Qnli-dev_cosine_f1_threshold": 0.7951650619506836, + "eval_Qnli-dev_cosine_precision": 0.5456919060052219, + "eval_Qnli-dev_cosine_recall": 0.885593220338983, + "eval_Qnli-dev_dot_accuracy": 0.640625, + "eval_Qnli-dev_dot_accuracy_threshold": 453.888427734375, + "eval_Qnli-dev_dot_ap": 0.6214011405733493, + "eval_Qnli-dev_dot_f1": 0.6637554585152838, + "eval_Qnli-dev_dot_f1_threshold": 385.6955261230469, + "eval_Qnli-dev_dot_precision": 0.5055432372505543, + "eval_Qnli-dev_dot_recall": 0.9661016949152542, + "eval_Qnli-dev_euclidean_accuracy": 0.666015625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.616241455078125, + "eval_Qnli-dev_euclidean_ap": 0.7139937882040119, + "eval_Qnli-dev_euclidean_f1": 0.6765676567656767, + "eval_Qnli-dev_euclidean_f1_threshold": 14.782005310058594, + "eval_Qnli-dev_euclidean_precision": 0.5540540540540541, + "eval_Qnli-dev_euclidean_recall": 0.8686440677966102, + "eval_Qnli-dev_manhattan_accuracy": 0.6640625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 199.71429443359375, + "eval_Qnli-dev_manhattan_ap": 0.7147017217346366, + "eval_Qnli-dev_manhattan_f1": 0.6812080536912752, + "eval_Qnli-dev_manhattan_f1_threshold": 259.9438781738281, + "eval_Qnli-dev_manhattan_precision": 0.5638888888888889, + "eval_Qnli-dev_manhattan_recall": 0.8601694915254238, + "eval_Qnli-dev_max_accuracy": 0.673828125, + "eval_Qnli-dev_max_accuracy_threshold": 453.888427734375, + "eval_Qnli-dev_max_ap": 0.7205450298424871, + "eval_Qnli-dev_max_f1": 0.6812080536912752, + "eval_Qnli-dev_max_f1_threshold": 385.6955261230469, + "eval_Qnli-dev_max_precision": 0.5638888888888889, + "eval_Qnli-dev_max_recall": 0.9661016949152542, + "eval_allNLI-dev_cosine_accuracy": 0.712890625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8952102065086365, + "eval_allNLI-dev_cosine_ap": 0.5866191424984795, + "eval_allNLI-dev_cosine_f1": 0.6153846153846154, + "eval_allNLI-dev_cosine_f1_threshold": 0.8231034874916077, + "eval_allNLI-dev_cosine_precision": 0.49645390070921985, + "eval_allNLI-dev_cosine_recall": 0.8092485549132948, + "eval_allNLI-dev_dot_accuracy": 0.673828125, + "eval_allNLI-dev_dot_accuracy_threshold": 433.2393493652344, + "eval_allNLI-dev_dot_ap": 0.46595992224599025, + "eval_allNLI-dev_dot_f1": 0.5708661417322834, + "eval_allNLI-dev_dot_f1_threshold": 373.5594482421875, + "eval_allNLI-dev_dot_precision": 0.43283582089552236, + "eval_allNLI-dev_dot_recall": 0.838150289017341, + "eval_allNLI-dev_euclidean_accuracy": 0.716796875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.566951751708984, + "eval_allNLI-dev_euclidean_ap": 0.5946647030558465, + "eval_allNLI-dev_euclidean_f1": 0.6064814814814814, + "eval_allNLI-dev_euclidean_f1_threshold": 12.611150741577148, + "eval_allNLI-dev_euclidean_precision": 0.5057915057915058, + "eval_allNLI-dev_euclidean_recall": 0.7572254335260116, + "eval_allNLI-dev_manhattan_accuracy": 0.72265625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 188.56045532226562, + "eval_allNLI-dev_manhattan_ap": 0.5967872765902448, + "eval_allNLI-dev_manhattan_f1": 0.6235565819861432, + "eval_allNLI-dev_manhattan_f1_threshold": 227.69821166992188, + "eval_allNLI-dev_manhattan_precision": 0.5192307692307693, + "eval_allNLI-dev_manhattan_recall": 0.7803468208092486, + "eval_allNLI-dev_max_accuracy": 0.72265625, + "eval_allNLI-dev_max_accuracy_threshold": 433.2393493652344, + "eval_allNLI-dev_max_ap": 0.5967872765902448, + "eval_allNLI-dev_max_f1": 0.6235565819861432, + "eval_allNLI-dev_max_f1_threshold": 373.5594482421875, + "eval_allNLI-dev_max_precision": 0.5192307692307693, + "eval_allNLI-dev_max_recall": 0.838150289017341, + "eval_sequential_score": 0.7205450298424871, + "eval_sts-test_pearson_cosine": 0.8068351179949242, + "eval_sts-test_pearson_dot": 0.7752161321253326, + "eval_sts-test_pearson_euclidean": 0.8352793694228442, + "eval_sts-test_pearson_manhattan": 0.8299592250249215, + "eval_sts-test_pearson_max": 0.8352793694228442, + "eval_sts-test_spearman_cosine": 0.8387028117836621, + "eval_sts-test_spearman_dot": 0.7567988294288159, + "eval_sts-test_spearman_euclidean": 0.831171219242751, + "eval_sts-test_spearman_manhattan": 0.8287938979838556, + "eval_sts-test_spearman_max": 0.8387028117836621, + "eval_vitaminc-pairs_loss": 3.6469686031341553, + "eval_vitaminc-pairs_runtime": 5.4931, + "eval_vitaminc-pairs_samples_per_second": 23.302, + "eval_vitaminc-pairs_steps_per_second": 0.364, + "step": 720 + }, + { + "epoch": 0.5585725368502715, + "eval_negation-triplets_loss": 1.8540340662002563, + "eval_negation-triplets_runtime": 1.0821, + "eval_negation-triplets_samples_per_second": 118.293, + "eval_negation-triplets_steps_per_second": 1.848, + "step": 720 + }, + { + "epoch": 0.5585725368502715, + "eval_scitail-pairs-pos_loss": 0.21250325441360474, + "eval_scitail-pairs-pos_runtime": 1.2752, + "eval_scitail-pairs-pos_samples_per_second": 100.38, + "eval_scitail-pairs-pos_steps_per_second": 1.568, + "step": 720 + }, + { + "epoch": 0.5585725368502715, + "eval_scitail-pairs-qa_loss": 0.05020624026656151, + "eval_scitail-pairs-qa_runtime": 0.7946, + "eval_scitail-pairs-qa_samples_per_second": 161.095, + "eval_scitail-pairs-qa_steps_per_second": 2.517, + "step": 720 + }, + { + "epoch": 0.5585725368502715, + "eval_xsum-pairs_loss": 0.5439064502716064, + "eval_xsum-pairs_runtime": 3.947, + "eval_xsum-pairs_samples_per_second": 32.43, + "eval_xsum-pairs_steps_per_second": 0.507, + "step": 720 + }, + { + "epoch": 0.5585725368502715, + "eval_sciq_pairs_loss": 0.07200054079294205, + "eval_sciq_pairs_runtime": 5.9421, + "eval_sciq_pairs_samples_per_second": 21.541, + "eval_sciq_pairs_steps_per_second": 0.337, + "step": 720 + }, + { + "epoch": 0.5585725368502715, + "eval_qasc_pairs_loss": 0.6132616996765137, + "eval_qasc_pairs_runtime": 0.9122, + "eval_qasc_pairs_samples_per_second": 140.313, + "eval_qasc_pairs_steps_per_second": 2.192, + "step": 720 + }, + { + "epoch": 0.5585725368502715, + "eval_openbookqa_pairs_loss": 1.5304473638534546, + "eval_openbookqa_pairs_runtime": 0.861, + "eval_openbookqa_pairs_samples_per_second": 148.663, + "eval_openbookqa_pairs_steps_per_second": 2.323, + "step": 720 + }, + { + "epoch": 0.5585725368502715, + "eval_msmarco_pairs_loss": 0.5928510427474976, + "eval_msmarco_pairs_runtime": 1.964, + "eval_msmarco_pairs_samples_per_second": 65.174, + "eval_msmarco_pairs_steps_per_second": 1.018, + "step": 720 + }, + { + "epoch": 0.5585725368502715, + "eval_nq_pairs_loss": 1.1065574884414673, + "eval_nq_pairs_runtime": 3.3464, + "eval_nq_pairs_samples_per_second": 38.25, + "eval_nq_pairs_steps_per_second": 0.598, + "step": 720 + }, + { + "epoch": 0.5585725368502715, + "eval_trivia_pairs_loss": 1.134360909461975, + "eval_trivia_pairs_runtime": 4.483, + "eval_trivia_pairs_samples_per_second": 28.553, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 720 + }, + { + "epoch": 0.5585725368502715, + "eval_gooaq_pairs_loss": 0.9950035810470581, + "eval_gooaq_pairs_runtime": 1.4614, + "eval_gooaq_pairs_samples_per_second": 87.585, + "eval_gooaq_pairs_steps_per_second": 1.369, + "step": 720 + }, + { + "epoch": 0.5585725368502715, + "eval_paws-pos_loss": 0.04947199672460556, + "eval_paws-pos_runtime": 1.037, + "eval_paws-pos_samples_per_second": 123.428, + "eval_paws-pos_steps_per_second": 1.929, + "step": 720 + }, + { + "epoch": 0.5585725368502715, + "eval_global_dataset_loss": 0.5364225506782532, + "eval_global_dataset_runtime": 19.4016, + "eval_global_dataset_samples_per_second": 21.441, + "eval_global_dataset_steps_per_second": 0.361, + "step": 720 + }, + { + "epoch": 0.5593483320403414, + "grad_norm": 3.8790650367736816, + "learning_rate": 2.952996474735605e-05, + "loss": 0.1448, + "step": 721 + }, + { + "epoch": 0.5601241272304112, + "grad_norm": 7.98465633392334, + "learning_rate": 2.9571092831962395e-05, + "loss": 0.523, + "step": 722 + }, + { + "epoch": 0.560899922420481, + "grad_norm": 9.091803550720215, + "learning_rate": 2.961222091656874e-05, + "loss": 0.6414, + "step": 723 + }, + { + "epoch": 0.5616757176105508, + "grad_norm": 12.962803840637207, + "learning_rate": 2.9653349001175086e-05, + "loss": 1.2545, + "step": 724 + }, + { + "epoch": 0.5624515128006207, + "grad_norm": 12.292401313781738, + "learning_rate": 2.969447708578143e-05, + "loss": 1.0442, + "step": 725 + }, + { + "epoch": 0.5632273079906904, + "grad_norm": 0.5314335823059082, + "learning_rate": 2.9735605170387776e-05, + "loss": 0.0114, + "step": 726 + }, + { + "epoch": 0.5640031031807603, + "grad_norm": 9.92776107788086, + "learning_rate": 2.9776733254994125e-05, + "loss": 0.6287, + "step": 727 + }, + { + "epoch": 0.5647788983708301, + "grad_norm": 10.079903602600098, + "learning_rate": 2.9817861339600467e-05, + "loss": 0.5608, + "step": 728 + }, + { + "epoch": 0.5655546935608999, + "grad_norm": 8.860121726989746, + "learning_rate": 2.9858989424206812e-05, + "loss": 0.3379, + "step": 729 + }, + { + "epoch": 0.5663304887509697, + "grad_norm": 7.067585468292236, + "learning_rate": 2.9900117508813157e-05, + "loss": 0.2763, + "step": 730 + }, + { + "epoch": 0.5671062839410396, + "grad_norm": 1.1004122495651245, + "learning_rate": 2.9941245593419503e-05, + "loss": 0.025, + "step": 731 + }, + { + "epoch": 0.5678820791311094, + "grad_norm": 16.58653450012207, + "learning_rate": 2.998237367802585e-05, + "loss": 1.3182, + "step": 732 + }, + { + "epoch": 0.5686578743211792, + "grad_norm": 9.295772552490234, + "learning_rate": 3.0023501762632193e-05, + "loss": 0.5533, + "step": 733 + }, + { + "epoch": 0.5694336695112491, + "grad_norm": 8.6797513961792, + "learning_rate": 3.0064629847238542e-05, + "loss": 0.4049, + "step": 734 + }, + { + "epoch": 0.5702094647013188, + "grad_norm": 2.2837955951690674, + "learning_rate": 3.0105757931844887e-05, + "loss": 0.0823, + "step": 735 + }, + { + "epoch": 0.5709852598913887, + "grad_norm": 10.1045503616333, + "learning_rate": 3.014688601645123e-05, + "loss": 0.4921, + "step": 736 + }, + { + "epoch": 0.5717610550814585, + "grad_norm": 9.599879264831543, + "learning_rate": 3.0188014101057578e-05, + "loss": 0.5957, + "step": 737 + }, + { + "epoch": 0.5725368502715283, + "grad_norm": 11.49008846282959, + "learning_rate": 3.0229142185663923e-05, + "loss": 0.5823, + "step": 738 + }, + { + "epoch": 0.5733126454615981, + "grad_norm": 8.599343299865723, + "learning_rate": 3.027027027027027e-05, + "loss": 0.3722, + "step": 739 + }, + { + "epoch": 0.574088440651668, + "grad_norm": 4.150580406188965, + "learning_rate": 3.0311398354876614e-05, + "loss": 0.1084, + "step": 740 + }, + { + "epoch": 0.574088440651668, + "eval_Qnli-dev_cosine_accuracy": 0.68359375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8398900628089905, + "eval_Qnli-dev_cosine_ap": 0.701964527309015, + "eval_Qnli-dev_cosine_f1": 0.6749226006191951, + "eval_Qnli-dev_cosine_f1_threshold": 0.7028575539588928, + "eval_Qnli-dev_cosine_precision": 0.5317073170731708, + "eval_Qnli-dev_cosine_recall": 0.923728813559322, + "eval_Qnli-dev_dot_accuracy": 0.638671875, + "eval_Qnli-dev_dot_accuracy_threshold": 434.2928161621094, + "eval_Qnli-dev_dot_ap": 0.6309465491829072, + "eval_Qnli-dev_dot_f1": 0.656, + "eval_Qnli-dev_dot_f1_threshold": 362.2476806640625, + "eval_Qnli-dev_dot_precision": 0.5269922879177378, + "eval_Qnli-dev_dot_recall": 0.8686440677966102, + "eval_Qnli-dev_euclidean_accuracy": 0.66796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.936728477478027, + "eval_Qnli-dev_euclidean_ap": 0.69981842752052, + "eval_Qnli-dev_euclidean_f1": 0.6823161189358372, + "eval_Qnli-dev_euclidean_f1_threshold": 17.49447250366211, + "eval_Qnli-dev_euclidean_precision": 0.5409429280397022, + "eval_Qnli-dev_euclidean_recall": 0.923728813559322, + "eval_Qnli-dev_manhattan_accuracy": 0.67578125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 231.48367309570312, + "eval_Qnli-dev_manhattan_ap": 0.7003099057246541, + "eval_Qnli-dev_manhattan_f1": 0.678125, + "eval_Qnli-dev_manhattan_f1_threshold": 312.58099365234375, + "eval_Qnli-dev_manhattan_precision": 0.5371287128712872, + "eval_Qnli-dev_manhattan_recall": 0.9194915254237288, + "eval_Qnli-dev_max_accuracy": 0.68359375, + "eval_Qnli-dev_max_accuracy_threshold": 434.2928161621094, + "eval_Qnli-dev_max_ap": 0.701964527309015, + "eval_Qnli-dev_max_f1": 0.6823161189358372, + "eval_Qnli-dev_max_f1_threshold": 362.2476806640625, + "eval_Qnli-dev_max_precision": 0.5409429280397022, + "eval_Qnli-dev_max_recall": 0.923728813559322, + "eval_allNLI-dev_cosine_accuracy": 0.70703125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8928638696670532, + "eval_allNLI-dev_cosine_ap": 0.5920124897509392, + "eval_allNLI-dev_cosine_f1": 0.6055045871559632, + "eval_allNLI-dev_cosine_f1_threshold": 0.8035704493522644, + "eval_allNLI-dev_cosine_precision": 0.5019011406844106, + "eval_allNLI-dev_cosine_recall": 0.7630057803468208, + "eval_allNLI-dev_dot_accuracy": 0.6796875, + "eval_allNLI-dev_dot_accuracy_threshold": 454.62860107421875, + "eval_allNLI-dev_dot_ap": 0.48208355686782006, + "eval_allNLI-dev_dot_f1": 0.5642105263157895, + "eval_allNLI-dev_dot_f1_threshold": 359.41827392578125, + "eval_allNLI-dev_dot_precision": 0.44370860927152317, + "eval_allNLI-dev_dot_recall": 0.7745664739884393, + "eval_allNLI-dev_euclidean_accuracy": 0.71875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 9.954824447631836, + "eval_allNLI-dev_euclidean_ap": 0.5969360708702632, + "eval_allNLI-dev_euclidean_f1": 0.6037735849056605, + "eval_allNLI-dev_euclidean_f1_threshold": 14.44923210144043, + "eval_allNLI-dev_euclidean_precision": 0.47368421052631576, + "eval_allNLI-dev_euclidean_recall": 0.8323699421965318, + "eval_allNLI-dev_manhattan_accuracy": 0.716796875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 180.33102416992188, + "eval_allNLI-dev_manhattan_ap": 0.599151505651083, + "eval_allNLI-dev_manhattan_f1": 0.6150341685649203, + "eval_allNLI-dev_manhattan_f1_threshold": 241.29034423828125, + "eval_allNLI-dev_manhattan_precision": 0.5075187969924813, + "eval_allNLI-dev_manhattan_recall": 0.7803468208092486, + "eval_allNLI-dev_max_accuracy": 0.71875, + "eval_allNLI-dev_max_accuracy_threshold": 454.62860107421875, + "eval_allNLI-dev_max_ap": 0.599151505651083, + "eval_allNLI-dev_max_f1": 0.6150341685649203, + "eval_allNLI-dev_max_f1_threshold": 359.41827392578125, + "eval_allNLI-dev_max_precision": 0.5075187969924813, + "eval_allNLI-dev_max_recall": 0.8323699421965318, + "eval_sequential_score": 0.701964527309015, + "eval_sts-test_pearson_cosine": 0.8093560512843387, + "eval_sts-test_pearson_dot": 0.768511375313704, + "eval_sts-test_pearson_euclidean": 0.8346196774403323, + "eval_sts-test_pearson_manhattan": 0.8336336376576433, + "eval_sts-test_pearson_max": 0.8346196774403323, + "eval_sts-test_spearman_cosine": 0.839810718761929, + "eval_sts-test_spearman_dot": 0.749576557606642, + "eval_sts-test_spearman_euclidean": 0.8316342157931065, + "eval_sts-test_spearman_manhattan": 0.8329256886873827, + "eval_sts-test_spearman_max": 0.839810718761929, + "eval_vitaminc-pairs_loss": 3.8233823776245117, + "eval_vitaminc-pairs_runtime": 5.4612, + "eval_vitaminc-pairs_samples_per_second": 23.438, + "eval_vitaminc-pairs_steps_per_second": 0.366, + "step": 740 + }, + { + "epoch": 0.574088440651668, + "eval_negation-triplets_loss": 1.911386251449585, + "eval_negation-triplets_runtime": 1.077, + "eval_negation-triplets_samples_per_second": 118.853, + "eval_negation-triplets_steps_per_second": 1.857, + "step": 740 + }, + { + "epoch": 0.574088440651668, + "eval_scitail-pairs-pos_loss": 0.21341906487941742, + "eval_scitail-pairs-pos_runtime": 1.2851, + "eval_scitail-pairs-pos_samples_per_second": 99.603, + "eval_scitail-pairs-pos_steps_per_second": 1.556, + "step": 740 + }, + { + "epoch": 0.574088440651668, + "eval_scitail-pairs-qa_loss": 0.029989801347255707, + "eval_scitail-pairs-qa_runtime": 0.7968, + "eval_scitail-pairs-qa_samples_per_second": 160.649, + "eval_scitail-pairs-qa_steps_per_second": 2.51, + "step": 740 + }, + { + "epoch": 0.574088440651668, + "eval_xsum-pairs_loss": 0.5320242047309875, + "eval_xsum-pairs_runtime": 3.9398, + "eval_xsum-pairs_samples_per_second": 32.489, + "eval_xsum-pairs_steps_per_second": 0.508, + "step": 740 + }, + { + "epoch": 0.574088440651668, + "eval_sciq_pairs_loss": 0.07748549431562424, + "eval_sciq_pairs_runtime": 6.0588, + "eval_sciq_pairs_samples_per_second": 21.126, + "eval_sciq_pairs_steps_per_second": 0.33, + "step": 740 + }, + { + "epoch": 0.574088440651668, + "eval_qasc_pairs_loss": 0.5939160585403442, + "eval_qasc_pairs_runtime": 0.9469, + "eval_qasc_pairs_samples_per_second": 135.174, + "eval_qasc_pairs_steps_per_second": 2.112, + "step": 740 + }, + { + "epoch": 0.574088440651668, + "eval_openbookqa_pairs_loss": 1.5577739477157593, + "eval_openbookqa_pairs_runtime": 0.8725, + "eval_openbookqa_pairs_samples_per_second": 146.701, + "eval_openbookqa_pairs_steps_per_second": 2.292, + "step": 740 + }, + { + "epoch": 0.574088440651668, + "eval_msmarco_pairs_loss": 0.6275253891944885, + "eval_msmarco_pairs_runtime": 1.9731, + "eval_msmarco_pairs_samples_per_second": 64.874, + "eval_msmarco_pairs_steps_per_second": 1.014, + "step": 740 + }, + { + "epoch": 0.574088440651668, + "eval_nq_pairs_loss": 1.280505657196045, + "eval_nq_pairs_runtime": 3.3422, + "eval_nq_pairs_samples_per_second": 38.298, + "eval_nq_pairs_steps_per_second": 0.598, + "step": 740 + }, + { + "epoch": 0.574088440651668, + "eval_trivia_pairs_loss": 1.201932668685913, + "eval_trivia_pairs_runtime": 4.4914, + "eval_trivia_pairs_samples_per_second": 28.499, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 740 + }, + { + "epoch": 0.574088440651668, + "eval_gooaq_pairs_loss": 1.1815366744995117, + "eval_gooaq_pairs_runtime": 1.474, + "eval_gooaq_pairs_samples_per_second": 86.838, + "eval_gooaq_pairs_steps_per_second": 1.357, + "step": 740 + }, + { + "epoch": 0.574088440651668, + "eval_paws-pos_loss": 0.04437920078635216, + "eval_paws-pos_runtime": 1.0358, + "eval_paws-pos_samples_per_second": 123.576, + "eval_paws-pos_steps_per_second": 1.931, + "step": 740 + }, + { + "epoch": 0.574088440651668, + "eval_global_dataset_loss": 0.5344300270080566, + "eval_global_dataset_runtime": 19.3994, + "eval_global_dataset_samples_per_second": 21.444, + "eval_global_dataset_steps_per_second": 0.361, + "step": 740 + }, + { + "epoch": 0.5748642358417377, + "grad_norm": 9.253707885742188, + "learning_rate": 3.0352526439482956e-05, + "loss": 0.528, + "step": 741 + }, + { + "epoch": 0.5756400310318076, + "grad_norm": 5.505712509155273, + "learning_rate": 3.0393654524089304e-05, + "loss": 0.1757, + "step": 742 + }, + { + "epoch": 0.5764158262218774, + "grad_norm": 9.570694923400879, + "learning_rate": 3.043478260869565e-05, + "loss": 0.5703, + "step": 743 + }, + { + "epoch": 0.5771916214119472, + "grad_norm": 12.39792537689209, + "learning_rate": 3.0475910693301995e-05, + "loss": 1.2171, + "step": 744 + }, + { + "epoch": 0.5779674166020171, + "grad_norm": 14.58137321472168, + "learning_rate": 3.051703877790834e-05, + "loss": 1.1231, + "step": 745 + }, + { + "epoch": 0.5787432117920869, + "grad_norm": 7.944159507751465, + "learning_rate": 3.0558166862514685e-05, + "loss": 0.3165, + "step": 746 + }, + { + "epoch": 0.5795190069821567, + "grad_norm": 12.121591567993164, + "learning_rate": 3.059929494712103e-05, + "loss": 0.6469, + "step": 747 + }, + { + "epoch": 0.5802948021722265, + "grad_norm": 18.77005958557129, + "learning_rate": 3.0640423031727376e-05, + "loss": 2.3595, + "step": 748 + }, + { + "epoch": 0.5810705973622964, + "grad_norm": 12.629343032836914, + "learning_rate": 3.068155111633372e-05, + "loss": 1.1556, + "step": 749 + }, + { + "epoch": 0.5818463925523661, + "grad_norm": 8.255861282348633, + "learning_rate": 3.0722679200940067e-05, + "loss": 0.3653, + "step": 750 + }, + { + "epoch": 0.582622187742436, + "grad_norm": 7.568626880645752, + "learning_rate": 3.076380728554641e-05, + "loss": 0.3635, + "step": 751 + }, + { + "epoch": 0.5833979829325058, + "grad_norm": 8.831418991088867, + "learning_rate": 3.080493537015276e-05, + "loss": 0.4417, + "step": 752 + }, + { + "epoch": 0.5841737781225756, + "grad_norm": 13.779094696044922, + "learning_rate": 3.08460634547591e-05, + "loss": 1.3804, + "step": 753 + }, + { + "epoch": 0.5849495733126454, + "grad_norm": 15.98426342010498, + "learning_rate": 3.088719153936545e-05, + "loss": 1.7223, + "step": 754 + }, + { + "epoch": 0.5857253685027153, + "grad_norm": 9.531943321228027, + "learning_rate": 3.092831962397179e-05, + "loss": 0.4985, + "step": 755 + }, + { + "epoch": 0.5865011636927852, + "grad_norm": 12.930288314819336, + "learning_rate": 3.0969447708578145e-05, + "loss": 1.1737, + "step": 756 + }, + { + "epoch": 0.5872769588828549, + "grad_norm": 10.649758338928223, + "learning_rate": 3.1010575793184484e-05, + "loss": 0.7157, + "step": 757 + }, + { + "epoch": 0.5880527540729248, + "grad_norm": 10.084814071655273, + "learning_rate": 3.105170387779083e-05, + "loss": 0.7424, + "step": 758 + }, + { + "epoch": 0.5888285492629945, + "grad_norm": 16.65734100341797, + "learning_rate": 3.109283196239718e-05, + "loss": 1.4451, + "step": 759 + }, + { + "epoch": 0.5896043444530644, + "grad_norm": 9.285470008850098, + "learning_rate": 3.113396004700352e-05, + "loss": 0.6282, + "step": 760 + }, + { + "epoch": 0.5896043444530644, + "eval_Qnli-dev_cosine_accuracy": 0.68359375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.877650797367096, + "eval_Qnli-dev_cosine_ap": 0.7098472636033686, + "eval_Qnli-dev_cosine_f1": 0.683464566929134, + "eval_Qnli-dev_cosine_f1_threshold": 0.7754403352737427, + "eval_Qnli-dev_cosine_precision": 0.543859649122807, + "eval_Qnli-dev_cosine_recall": 0.9194915254237288, + "eval_Qnli-dev_dot_accuracy": 0.62890625, + "eval_Qnli-dev_dot_accuracy_threshold": 501.2518310546875, + "eval_Qnli-dev_dot_ap": 0.6071012588934351, + "eval_Qnli-dev_dot_f1": 0.6666666666666667, + "eval_Qnli-dev_dot_f1_threshold": 413.00408935546875, + "eval_Qnli-dev_dot_precision": 0.5228915662650603, + "eval_Qnli-dev_dot_recall": 0.9194915254237288, + "eval_Qnli-dev_euclidean_accuracy": 0.68359375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.836280822753906, + "eval_Qnli-dev_euclidean_ap": 0.7140022543305473, + "eval_Qnli-dev_euclidean_f1": 0.688888888888889, + "eval_Qnli-dev_euclidean_f1_threshold": 15.903035163879395, + "eval_Qnli-dev_euclidean_precision": 0.550761421319797, + "eval_Qnli-dev_euclidean_recall": 0.9194915254237288, + "eval_Qnli-dev_manhattan_accuracy": 0.68359375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 206.20899963378906, + "eval_Qnli-dev_manhattan_ap": 0.7106636159518751, + "eval_Qnli-dev_manhattan_f1": 0.6893353941267387, + "eval_Qnli-dev_manhattan_f1_threshold": 290.93365478515625, + "eval_Qnli-dev_manhattan_precision": 0.5425790754257908, + "eval_Qnli-dev_manhattan_recall": 0.9449152542372882, + "eval_Qnli-dev_max_accuracy": 0.68359375, + "eval_Qnli-dev_max_accuracy_threshold": 501.2518310546875, + "eval_Qnli-dev_max_ap": 0.7140022543305473, + "eval_Qnli-dev_max_f1": 0.6893353941267387, + "eval_Qnli-dev_max_f1_threshold": 413.00408935546875, + "eval_Qnli-dev_max_precision": 0.550761421319797, + "eval_Qnli-dev_max_recall": 0.9449152542372882, + "eval_allNLI-dev_cosine_accuracy": 0.71484375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8954986333847046, + "eval_allNLI-dev_cosine_ap": 0.5995921924506478, + "eval_allNLI-dev_cosine_f1": 0.620985010706638, + "eval_allNLI-dev_cosine_f1_threshold": 0.8494750261306763, + "eval_allNLI-dev_cosine_precision": 0.4931972789115646, + "eval_allNLI-dev_cosine_recall": 0.838150289017341, + "eval_allNLI-dev_dot_accuracy": 0.677734375, + "eval_allNLI-dev_dot_accuracy_threshold": 490.69891357421875, + "eval_allNLI-dev_dot_ap": 0.471534213810263, + "eval_allNLI-dev_dot_f1": 0.5616161616161616, + "eval_allNLI-dev_dot_f1_threshold": 422.97576904296875, + "eval_allNLI-dev_dot_precision": 0.43167701863354035, + "eval_allNLI-dev_dot_recall": 0.8034682080924855, + "eval_allNLI-dev_euclidean_accuracy": 0.716796875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.38502025604248, + "eval_allNLI-dev_euclidean_ap": 0.6013805829707015, + "eval_allNLI-dev_euclidean_f1": 0.6225165562913907, + "eval_allNLI-dev_euclidean_f1_threshold": 12.259134292602539, + "eval_allNLI-dev_euclidean_precision": 0.5035714285714286, + "eval_allNLI-dev_euclidean_recall": 0.815028901734104, + "eval_allNLI-dev_manhattan_accuracy": 0.7265625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 190.51132202148438, + "eval_allNLI-dev_manhattan_ap": 0.6036438362676858, + "eval_allNLI-dev_manhattan_f1": 0.6363636363636364, + "eval_allNLI-dev_manhattan_f1_threshold": 223.367431640625, + "eval_allNLI-dev_manhattan_precision": 0.5086505190311419, + "eval_allNLI-dev_manhattan_recall": 0.8497109826589595, + "eval_allNLI-dev_max_accuracy": 0.7265625, + "eval_allNLI-dev_max_accuracy_threshold": 490.69891357421875, + "eval_allNLI-dev_max_ap": 0.6036438362676858, + "eval_allNLI-dev_max_f1": 0.6363636363636364, + "eval_allNLI-dev_max_f1_threshold": 422.97576904296875, + "eval_allNLI-dev_max_precision": 0.5086505190311419, + "eval_allNLI-dev_max_recall": 0.8497109826589595, + "eval_sequential_score": 0.7140022543305473, + "eval_sts-test_pearson_cosine": 0.804745942535201, + "eval_sts-test_pearson_dot": 0.7443646352401735, + "eval_sts-test_pearson_euclidean": 0.8272401719809497, + "eval_sts-test_pearson_manhattan": 0.8265389172613282, + "eval_sts-test_pearson_max": 0.8272401719809497, + "eval_sts-test_spearman_cosine": 0.8358944463750448, + "eval_sts-test_spearman_dot": 0.7358845959251136, + "eval_sts-test_spearman_euclidean": 0.8254604660253332, + "eval_sts-test_spearman_manhattan": 0.8263797911880966, + "eval_sts-test_spearman_max": 0.8358944463750448, + "eval_vitaminc-pairs_loss": 3.7820026874542236, + "eval_vitaminc-pairs_runtime": 5.4667, + "eval_vitaminc-pairs_samples_per_second": 23.414, + "eval_vitaminc-pairs_steps_per_second": 0.366, + "step": 760 + }, + { + "epoch": 0.5896043444530644, + "eval_negation-triplets_loss": 1.847720742225647, + "eval_negation-triplets_runtime": 1.073, + "eval_negation-triplets_samples_per_second": 119.29, + "eval_negation-triplets_steps_per_second": 1.864, + "step": 760 + }, + { + "epoch": 0.5896043444530644, + "eval_scitail-pairs-pos_loss": 0.2132282555103302, + "eval_scitail-pairs-pos_runtime": 1.3257, + "eval_scitail-pairs-pos_samples_per_second": 96.555, + "eval_scitail-pairs-pos_steps_per_second": 1.509, + "step": 760 + }, + { + "epoch": 0.5896043444530644, + "eval_scitail-pairs-qa_loss": 0.037515174597501755, + "eval_scitail-pairs-qa_runtime": 0.7803, + "eval_scitail-pairs-qa_samples_per_second": 164.041, + "eval_scitail-pairs-qa_steps_per_second": 2.563, + "step": 760 + }, + { + "epoch": 0.5896043444530644, + "eval_xsum-pairs_loss": 0.5676081776618958, + "eval_xsum-pairs_runtime": 3.9279, + "eval_xsum-pairs_samples_per_second": 32.588, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 760 + }, + { + "epoch": 0.5896043444530644, + "eval_sciq_pairs_loss": 0.07753700762987137, + "eval_sciq_pairs_runtime": 5.9445, + "eval_sciq_pairs_samples_per_second": 21.532, + "eval_sciq_pairs_steps_per_second": 0.336, + "step": 760 + }, + { + "epoch": 0.5896043444530644, + "eval_qasc_pairs_loss": 0.639065146446228, + "eval_qasc_pairs_runtime": 0.9162, + "eval_qasc_pairs_samples_per_second": 139.71, + "eval_qasc_pairs_steps_per_second": 2.183, + "step": 760 + }, + { + "epoch": 0.5896043444530644, + "eval_openbookqa_pairs_loss": 1.440195918083191, + "eval_openbookqa_pairs_runtime": 0.8723, + "eval_openbookqa_pairs_samples_per_second": 146.732, + "eval_openbookqa_pairs_steps_per_second": 2.293, + "step": 760 + }, + { + "epoch": 0.5896043444530644, + "eval_msmarco_pairs_loss": 0.6456062197685242, + "eval_msmarco_pairs_runtime": 1.9747, + "eval_msmarco_pairs_samples_per_second": 64.82, + "eval_msmarco_pairs_steps_per_second": 1.013, + "step": 760 + }, + { + "epoch": 0.5896043444530644, + "eval_nq_pairs_loss": 1.2071092128753662, + "eval_nq_pairs_runtime": 3.3352, + "eval_nq_pairs_samples_per_second": 38.379, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 760 + }, + { + "epoch": 0.5896043444530644, + "eval_trivia_pairs_loss": 1.0429044961929321, + "eval_trivia_pairs_runtime": 4.487, + "eval_trivia_pairs_samples_per_second": 28.527, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 760 + }, + { + "epoch": 0.5896043444530644, + "eval_gooaq_pairs_loss": 1.0560115575790405, + "eval_gooaq_pairs_runtime": 1.4648, + "eval_gooaq_pairs_samples_per_second": 87.384, + "eval_gooaq_pairs_steps_per_second": 1.365, + "step": 760 + }, + { + "epoch": 0.5896043444530644, + "eval_paws-pos_loss": 0.0502048097550869, + "eval_paws-pos_runtime": 1.0248, + "eval_paws-pos_samples_per_second": 124.903, + "eval_paws-pos_steps_per_second": 1.952, + "step": 760 + }, + { + "epoch": 0.5896043444530644, + "eval_global_dataset_loss": 0.5463089942932129, + "eval_global_dataset_runtime": 19.3988, + "eval_global_dataset_samples_per_second": 21.445, + "eval_global_dataset_steps_per_second": 0.361, + "step": 760 + }, + { + "epoch": 0.5903801396431342, + "grad_norm": 12.11174201965332, + "learning_rate": 3.117508813160987e-05, + "loss": 1.3516, + "step": 761 + }, + { + "epoch": 0.591155934833204, + "grad_norm": 6.552314281463623, + "learning_rate": 3.121621621621621e-05, + "loss": 0.3012, + "step": 762 + }, + { + "epoch": 0.5919317300232738, + "grad_norm": 10.677921295166016, + "learning_rate": 3.1257344300822555e-05, + "loss": 1.1564, + "step": 763 + }, + { + "epoch": 0.5927075252133437, + "grad_norm": 6.8679585456848145, + "learning_rate": 3.129847238542891e-05, + "loss": 0.565, + "step": 764 + }, + { + "epoch": 0.5934833204034135, + "grad_norm": 9.756589889526367, + "learning_rate": 3.1339600470035246e-05, + "loss": 0.6086, + "step": 765 + }, + { + "epoch": 0.5942591155934833, + "grad_norm": 7.003457069396973, + "learning_rate": 3.13807285546416e-05, + "loss": 0.4313, + "step": 766 + }, + { + "epoch": 0.5950349107835532, + "grad_norm": 7.544405460357666, + "learning_rate": 3.142185663924794e-05, + "loss": 0.7129, + "step": 767 + }, + { + "epoch": 0.595810705973623, + "grad_norm": 6.581875801086426, + "learning_rate": 3.146298472385429e-05, + "loss": 0.282, + "step": 768 + }, + { + "epoch": 0.5965865011636928, + "grad_norm": 6.3391313552856445, + "learning_rate": 3.1504112808460634e-05, + "loss": 0.2914, + "step": 769 + }, + { + "epoch": 0.5973622963537626, + "grad_norm": 14.847792625427246, + "learning_rate": 3.154524089306698e-05, + "loss": 1.248, + "step": 770 + }, + { + "epoch": 0.5981380915438325, + "grad_norm": 7.190470218658447, + "learning_rate": 3.1586368977673324e-05, + "loss": 0.3477, + "step": 771 + }, + { + "epoch": 0.5989138867339022, + "grad_norm": 11.779085159301758, + "learning_rate": 3.162749706227967e-05, + "loss": 1.3427, + "step": 772 + }, + { + "epoch": 0.5996896819239721, + "grad_norm": 15.973305702209473, + "learning_rate": 3.1668625146886015e-05, + "loss": 1.5705, + "step": 773 + }, + { + "epoch": 0.6004654771140419, + "grad_norm": 3.4099512100219727, + "learning_rate": 3.170975323149236e-05, + "loss": 0.1141, + "step": 774 + }, + { + "epoch": 0.6012412723041117, + "grad_norm": 0.6680812239646912, + "learning_rate": 3.1750881316098706e-05, + "loss": 0.0123, + "step": 775 + }, + { + "epoch": 0.6020170674941815, + "grad_norm": 6.4536051750183105, + "learning_rate": 3.179200940070505e-05, + "loss": 0.3727, + "step": 776 + }, + { + "epoch": 0.6027928626842514, + "grad_norm": 8.789738655090332, + "learning_rate": 3.1833137485311396e-05, + "loss": 0.5106, + "step": 777 + }, + { + "epoch": 0.6035686578743211, + "grad_norm": 16.27378273010254, + "learning_rate": 3.187426556991774e-05, + "loss": 1.3653, + "step": 778 + }, + { + "epoch": 0.604344453064391, + "grad_norm": 13.552659034729004, + "learning_rate": 3.191539365452409e-05, + "loss": 1.0364, + "step": 779 + }, + { + "epoch": 0.6051202482544609, + "grad_norm": 8.921616554260254, + "learning_rate": 3.195652173913043e-05, + "loss": 0.4358, + "step": 780 + }, + { + "epoch": 0.6051202482544609, + "eval_Qnli-dev_cosine_accuracy": 0.6796875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8365581035614014, + "eval_Qnli-dev_cosine_ap": 0.7109091471356163, + "eval_Qnli-dev_cosine_f1": 0.6774716369529984, + "eval_Qnli-dev_cosine_f1_threshold": 0.7240636944770813, + "eval_Qnli-dev_cosine_precision": 0.5485564304461942, + "eval_Qnli-dev_cosine_recall": 0.885593220338983, + "eval_Qnli-dev_dot_accuracy": 0.6484375, + "eval_Qnli-dev_dot_accuracy_threshold": 440.2427673339844, + "eval_Qnli-dev_dot_ap": 0.6363237135922722, + "eval_Qnli-dev_dot_f1": 0.6584615384615385, + "eval_Qnli-dev_dot_f1_threshold": 337.6985778808594, + "eval_Qnli-dev_dot_precision": 0.5169082125603864, + "eval_Qnli-dev_dot_recall": 0.9067796610169492, + "eval_Qnli-dev_euclidean_accuracy": 0.66796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.771133422851562, + "eval_Qnli-dev_euclidean_ap": 0.7028785095868835, + "eval_Qnli-dev_euclidean_f1": 0.6877076411960132, + "eval_Qnli-dev_euclidean_f1_threshold": 16.495817184448242, + "eval_Qnli-dev_euclidean_precision": 0.5655737704918032, + "eval_Qnli-dev_euclidean_recall": 0.8771186440677966, + "eval_Qnli-dev_manhattan_accuracy": 0.681640625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 230.85598754882812, + "eval_Qnli-dev_manhattan_ap": 0.6980767172326623, + "eval_Qnli-dev_manhattan_f1": 0.6842105263157895, + "eval_Qnli-dev_manhattan_f1_threshold": 296.935791015625, + "eval_Qnli-dev_manhattan_precision": 0.5591397849462365, + "eval_Qnli-dev_manhattan_recall": 0.8813559322033898, + "eval_Qnli-dev_max_accuracy": 0.681640625, + "eval_Qnli-dev_max_accuracy_threshold": 440.2427673339844, + "eval_Qnli-dev_max_ap": 0.7109091471356163, + "eval_Qnli-dev_max_f1": 0.6877076411960132, + "eval_Qnli-dev_max_f1_threshold": 337.6985778808594, + "eval_Qnli-dev_max_precision": 0.5655737704918032, + "eval_Qnli-dev_max_recall": 0.9067796610169492, + "eval_allNLI-dev_cosine_accuracy": 0.724609375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9060673117637634, + "eval_allNLI-dev_cosine_ap": 0.5995291187250216, + "eval_allNLI-dev_cosine_f1": 0.6070588235294118, + "eval_allNLI-dev_cosine_f1_threshold": 0.828079342842102, + "eval_allNLI-dev_cosine_precision": 0.5119047619047619, + "eval_allNLI-dev_cosine_recall": 0.7456647398843931, + "eval_allNLI-dev_dot_accuracy": 0.669921875, + "eval_allNLI-dev_dot_accuracy_threshold": 486.9648742675781, + "eval_allNLI-dev_dot_ap": 0.46196588529743243, + "eval_allNLI-dev_dot_f1": 0.553415061295972, + "eval_allNLI-dev_dot_f1_threshold": 324.0147705078125, + "eval_allNLI-dev_dot_precision": 0.3969849246231156, + "eval_allNLI-dev_dot_recall": 0.9132947976878613, + "eval_allNLI-dev_euclidean_accuracy": 0.728515625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 9.159957885742188, + "eval_allNLI-dev_euclidean_ap": 0.6074464508937693, + "eval_allNLI-dev_euclidean_f1": 0.615071283095723, + "eval_allNLI-dev_euclidean_f1_threshold": 14.04478931427002, + "eval_allNLI-dev_euclidean_precision": 0.4748427672955975, + "eval_allNLI-dev_euclidean_recall": 0.8728323699421965, + "eval_allNLI-dev_manhattan_accuracy": 0.73046875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 196.75416564941406, + "eval_allNLI-dev_manhattan_ap": 0.6073745409600696, + "eval_allNLI-dev_manhattan_f1": 0.6157205240174674, + "eval_allNLI-dev_manhattan_f1_threshold": 240.8726806640625, + "eval_allNLI-dev_manhattan_precision": 0.49473684210526314, + "eval_allNLI-dev_manhattan_recall": 0.815028901734104, + "eval_allNLI-dev_max_accuracy": 0.73046875, + "eval_allNLI-dev_max_accuracy_threshold": 486.9648742675781, + "eval_allNLI-dev_max_ap": 0.6074464508937693, + "eval_allNLI-dev_max_f1": 0.6157205240174674, + "eval_allNLI-dev_max_f1_threshold": 324.0147705078125, + "eval_allNLI-dev_max_precision": 0.5119047619047619, + "eval_allNLI-dev_max_recall": 0.9132947976878613, + "eval_sequential_score": 0.7109091471356163, + "eval_sts-test_pearson_cosine": 0.8102869734375284, + "eval_sts-test_pearson_dot": 0.7802602634945165, + "eval_sts-test_pearson_euclidean": 0.8366414194881395, + "eval_sts-test_pearson_manhattan": 0.8360601319983859, + "eval_sts-test_pearson_max": 0.8366414194881395, + "eval_sts-test_spearman_cosine": 0.8431522059475021, + "eval_sts-test_spearman_dot": 0.7624533303106646, + "eval_sts-test_spearman_euclidean": 0.8333223676265742, + "eval_sts-test_spearman_manhattan": 0.8343712476910657, + "eval_sts-test_spearman_max": 0.8431522059475021, + "eval_vitaminc-pairs_loss": 4.127828598022461, + "eval_vitaminc-pairs_runtime": 5.5473, + "eval_vitaminc-pairs_samples_per_second": 23.074, + "eval_vitaminc-pairs_steps_per_second": 0.361, + "step": 780 + }, + { + "epoch": 0.6051202482544609, + "eval_negation-triplets_loss": 1.8213467597961426, + "eval_negation-triplets_runtime": 1.1124, + "eval_negation-triplets_samples_per_second": 115.07, + "eval_negation-triplets_steps_per_second": 1.798, + "step": 780 + }, + { + "epoch": 0.6051202482544609, + "eval_scitail-pairs-pos_loss": 0.2154824137687683, + "eval_scitail-pairs-pos_runtime": 1.372, + "eval_scitail-pairs-pos_samples_per_second": 93.295, + "eval_scitail-pairs-pos_steps_per_second": 1.458, + "step": 780 + }, + { + "epoch": 0.6051202482544609, + "eval_scitail-pairs-qa_loss": 0.022496070712804794, + "eval_scitail-pairs-qa_runtime": 0.8115, + "eval_scitail-pairs-qa_samples_per_second": 157.73, + "eval_scitail-pairs-qa_steps_per_second": 2.465, + "step": 780 + }, + { + "epoch": 0.6051202482544609, + "eval_xsum-pairs_loss": 0.45196831226348877, + "eval_xsum-pairs_runtime": 3.9421, + "eval_xsum-pairs_samples_per_second": 32.47, + "eval_xsum-pairs_steps_per_second": 0.507, + "step": 780 + }, + { + "epoch": 0.6051202482544609, + "eval_sciq_pairs_loss": 0.08607526868581772, + "eval_sciq_pairs_runtime": 6.1291, + "eval_sciq_pairs_samples_per_second": 20.884, + "eval_sciq_pairs_steps_per_second": 0.326, + "step": 780 + }, + { + "epoch": 0.6051202482544609, + "eval_qasc_pairs_loss": 0.6608518958091736, + "eval_qasc_pairs_runtime": 0.9304, + "eval_qasc_pairs_samples_per_second": 137.573, + "eval_qasc_pairs_steps_per_second": 2.15, + "step": 780 + }, + { + "epoch": 0.6051202482544609, + "eval_openbookqa_pairs_loss": 1.5025335550308228, + "eval_openbookqa_pairs_runtime": 0.8706, + "eval_openbookqa_pairs_samples_per_second": 147.029, + "eval_openbookqa_pairs_steps_per_second": 2.297, + "step": 780 + }, + { + "epoch": 0.6051202482544609, + "eval_msmarco_pairs_loss": 0.6596011519432068, + "eval_msmarco_pairs_runtime": 1.9904, + "eval_msmarco_pairs_samples_per_second": 64.309, + "eval_msmarco_pairs_steps_per_second": 1.005, + "step": 780 + }, + { + "epoch": 0.6051202482544609, + "eval_nq_pairs_loss": 1.3736824989318848, + "eval_nq_pairs_runtime": 3.362, + "eval_nq_pairs_samples_per_second": 38.073, + "eval_nq_pairs_steps_per_second": 0.595, + "step": 780 + }, + { + "epoch": 0.6051202482544609, + "eval_trivia_pairs_loss": 1.0609238147735596, + "eval_trivia_pairs_runtime": 4.497, + "eval_trivia_pairs_samples_per_second": 28.464, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 780 + }, + { + "epoch": 0.6051202482544609, + "eval_gooaq_pairs_loss": 1.254521369934082, + "eval_gooaq_pairs_runtime": 1.4829, + "eval_gooaq_pairs_samples_per_second": 86.316, + "eval_gooaq_pairs_steps_per_second": 1.349, + "step": 780 + }, + { + "epoch": 0.6051202482544609, + "eval_paws-pos_loss": 0.04701700806617737, + "eval_paws-pos_runtime": 1.0641, + "eval_paws-pos_samples_per_second": 120.287, + "eval_paws-pos_steps_per_second": 1.879, + "step": 780 + }, + { + "epoch": 0.6051202482544609, + "eval_global_dataset_loss": 0.5899882316589355, + "eval_global_dataset_runtime": 19.4747, + "eval_global_dataset_samples_per_second": 21.361, + "eval_global_dataset_steps_per_second": 0.359, + "step": 780 + }, + { + "epoch": 0.6058960434445306, + "grad_norm": 1.117108702659607, + "learning_rate": 3.199764982373678e-05, + "loss": 0.0289, + "step": 781 + }, + { + "epoch": 0.6066718386346005, + "grad_norm": 1.0899808406829834, + "learning_rate": 3.203877790834312e-05, + "loss": 0.0641, + "step": 782 + }, + { + "epoch": 0.6074476338246703, + "grad_norm": 12.055098533630371, + "learning_rate": 3.207990599294947e-05, + "loss": 0.8557, + "step": 783 + }, + { + "epoch": 0.6082234290147401, + "grad_norm": 12.80833911895752, + "learning_rate": 3.212103407755581e-05, + "loss": 0.789, + "step": 784 + }, + { + "epoch": 0.6089992242048099, + "grad_norm": 12.154915809631348, + "learning_rate": 3.216216216216216e-05, + "loss": 0.5833, + "step": 785 + }, + { + "epoch": 0.6097750193948798, + "grad_norm": 11.571723937988281, + "learning_rate": 3.2203290246768504e-05, + "loss": 1.1265, + "step": 786 + }, + { + "epoch": 0.6105508145849495, + "grad_norm": 8.236241340637207, + "learning_rate": 3.224441833137485e-05, + "loss": 0.4488, + "step": 787 + }, + { + "epoch": 0.6113266097750194, + "grad_norm": 16.71768569946289, + "learning_rate": 3.2285546415981194e-05, + "loss": 1.2828, + "step": 788 + }, + { + "epoch": 0.6121024049650892, + "grad_norm": 2.64467191696167, + "learning_rate": 3.232667450058754e-05, + "loss": 0.0953, + "step": 789 + }, + { + "epoch": 0.612878200155159, + "grad_norm": 8.205010414123535, + "learning_rate": 3.2367802585193885e-05, + "loss": 0.647, + "step": 790 + }, + { + "epoch": 0.6136539953452289, + "grad_norm": 9.149142265319824, + "learning_rate": 3.240893066980023e-05, + "loss": 0.4797, + "step": 791 + }, + { + "epoch": 0.6144297905352987, + "grad_norm": 6.987081527709961, + "learning_rate": 3.2450058754406575e-05, + "loss": 0.316, + "step": 792 + }, + { + "epoch": 0.6152055857253685, + "grad_norm": 0.9240909218788147, + "learning_rate": 3.249118683901292e-05, + "loss": 0.0332, + "step": 793 + }, + { + "epoch": 0.6159813809154383, + "grad_norm": 10.872102737426758, + "learning_rate": 3.2532314923619266e-05, + "loss": 0.7116, + "step": 794 + }, + { + "epoch": 0.6167571761055082, + "grad_norm": 9.33710765838623, + "learning_rate": 3.257344300822562e-05, + "loss": 0.6129, + "step": 795 + }, + { + "epoch": 0.6175329712955779, + "grad_norm": 7.869011402130127, + "learning_rate": 3.2614571092831957e-05, + "loss": 0.362, + "step": 796 + }, + { + "epoch": 0.6183087664856478, + "grad_norm": 20.241748809814453, + "learning_rate": 3.26556991774383e-05, + "loss": 1.9527, + "step": 797 + }, + { + "epoch": 0.6190845616757176, + "grad_norm": 19.427724838256836, + "learning_rate": 3.2696827262044654e-05, + "loss": 1.6134, + "step": 798 + }, + { + "epoch": 0.6198603568657874, + "grad_norm": 12.748085975646973, + "learning_rate": 3.273795534665099e-05, + "loss": 0.9555, + "step": 799 + }, + { + "epoch": 0.6206361520558572, + "grad_norm": 13.742033004760742, + "learning_rate": 3.2779083431257345e-05, + "loss": 0.924, + "step": 800 + }, + { + "epoch": 0.6206361520558572, + "eval_Qnli-dev_cosine_accuracy": 0.66796875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.839040219783783, + "eval_Qnli-dev_cosine_ap": 0.7004889223738922, + "eval_Qnli-dev_cosine_f1": 0.6677966101694917, + "eval_Qnli-dev_cosine_f1_threshold": 0.7544183731079102, + "eval_Qnli-dev_cosine_precision": 0.556497175141243, + "eval_Qnli-dev_cosine_recall": 0.8347457627118644, + "eval_Qnli-dev_dot_accuracy": 0.634765625, + "eval_Qnli-dev_dot_accuracy_threshold": 440.27606201171875, + "eval_Qnli-dev_dot_ap": 0.6234031145909478, + "eval_Qnli-dev_dot_f1": 0.6539027982326951, + "eval_Qnli-dev_dot_f1_threshold": 328.9981689453125, + "eval_Qnli-dev_dot_precision": 0.5011286681715575, + "eval_Qnli-dev_dot_recall": 0.940677966101695, + "eval_Qnli-dev_euclidean_accuracy": 0.669921875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.322534561157227, + "eval_Qnli-dev_euclidean_ap": 0.6983872470433685, + "eval_Qnli-dev_euclidean_f1": 0.6765676567656767, + "eval_Qnli-dev_euclidean_f1_threshold": 16.18321990966797, + "eval_Qnli-dev_euclidean_precision": 0.5540540540540541, + "eval_Qnli-dev_euclidean_recall": 0.8686440677966102, + "eval_Qnli-dev_manhattan_accuracy": 0.6796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 228.90676879882812, + "eval_Qnli-dev_manhattan_ap": 0.6917529117366472, + "eval_Qnli-dev_manhattan_f1": 0.6793650793650794, + "eval_Qnli-dev_manhattan_f1_threshold": 300.86181640625, + "eval_Qnli-dev_manhattan_precision": 0.5431472081218274, + "eval_Qnli-dev_manhattan_recall": 0.9067796610169492, + "eval_Qnli-dev_max_accuracy": 0.6796875, + "eval_Qnli-dev_max_accuracy_threshold": 440.27606201171875, + "eval_Qnli-dev_max_ap": 0.7004889223738922, + "eval_Qnli-dev_max_f1": 0.6793650793650794, + "eval_Qnli-dev_max_f1_threshold": 328.9981689453125, + "eval_Qnli-dev_max_precision": 0.556497175141243, + "eval_Qnli-dev_max_recall": 0.940677966101695, + "eval_allNLI-dev_cosine_accuracy": 0.70703125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.875653088092804, + "eval_allNLI-dev_cosine_ap": 0.5807714821661151, + "eval_allNLI-dev_cosine_f1": 0.5968992248062016, + "eval_allNLI-dev_cosine_f1_threshold": 0.7751729488372803, + "eval_allNLI-dev_cosine_precision": 0.4489795918367347, + "eval_allNLI-dev_cosine_recall": 0.8901734104046243, + "eval_allNLI-dev_dot_accuracy": 0.671875, + "eval_allNLI-dev_dot_accuracy_threshold": 454.7177429199219, + "eval_allNLI-dev_dot_ap": 0.4489635586483933, + "eval_allNLI-dev_dot_f1": 0.5490196078431373, + "eval_allNLI-dev_dot_f1_threshold": 362.77105712890625, + "eval_allNLI-dev_dot_precision": 0.4405594405594406, + "eval_allNLI-dev_dot_recall": 0.7283236994219653, + "eval_allNLI-dev_euclidean_accuracy": 0.7109375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 9.839738845825195, + "eval_allNLI-dev_euclidean_ap": 0.5890042738121801, + "eval_allNLI-dev_euclidean_f1": 0.6083333333333334, + "eval_allNLI-dev_euclidean_f1_threshold": 13.389555931091309, + "eval_allNLI-dev_euclidean_precision": 0.4755700325732899, + "eval_allNLI-dev_euclidean_recall": 0.8439306358381503, + "eval_allNLI-dev_manhattan_accuracy": 0.72265625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 181.52691650390625, + "eval_allNLI-dev_manhattan_ap": 0.5955122415793616, + "eval_allNLI-dev_manhattan_f1": 0.6104417670682731, + "eval_allNLI-dev_manhattan_f1_threshold": 247.3948974609375, + "eval_allNLI-dev_manhattan_precision": 0.4676923076923077, + "eval_allNLI-dev_manhattan_recall": 0.8786127167630058, + "eval_allNLI-dev_max_accuracy": 0.72265625, + "eval_allNLI-dev_max_accuracy_threshold": 454.7177429199219, + "eval_allNLI-dev_max_ap": 0.5955122415793616, + "eval_allNLI-dev_max_f1": 0.6104417670682731, + "eval_allNLI-dev_max_f1_threshold": 362.77105712890625, + "eval_allNLI-dev_max_precision": 0.4755700325732899, + "eval_allNLI-dev_max_recall": 0.8901734104046243, + "eval_sequential_score": 0.7004889223738922, + "eval_sts-test_pearson_cosine": 0.8084181537766624, + "eval_sts-test_pearson_dot": 0.7575919585001736, + "eval_sts-test_pearson_euclidean": 0.8413080126015533, + "eval_sts-test_pearson_manhattan": 0.8403622379229991, + "eval_sts-test_pearson_max": 0.8413080126015533, + "eval_sts-test_spearman_cosine": 0.8464401706244509, + "eval_sts-test_spearman_dot": 0.7350349037250751, + "eval_sts-test_spearman_euclidean": 0.8402877672083613, + "eval_sts-test_spearman_manhattan": 0.8388123181202868, + "eval_sts-test_spearman_max": 0.8464401706244509, + "eval_vitaminc-pairs_loss": 3.7932844161987305, + "eval_vitaminc-pairs_runtime": 5.5254, + "eval_vitaminc-pairs_samples_per_second": 23.166, + "eval_vitaminc-pairs_steps_per_second": 0.362, + "step": 800 + }, + { + "epoch": 0.6206361520558572, + "eval_negation-triplets_loss": 1.8624801635742188, + "eval_negation-triplets_runtime": 1.0838, + "eval_negation-triplets_samples_per_second": 118.108, + "eval_negation-triplets_steps_per_second": 1.845, + "step": 800 + }, + { + "epoch": 0.6206361520558572, + "eval_scitail-pairs-pos_loss": 0.1760130673646927, + "eval_scitail-pairs-pos_runtime": 1.2866, + "eval_scitail-pairs-pos_samples_per_second": 99.49, + "eval_scitail-pairs-pos_steps_per_second": 1.555, + "step": 800 + }, + { + "epoch": 0.6206361520558572, + "eval_scitail-pairs-qa_loss": 0.02360847406089306, + "eval_scitail-pairs-qa_runtime": 0.7762, + "eval_scitail-pairs-qa_samples_per_second": 164.903, + "eval_scitail-pairs-qa_steps_per_second": 2.577, + "step": 800 + }, + { + "epoch": 0.6206361520558572, + "eval_xsum-pairs_loss": 0.42321959137916565, + "eval_xsum-pairs_runtime": 3.9268, + "eval_xsum-pairs_samples_per_second": 32.597, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 800 + }, + { + "epoch": 0.6206361520558572, + "eval_sciq_pairs_loss": 0.0696462094783783, + "eval_sciq_pairs_runtime": 5.9639, + "eval_sciq_pairs_samples_per_second": 21.463, + "eval_sciq_pairs_steps_per_second": 0.335, + "step": 800 + }, + { + "epoch": 0.6206361520558572, + "eval_qasc_pairs_loss": 0.5861497521400452, + "eval_qasc_pairs_runtime": 0.9105, + "eval_qasc_pairs_samples_per_second": 140.583, + "eval_qasc_pairs_steps_per_second": 2.197, + "step": 800 + }, + { + "epoch": 0.6206361520558572, + "eval_openbookqa_pairs_loss": 1.5731606483459473, + "eval_openbookqa_pairs_runtime": 0.8645, + "eval_openbookqa_pairs_samples_per_second": 148.068, + "eval_openbookqa_pairs_steps_per_second": 2.314, + "step": 800 + }, + { + "epoch": 0.6206361520558572, + "eval_msmarco_pairs_loss": 0.6926800012588501, + "eval_msmarco_pairs_runtime": 1.9805, + "eval_msmarco_pairs_samples_per_second": 64.63, + "eval_msmarco_pairs_steps_per_second": 1.01, + "step": 800 + }, + { + "epoch": 0.6206361520558572, + "eval_nq_pairs_loss": 1.2808607816696167, + "eval_nq_pairs_runtime": 3.7309, + "eval_nq_pairs_samples_per_second": 34.308, + "eval_nq_pairs_steps_per_second": 0.536, + "step": 800 + }, + { + "epoch": 0.6206361520558572, + "eval_trivia_pairs_loss": 1.0128430128097534, + "eval_trivia_pairs_runtime": 4.5157, + "eval_trivia_pairs_samples_per_second": 28.345, + "eval_trivia_pairs_steps_per_second": 0.443, + "step": 800 + }, + { + "epoch": 0.6206361520558572, + "eval_gooaq_pairs_loss": 1.2501243352890015, + "eval_gooaq_pairs_runtime": 1.4597, + "eval_gooaq_pairs_samples_per_second": 87.692, + "eval_gooaq_pairs_steps_per_second": 1.37, + "step": 800 + }, + { + "epoch": 0.6206361520558572, + "eval_paws-pos_loss": 0.04498037323355675, + "eval_paws-pos_runtime": 1.0305, + "eval_paws-pos_samples_per_second": 124.213, + "eval_paws-pos_steps_per_second": 1.941, + "step": 800 + }, + { + "epoch": 0.6206361520558572, + "eval_global_dataset_loss": 0.537517249584198, + "eval_global_dataset_runtime": 19.4173, + "eval_global_dataset_samples_per_second": 21.424, + "eval_global_dataset_steps_per_second": 0.361, + "step": 800 + }, + { + "epoch": 0.6214119472459271, + "grad_norm": 8.825736999511719, + "learning_rate": 3.282021151586368e-05, + "loss": 0.4939, + "step": 801 + }, + { + "epoch": 0.622187742435997, + "grad_norm": 1.2253526449203491, + "learning_rate": 3.2861339600470035e-05, + "loss": 0.0407, + "step": 802 + }, + { + "epoch": 0.6229635376260667, + "grad_norm": 8.241771697998047, + "learning_rate": 3.290246768507638e-05, + "loss": 0.3288, + "step": 803 + }, + { + "epoch": 0.6237393328161366, + "grad_norm": 9.082040786743164, + "learning_rate": 3.294359576968272e-05, + "loss": 0.3719, + "step": 804 + }, + { + "epoch": 0.6245151280062063, + "grad_norm": 8.657794952392578, + "learning_rate": 3.298472385428907e-05, + "loss": 0.5764, + "step": 805 + }, + { + "epoch": 0.6252909231962762, + "grad_norm": 4.004289150238037, + "learning_rate": 3.3025851938895416e-05, + "loss": 0.1127, + "step": 806 + }, + { + "epoch": 0.626066718386346, + "grad_norm": 10.909632682800293, + "learning_rate": 3.306698002350176e-05, + "loss": 0.6597, + "step": 807 + }, + { + "epoch": 0.6268425135764158, + "grad_norm": 7.65082311630249, + "learning_rate": 3.310810810810811e-05, + "loss": 0.325, + "step": 808 + }, + { + "epoch": 0.6276183087664856, + "grad_norm": 11.932289123535156, + "learning_rate": 3.314923619271445e-05, + "loss": 0.7185, + "step": 809 + }, + { + "epoch": 0.6283941039565555, + "grad_norm": 15.933886528015137, + "learning_rate": 3.31903642773208e-05, + "loss": 1.4173, + "step": 810 + }, + { + "epoch": 0.6291698991466252, + "grad_norm": 15.840822219848633, + "learning_rate": 3.323149236192714e-05, + "loss": 1.3152, + "step": 811 + }, + { + "epoch": 0.6299456943366951, + "grad_norm": 7.007137775421143, + "learning_rate": 3.327262044653349e-05, + "loss": 0.3601, + "step": 812 + }, + { + "epoch": 0.630721489526765, + "grad_norm": 9.546051979064941, + "learning_rate": 3.331374853113983e-05, + "loss": 0.5431, + "step": 813 + }, + { + "epoch": 0.6314972847168348, + "grad_norm": 1.17055344581604, + "learning_rate": 3.335487661574618e-05, + "loss": 0.05, + "step": 814 + }, + { + "epoch": 0.6322730799069046, + "grad_norm": 15.630793571472168, + "learning_rate": 3.3396004700352524e-05, + "loss": 1.3792, + "step": 815 + }, + { + "epoch": 0.6330488750969744, + "grad_norm": 10.699373245239258, + "learning_rate": 3.343713278495887e-05, + "loss": 0.6219, + "step": 816 + }, + { + "epoch": 0.6338246702870443, + "grad_norm": 15.029419898986816, + "learning_rate": 3.3478260869565214e-05, + "loss": 1.4932, + "step": 817 + }, + { + "epoch": 0.634600465477114, + "grad_norm": 8.618091583251953, + "learning_rate": 3.351938895417156e-05, + "loss": 0.541, + "step": 818 + }, + { + "epoch": 0.6353762606671839, + "grad_norm": 7.403913497924805, + "learning_rate": 3.3560517038777905e-05, + "loss": 0.3624, + "step": 819 + }, + { + "epoch": 0.6361520558572537, + "grad_norm": 7.243978500366211, + "learning_rate": 3.360164512338425e-05, + "loss": 0.405, + "step": 820 + }, + { + "epoch": 0.6361520558572537, + "eval_Qnli-dev_cosine_accuracy": 0.67578125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8632968664169312, + "eval_Qnli-dev_cosine_ap": 0.7105096682981787, + "eval_Qnli-dev_cosine_f1": 0.682615629984051, + "eval_Qnli-dev_cosine_f1_threshold": 0.7685413360595703, + "eval_Qnli-dev_cosine_precision": 0.5473145780051151, + "eval_Qnli-dev_cosine_recall": 0.9067796610169492, + "eval_Qnli-dev_dot_accuracy": 0.65234375, + "eval_Qnli-dev_dot_accuracy_threshold": 469.4715576171875, + "eval_Qnli-dev_dot_ap": 0.6164343060349063, + "eval_Qnli-dev_dot_f1": 0.6666666666666667, + "eval_Qnli-dev_dot_f1_threshold": 392.5911865234375, + "eval_Qnli-dev_dot_precision": 0.5256723716381418, + "eval_Qnli-dev_dot_recall": 0.9110169491525424, + "eval_Qnli-dev_euclidean_accuracy": 0.669921875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.602500915527344, + "eval_Qnli-dev_euclidean_ap": 0.7000730595374474, + "eval_Qnli-dev_euclidean_f1": 0.6793650793650794, + "eval_Qnli-dev_euclidean_f1_threshold": 16.257261276245117, + "eval_Qnli-dev_euclidean_precision": 0.5431472081218274, + "eval_Qnli-dev_euclidean_recall": 0.9067796610169492, + "eval_Qnli-dev_manhattan_accuracy": 0.6796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 214.744140625, + "eval_Qnli-dev_manhattan_ap": 0.6940773028961578, + "eval_Qnli-dev_manhattan_f1": 0.6864686468646866, + "eval_Qnli-dev_manhattan_f1_threshold": 277.89483642578125, + "eval_Qnli-dev_manhattan_precision": 0.5621621621621622, + "eval_Qnli-dev_manhattan_recall": 0.8813559322033898, + "eval_Qnli-dev_max_accuracy": 0.6796875, + "eval_Qnli-dev_max_accuracy_threshold": 469.4715576171875, + "eval_Qnli-dev_max_ap": 0.7105096682981787, + "eval_Qnli-dev_max_f1": 0.6864686468646866, + "eval_Qnli-dev_max_f1_threshold": 392.5911865234375, + "eval_Qnli-dev_max_precision": 0.5621621621621622, + "eval_Qnli-dev_max_recall": 0.9110169491525424, + "eval_allNLI-dev_cosine_accuracy": 0.7109375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9248532056808472, + "eval_allNLI-dev_cosine_ap": 0.5891613439882775, + "eval_allNLI-dev_cosine_f1": 0.5904255319148936, + "eval_allNLI-dev_cosine_f1_threshold": 0.8689008951187134, + "eval_allNLI-dev_cosine_precision": 0.5467980295566502, + "eval_allNLI-dev_cosine_recall": 0.6416184971098265, + "eval_allNLI-dev_dot_accuracy": 0.666015625, + "eval_allNLI-dev_dot_accuracy_threshold": 476.4175720214844, + "eval_allNLI-dev_dot_ap": 0.44362563057943183, + "eval_allNLI-dev_dot_f1": 0.5435168738898757, + "eval_allNLI-dev_dot_f1_threshold": 388.515625, + "eval_allNLI-dev_dot_precision": 0.3923076923076923, + "eval_allNLI-dev_dot_recall": 0.884393063583815, + "eval_allNLI-dev_euclidean_accuracy": 0.720703125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 8.71603775024414, + "eval_allNLI-dev_euclidean_ap": 0.5995995762619067, + "eval_allNLI-dev_euclidean_f1": 0.5978260869565217, + "eval_allNLI-dev_euclidean_f1_threshold": 11.47945499420166, + "eval_allNLI-dev_euclidean_precision": 0.5641025641025641, + "eval_allNLI-dev_euclidean_recall": 0.6358381502890174, + "eval_allNLI-dev_manhattan_accuracy": 0.71484375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 164.6937255859375, + "eval_allNLI-dev_manhattan_ap": 0.5987554652648344, + "eval_allNLI-dev_manhattan_f1": 0.6066838046272494, + "eval_allNLI-dev_manhattan_f1_threshold": 211.4429168701172, + "eval_allNLI-dev_manhattan_precision": 0.5462962962962963, + "eval_allNLI-dev_manhattan_recall": 0.6820809248554913, + "eval_allNLI-dev_max_accuracy": 0.720703125, + "eval_allNLI-dev_max_accuracy_threshold": 476.4175720214844, + "eval_allNLI-dev_max_ap": 0.5995995762619067, + "eval_allNLI-dev_max_f1": 0.6066838046272494, + "eval_allNLI-dev_max_f1_threshold": 388.515625, + "eval_allNLI-dev_max_precision": 0.5641025641025641, + "eval_allNLI-dev_max_recall": 0.884393063583815, + "eval_sequential_score": 0.7105096682981787, + "eval_sts-test_pearson_cosine": 0.8114975593444413, + "eval_sts-test_pearson_dot": 0.7777542583389272, + "eval_sts-test_pearson_euclidean": 0.8402442512729632, + "eval_sts-test_pearson_manhattan": 0.839058121041468, + "eval_sts-test_pearson_max": 0.8402442512729632, + "eval_sts-test_spearman_cosine": 0.8450279550687225, + "eval_sts-test_spearman_dot": 0.7640361376748169, + "eval_sts-test_spearman_euclidean": 0.8344748468611869, + "eval_sts-test_spearman_manhattan": 0.8344584678995177, + "eval_sts-test_spearman_max": 0.8450279550687225, + "eval_vitaminc-pairs_loss": 4.192330360412598, + "eval_vitaminc-pairs_runtime": 5.4934, + "eval_vitaminc-pairs_samples_per_second": 23.301, + "eval_vitaminc-pairs_steps_per_second": 0.364, + "step": 820 + }, + { + "epoch": 0.6361520558572537, + "eval_negation-triplets_loss": 1.776087999343872, + "eval_negation-triplets_runtime": 1.0852, + "eval_negation-triplets_samples_per_second": 117.952, + "eval_negation-triplets_steps_per_second": 1.843, + "step": 820 + }, + { + "epoch": 0.6361520558572537, + "eval_scitail-pairs-pos_loss": 0.16301314532756805, + "eval_scitail-pairs-pos_runtime": 1.2781, + "eval_scitail-pairs-pos_samples_per_second": 100.147, + "eval_scitail-pairs-pos_steps_per_second": 1.565, + "step": 820 + }, + { + "epoch": 0.6361520558572537, + "eval_scitail-pairs-qa_loss": 0.03246322646737099, + "eval_scitail-pairs-qa_runtime": 0.7859, + "eval_scitail-pairs-qa_samples_per_second": 162.869, + "eval_scitail-pairs-qa_steps_per_second": 2.545, + "step": 820 + }, + { + "epoch": 0.6361520558572537, + "eval_xsum-pairs_loss": 0.4296410381793976, + "eval_xsum-pairs_runtime": 3.9228, + "eval_xsum-pairs_samples_per_second": 32.629, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 820 + }, + { + "epoch": 0.6361520558572537, + "eval_sciq_pairs_loss": 0.07025601714849472, + "eval_sciq_pairs_runtime": 5.9689, + "eval_sciq_pairs_samples_per_second": 21.445, + "eval_sciq_pairs_steps_per_second": 0.335, + "step": 820 + }, + { + "epoch": 0.6361520558572537, + "eval_qasc_pairs_loss": 0.6239461302757263, + "eval_qasc_pairs_runtime": 0.9107, + "eval_qasc_pairs_samples_per_second": 140.545, + "eval_qasc_pairs_steps_per_second": 2.196, + "step": 820 + }, + { + "epoch": 0.6361520558572537, + "eval_openbookqa_pairs_loss": 1.4937922954559326, + "eval_openbookqa_pairs_runtime": 0.8682, + "eval_openbookqa_pairs_samples_per_second": 147.425, + "eval_openbookqa_pairs_steps_per_second": 2.304, + "step": 820 + }, + { + "epoch": 0.6361520558572537, + "eval_msmarco_pairs_loss": 0.616762101650238, + "eval_msmarco_pairs_runtime": 1.9623, + "eval_msmarco_pairs_samples_per_second": 65.23, + "eval_msmarco_pairs_steps_per_second": 1.019, + "step": 820 + }, + { + "epoch": 0.6361520558572537, + "eval_nq_pairs_loss": 1.1604129076004028, + "eval_nq_pairs_runtime": 3.3331, + "eval_nq_pairs_samples_per_second": 38.403, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 820 + }, + { + "epoch": 0.6361520558572537, + "eval_trivia_pairs_loss": 1.0031338930130005, + "eval_trivia_pairs_runtime": 4.4801, + "eval_trivia_pairs_samples_per_second": 28.571, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 820 + }, + { + "epoch": 0.6361520558572537, + "eval_gooaq_pairs_loss": 1.1163551807403564, + "eval_gooaq_pairs_runtime": 1.4598, + "eval_gooaq_pairs_samples_per_second": 87.681, + "eval_gooaq_pairs_steps_per_second": 1.37, + "step": 820 + }, + { + "epoch": 0.6361520558572537, + "eval_paws-pos_loss": 0.04240386188030243, + "eval_paws-pos_runtime": 1.0373, + "eval_paws-pos_samples_per_second": 123.399, + "eval_paws-pos_steps_per_second": 1.928, + "step": 820 + }, + { + "epoch": 0.6361520558572537, + "eval_global_dataset_loss": 0.5452271103858948, + "eval_global_dataset_runtime": 19.4483, + "eval_global_dataset_samples_per_second": 21.39, + "eval_global_dataset_steps_per_second": 0.36, + "step": 820 + }, + { + "epoch": 0.6369278510473235, + "grad_norm": 8.425275802612305, + "learning_rate": 3.3642773207990596e-05, + "loss": 0.4803, + "step": 821 + }, + { + "epoch": 0.6377036462373933, + "grad_norm": 2.8848042488098145, + "learning_rate": 3.368390129259694e-05, + "loss": 0.0942, + "step": 822 + }, + { + "epoch": 0.6384794414274632, + "grad_norm": 7.95778226852417, + "learning_rate": 3.3725029377203286e-05, + "loss": 0.52, + "step": 823 + }, + { + "epoch": 0.6392552366175329, + "grad_norm": 7.62559175491333, + "learning_rate": 3.376615746180963e-05, + "loss": 0.3963, + "step": 824 + }, + { + "epoch": 0.6400310318076028, + "grad_norm": 8.555058479309082, + "learning_rate": 3.380728554641598e-05, + "loss": 0.4374, + "step": 825 + }, + { + "epoch": 0.6408068269976727, + "grad_norm": 14.613774299621582, + "learning_rate": 3.384841363102232e-05, + "loss": 1.3948, + "step": 826 + }, + { + "epoch": 0.6415826221877424, + "grad_norm": 1.0478379726409912, + "learning_rate": 3.388954171562867e-05, + "loss": 0.0138, + "step": 827 + }, + { + "epoch": 0.6423584173778123, + "grad_norm": 7.231791019439697, + "learning_rate": 3.393066980023501e-05, + "loss": 0.3117, + "step": 828 + }, + { + "epoch": 0.6431342125678821, + "grad_norm": 7.956910610198975, + "learning_rate": 3.3971797884841365e-05, + "loss": 0.32, + "step": 829 + }, + { + "epoch": 0.6439100077579519, + "grad_norm": 10.23410415649414, + "learning_rate": 3.40129259694477e-05, + "loss": 0.6437, + "step": 830 + }, + { + "epoch": 0.6446858029480217, + "grad_norm": 12.386017799377441, + "learning_rate": 3.4054054054054055e-05, + "loss": 1.3787, + "step": 831 + }, + { + "epoch": 0.6454615981380916, + "grad_norm": 13.269991874694824, + "learning_rate": 3.4095182138660394e-05, + "loss": 1.1593, + "step": 832 + }, + { + "epoch": 0.6462373933281613, + "grad_norm": 7.184254169464111, + "learning_rate": 3.413631022326674e-05, + "loss": 0.3902, + "step": 833 + }, + { + "epoch": 0.6470131885182312, + "grad_norm": 9.21146297454834, + "learning_rate": 3.417743830787309e-05, + "loss": 0.4275, + "step": 834 + }, + { + "epoch": 0.647788983708301, + "grad_norm": 2.9569036960601807, + "learning_rate": 3.421856639247943e-05, + "loss": 0.087, + "step": 835 + }, + { + "epoch": 0.6485647788983708, + "grad_norm": 9.718716621398926, + "learning_rate": 3.425969447708578e-05, + "loss": 0.4947, + "step": 836 + }, + { + "epoch": 0.6493405740884407, + "grad_norm": 10.130070686340332, + "learning_rate": 3.430082256169213e-05, + "loss": 0.6006, + "step": 837 + }, + { + "epoch": 0.6501163692785105, + "grad_norm": 7.984280586242676, + "learning_rate": 3.4341950646298466e-05, + "loss": 0.3932, + "step": 838 + }, + { + "epoch": 0.6508921644685803, + "grad_norm": 13.934252738952637, + "learning_rate": 3.438307873090482e-05, + "loss": 1.2202, + "step": 839 + }, + { + "epoch": 0.6516679596586501, + "grad_norm": 11.00302505493164, + "learning_rate": 3.4424206815511156e-05, + "loss": 1.5607, + "step": 840 + }, + { + "epoch": 0.6516679596586501, + "eval_Qnli-dev_cosine_accuracy": 0.67578125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8505039215087891, + "eval_Qnli-dev_cosine_ap": 0.7119426569877745, + "eval_Qnli-dev_cosine_f1": 0.6843971631205673, + "eval_Qnli-dev_cosine_f1_threshold": 0.7890456914901733, + "eval_Qnli-dev_cosine_precision": 0.5884146341463414, + "eval_Qnli-dev_cosine_recall": 0.8177966101694916, + "eval_Qnli-dev_dot_accuracy": 0.650390625, + "eval_Qnli-dev_dot_accuracy_threshold": 425.3107604980469, + "eval_Qnli-dev_dot_ap": 0.6269845366138903, + "eval_Qnli-dev_dot_f1": 0.6686746987951808, + "eval_Qnli-dev_dot_f1_threshold": 347.7598571777344, + "eval_Qnli-dev_dot_precision": 0.5186915887850467, + "eval_Qnli-dev_dot_recall": 0.940677966101695, + "eval_Qnli-dev_euclidean_accuracy": 0.66796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.454263687133789, + "eval_Qnli-dev_euclidean_ap": 0.7074799488506723, + "eval_Qnli-dev_euclidean_f1": 0.6893039049235994, + "eval_Qnli-dev_euclidean_f1_threshold": 15.199814796447754, + "eval_Qnli-dev_euclidean_precision": 0.5750708215297451, + "eval_Qnli-dev_euclidean_recall": 0.8601694915254238, + "eval_Qnli-dev_manhattan_accuracy": 0.6796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 218.44036865234375, + "eval_Qnli-dev_manhattan_ap": 0.7004108438466783, + "eval_Qnli-dev_manhattan_f1": 0.6884176182707994, + "eval_Qnli-dev_manhattan_f1_threshold": 279.2939758300781, + "eval_Qnli-dev_manhattan_precision": 0.5596816976127321, + "eval_Qnli-dev_manhattan_recall": 0.8940677966101694, + "eval_Qnli-dev_max_accuracy": 0.6796875, + "eval_Qnli-dev_max_accuracy_threshold": 425.3107604980469, + "eval_Qnli-dev_max_ap": 0.7119426569877745, + "eval_Qnli-dev_max_f1": 0.6893039049235994, + "eval_Qnli-dev_max_f1_threshold": 347.7598571777344, + "eval_Qnli-dev_max_precision": 0.5884146341463414, + "eval_Qnli-dev_max_recall": 0.940677966101695, + "eval_allNLI-dev_cosine_accuracy": 0.70703125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8842862844467163, + "eval_allNLI-dev_cosine_ap": 0.59212475826102, + "eval_allNLI-dev_cosine_f1": 0.6048565121412803, + "eval_allNLI-dev_cosine_f1_threshold": 0.812332034111023, + "eval_allNLI-dev_cosine_precision": 0.48928571428571427, + "eval_allNLI-dev_cosine_recall": 0.791907514450867, + "eval_allNLI-dev_dot_accuracy": 0.677734375, + "eval_allNLI-dev_dot_accuracy_threshold": 451.3145751953125, + "eval_allNLI-dev_dot_ap": 0.4639064297219187, + "eval_allNLI-dev_dot_f1": 0.580896686159844, + "eval_allNLI-dev_dot_f1_threshold": 352.40753173828125, + "eval_allNLI-dev_dot_precision": 0.43823529411764706, + "eval_allNLI-dev_dot_recall": 0.861271676300578, + "eval_allNLI-dev_euclidean_accuracy": 0.720703125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.022924423217773, + "eval_allNLI-dev_euclidean_ap": 0.5993217762757014, + "eval_allNLI-dev_euclidean_f1": 0.6179245283018867, + "eval_allNLI-dev_euclidean_f1_threshold": 12.640939712524414, + "eval_allNLI-dev_euclidean_precision": 0.5219123505976095, + "eval_allNLI-dev_euclidean_recall": 0.7572254335260116, + "eval_allNLI-dev_manhattan_accuracy": 0.720703125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 186.68746948242188, + "eval_allNLI-dev_manhattan_ap": 0.602591177953656, + "eval_allNLI-dev_manhattan_f1": 0.6269315673289184, + "eval_allNLI-dev_manhattan_f1_threshold": 235.98388671875, + "eval_allNLI-dev_manhattan_precision": 0.5071428571428571, + "eval_allNLI-dev_manhattan_recall": 0.8208092485549133, + "eval_allNLI-dev_max_accuracy": 0.720703125, + "eval_allNLI-dev_max_accuracy_threshold": 451.3145751953125, + "eval_allNLI-dev_max_ap": 0.602591177953656, + "eval_allNLI-dev_max_f1": 0.6269315673289184, + "eval_allNLI-dev_max_f1_threshold": 352.40753173828125, + "eval_allNLI-dev_max_precision": 0.5219123505976095, + "eval_allNLI-dev_max_recall": 0.861271676300578, + "eval_sequential_score": 0.7119426569877745, + "eval_sts-test_pearson_cosine": 0.8056133897321561, + "eval_sts-test_pearson_dot": 0.7788407478193802, + "eval_sts-test_pearson_euclidean": 0.8292731923559498, + "eval_sts-test_pearson_manhattan": 0.8265002974859611, + "eval_sts-test_pearson_max": 0.8292731923559498, + "eval_sts-test_spearman_cosine": 0.8339219897779575, + "eval_sts-test_spearman_dot": 0.7616930958821476, + "eval_sts-test_spearman_euclidean": 0.8240892873763006, + "eval_sts-test_spearman_manhattan": 0.8228119968736672, + "eval_sts-test_spearman_max": 0.8339219897779575, + "eval_vitaminc-pairs_loss": 3.695993423461914, + "eval_vitaminc-pairs_runtime": 5.5055, + "eval_vitaminc-pairs_samples_per_second": 23.249, + "eval_vitaminc-pairs_steps_per_second": 0.363, + "step": 840 + }, + { + "epoch": 0.6516679596586501, + "eval_negation-triplets_loss": 1.801641821861267, + "eval_negation-triplets_runtime": 1.1276, + "eval_negation-triplets_samples_per_second": 113.52, + "eval_negation-triplets_steps_per_second": 1.774, + "step": 840 + }, + { + "epoch": 0.6516679596586501, + "eval_scitail-pairs-pos_loss": 0.158038929104805, + "eval_scitail-pairs-pos_runtime": 1.3045, + "eval_scitail-pairs-pos_samples_per_second": 98.126, + "eval_scitail-pairs-pos_steps_per_second": 1.533, + "step": 840 + }, + { + "epoch": 0.6516679596586501, + "eval_scitail-pairs-qa_loss": 0.03073505125939846, + "eval_scitail-pairs-qa_runtime": 0.782, + "eval_scitail-pairs-qa_samples_per_second": 163.682, + "eval_scitail-pairs-qa_steps_per_second": 2.558, + "step": 840 + }, + { + "epoch": 0.6516679596586501, + "eval_xsum-pairs_loss": 0.39707228541374207, + "eval_xsum-pairs_runtime": 3.9314, + "eval_xsum-pairs_samples_per_second": 32.558, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 840 + }, + { + "epoch": 0.6516679596586501, + "eval_sciq_pairs_loss": 0.06643799692392349, + "eval_sciq_pairs_runtime": 6.0572, + "eval_sciq_pairs_samples_per_second": 21.132, + "eval_sciq_pairs_steps_per_second": 0.33, + "step": 840 + }, + { + "epoch": 0.6516679596586501, + "eval_qasc_pairs_loss": 0.6159921288490295, + "eval_qasc_pairs_runtime": 0.9365, + "eval_qasc_pairs_samples_per_second": 136.683, + "eval_qasc_pairs_steps_per_second": 2.136, + "step": 840 + }, + { + "epoch": 0.6516679596586501, + "eval_openbookqa_pairs_loss": 1.4330188035964966, + "eval_openbookqa_pairs_runtime": 0.8645, + "eval_openbookqa_pairs_samples_per_second": 148.06, + "eval_openbookqa_pairs_steps_per_second": 2.313, + "step": 840 + }, + { + "epoch": 0.6516679596586501, + "eval_msmarco_pairs_loss": 0.627158522605896, + "eval_msmarco_pairs_runtime": 1.9688, + "eval_msmarco_pairs_samples_per_second": 65.014, + "eval_msmarco_pairs_steps_per_second": 1.016, + "step": 840 + }, + { + "epoch": 0.6516679596586501, + "eval_nq_pairs_loss": 1.007352590560913, + "eval_nq_pairs_runtime": 3.3458, + "eval_nq_pairs_samples_per_second": 38.256, + "eval_nq_pairs_steps_per_second": 0.598, + "step": 840 + }, + { + "epoch": 0.6516679596586501, + "eval_trivia_pairs_loss": 1.168574333190918, + "eval_trivia_pairs_runtime": 4.4881, + "eval_trivia_pairs_samples_per_second": 28.52, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 840 + }, + { + "epoch": 0.6516679596586501, + "eval_gooaq_pairs_loss": 1.0054221153259277, + "eval_gooaq_pairs_runtime": 1.478, + "eval_gooaq_pairs_samples_per_second": 86.604, + "eval_gooaq_pairs_steps_per_second": 1.353, + "step": 840 + }, + { + "epoch": 0.6516679596586501, + "eval_paws-pos_loss": 0.044977836310863495, + "eval_paws-pos_runtime": 1.0436, + "eval_paws-pos_samples_per_second": 122.654, + "eval_paws-pos_steps_per_second": 1.916, + "step": 840 + }, + { + "epoch": 0.6516679596586501, + "eval_global_dataset_loss": 0.49039849638938904, + "eval_global_dataset_runtime": 19.427, + "eval_global_dataset_samples_per_second": 21.414, + "eval_global_dataset_steps_per_second": 0.36, + "step": 840 + }, + { + "epoch": 0.65244375484872, + "grad_norm": 0.8065465688705444, + "learning_rate": 3.446533490011751e-05, + "loss": 0.053, + "step": 841 + }, + { + "epoch": 0.6532195500387897, + "grad_norm": 13.056890487670898, + "learning_rate": 3.4506462984723853e-05, + "loss": 1.4751, + "step": 842 + }, + { + "epoch": 0.6539953452288596, + "grad_norm": 11.409846305847168, + "learning_rate": 3.454759106933019e-05, + "loss": 1.2965, + "step": 843 + }, + { + "epoch": 0.6547711404189294, + "grad_norm": 12.87276554107666, + "learning_rate": 3.4588719153936544e-05, + "loss": 1.2676, + "step": 844 + }, + { + "epoch": 0.6555469356089992, + "grad_norm": 16.317293167114258, + "learning_rate": 3.462984723854289e-05, + "loss": 2.5868, + "step": 845 + }, + { + "epoch": 0.656322730799069, + "grad_norm": 11.126489639282227, + "learning_rate": 3.4670975323149235e-05, + "loss": 1.1625, + "step": 846 + }, + { + "epoch": 0.6570985259891389, + "grad_norm": 9.192584037780762, + "learning_rate": 3.471210340775558e-05, + "loss": 0.4887, + "step": 847 + }, + { + "epoch": 0.6578743211792087, + "grad_norm": 9.569890022277832, + "learning_rate": 3.4753231492361925e-05, + "loss": 1.0127, + "step": 848 + }, + { + "epoch": 0.6586501163692785, + "grad_norm": 11.920380592346191, + "learning_rate": 3.479435957696827e-05, + "loss": 1.4533, + "step": 849 + }, + { + "epoch": 0.6594259115593484, + "grad_norm": 10.317612648010254, + "learning_rate": 3.4835487661574616e-05, + "loss": 0.8889, + "step": 850 + }, + { + "epoch": 0.6602017067494181, + "grad_norm": 10.163758277893066, + "learning_rate": 3.487661574618096e-05, + "loss": 1.0069, + "step": 851 + }, + { + "epoch": 0.660977501939488, + "grad_norm": 6.893292427062988, + "learning_rate": 3.4917743830787306e-05, + "loss": 0.4064, + "step": 852 + }, + { + "epoch": 0.6617532971295578, + "grad_norm": 5.340035438537598, + "learning_rate": 3.495887191539365e-05, + "loss": 0.2501, + "step": 853 + }, + { + "epoch": 0.6625290923196276, + "grad_norm": 8.26956844329834, + "learning_rate": 3.5e-05, + "loss": 0.4212, + "step": 854 + }, + { + "epoch": 0.6633048875096974, + "grad_norm": 7.952808856964111, + "learning_rate": 3.499997394053394e-05, + "loss": 0.4115, + "step": 855 + }, + { + "epoch": 0.6640806826997673, + "grad_norm": 9.371671676635742, + "learning_rate": 3.499989576222202e-05, + "loss": 0.3841, + "step": 856 + }, + { + "epoch": 0.664856477889837, + "grad_norm": 12.57394027709961, + "learning_rate": 3.499976546532293e-05, + "loss": 0.9916, + "step": 857 + }, + { + "epoch": 0.6656322730799069, + "grad_norm": 2.8098905086517334, + "learning_rate": 3.4999583050267844e-05, + "loss": 0.0504, + "step": 858 + }, + { + "epoch": 0.6664080682699767, + "grad_norm": 7.520341396331787, + "learning_rate": 3.49993485176604e-05, + "loss": 0.3424, + "step": 859 + }, + { + "epoch": 0.6671838634600465, + "grad_norm": 14.273913383483887, + "learning_rate": 3.499906186827669e-05, + "loss": 1.2562, + "step": 860 + }, + { + "epoch": 0.6671838634600465, + "eval_Qnli-dev_cosine_accuracy": 0.68359375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8516297936439514, + "eval_Qnli-dev_cosine_ap": 0.7057068833691326, + "eval_Qnli-dev_cosine_f1": 0.6788432267884322, + "eval_Qnli-dev_cosine_f1_threshold": 0.6823035478591919, + "eval_Qnli-dev_cosine_precision": 0.5296912114014252, + "eval_Qnli-dev_cosine_recall": 0.9449152542372882, + "eval_Qnli-dev_dot_accuracy": 0.650390625, + "eval_Qnli-dev_dot_accuracy_threshold": 422.2347412109375, + "eval_Qnli-dev_dot_ap": 0.6059597079421682, + "eval_Qnli-dev_dot_f1": 0.6576980568011958, + "eval_Qnli-dev_dot_f1_threshold": 331.4080810546875, + "eval_Qnli-dev_dot_precision": 0.5080831408775982, + "eval_Qnli-dev_dot_recall": 0.9322033898305084, + "eval_Qnli-dev_euclidean_accuracy": 0.671875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.281379699707031, + "eval_Qnli-dev_euclidean_ap": 0.704767518205549, + "eval_Qnli-dev_euclidean_f1": 0.6855345911949685, + "eval_Qnli-dev_euclidean_f1_threshold": 17.382638931274414, + "eval_Qnli-dev_euclidean_precision": 0.545, + "eval_Qnli-dev_euclidean_recall": 0.923728813559322, + "eval_Qnli-dev_manhattan_accuracy": 0.673828125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 223.9913330078125, + "eval_Qnli-dev_manhattan_ap": 0.707493799344846, + "eval_Qnli-dev_manhattan_f1": 0.6884735202492211, + "eval_Qnli-dev_manhattan_f1_threshold": 317.806640625, + "eval_Qnli-dev_manhattan_precision": 0.5443349753694581, + "eval_Qnli-dev_manhattan_recall": 0.9364406779661016, + "eval_Qnli-dev_max_accuracy": 0.68359375, + "eval_Qnli-dev_max_accuracy_threshold": 422.2347412109375, + "eval_Qnli-dev_max_ap": 0.707493799344846, + "eval_Qnli-dev_max_f1": 0.6884735202492211, + "eval_Qnli-dev_max_f1_threshold": 331.4080810546875, + "eval_Qnli-dev_max_precision": 0.545, + "eval_Qnli-dev_max_recall": 0.9449152542372882, + "eval_allNLI-dev_cosine_accuracy": 0.708984375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9153469800949097, + "eval_allNLI-dev_cosine_ap": 0.5901974841145107, + "eval_allNLI-dev_cosine_f1": 0.5955734406438632, + "eval_allNLI-dev_cosine_f1_threshold": 0.7628065943717957, + "eval_allNLI-dev_cosine_precision": 0.4567901234567901, + "eval_allNLI-dev_cosine_recall": 0.8554913294797688, + "eval_allNLI-dev_dot_accuracy": 0.6796875, + "eval_allNLI-dev_dot_accuracy_threshold": 432.1881408691406, + "eval_allNLI-dev_dot_ap": 0.4680754372531527, + "eval_allNLI-dev_dot_f1": 0.5595238095238095, + "eval_allNLI-dev_dot_f1_threshold": 336.7421569824219, + "eval_allNLI-dev_dot_precision": 0.4259818731117825, + "eval_allNLI-dev_dot_recall": 0.815028901734104, + "eval_allNLI-dev_euclidean_accuracy": 0.71484375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.299095153808594, + "eval_allNLI-dev_euclidean_ap": 0.5926947177528008, + "eval_allNLI-dev_euclidean_f1": 0.6033755274261603, + "eval_allNLI-dev_euclidean_f1_threshold": 14.113641738891602, + "eval_allNLI-dev_euclidean_precision": 0.4750830564784053, + "eval_allNLI-dev_euclidean_recall": 0.8265895953757225, + "eval_allNLI-dev_manhattan_accuracy": 0.716796875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 189.2566375732422, + "eval_allNLI-dev_manhattan_ap": 0.5955858571526622, + "eval_allNLI-dev_manhattan_f1": 0.606694560669456, + "eval_allNLI-dev_manhattan_f1_threshold": 257.765625, + "eval_allNLI-dev_manhattan_precision": 0.47540983606557374, + "eval_allNLI-dev_manhattan_recall": 0.838150289017341, + "eval_allNLI-dev_max_accuracy": 0.716796875, + "eval_allNLI-dev_max_accuracy_threshold": 432.1881408691406, + "eval_allNLI-dev_max_ap": 0.5955858571526622, + "eval_allNLI-dev_max_f1": 0.606694560669456, + "eval_allNLI-dev_max_f1_threshold": 336.7421569824219, + "eval_allNLI-dev_max_precision": 0.47540983606557374, + "eval_allNLI-dev_max_recall": 0.8554913294797688, + "eval_sequential_score": 0.707493799344846, + "eval_sts-test_pearson_cosine": 0.813022556628032, + "eval_sts-test_pearson_dot": 0.7859908850845266, + "eval_sts-test_pearson_euclidean": 0.8331509505788419, + "eval_sts-test_pearson_manhattan": 0.831921399982832, + "eval_sts-test_pearson_max": 0.8331509505788419, + "eval_sts-test_spearman_cosine": 0.8396164982874904, + "eval_sts-test_spearman_dot": 0.7781822351569454, + "eval_sts-test_spearman_euclidean": 0.8288816104561828, + "eval_sts-test_spearman_manhattan": 0.8279373588408243, + "eval_sts-test_spearman_max": 0.8396164982874904, + "eval_vitaminc-pairs_loss": 3.792736291885376, + "eval_vitaminc-pairs_runtime": 5.5319, + "eval_vitaminc-pairs_samples_per_second": 23.138, + "eval_vitaminc-pairs_steps_per_second": 0.362, + "step": 860 + }, + { + "epoch": 0.6671838634600465, + "eval_negation-triplets_loss": 1.7961204051971436, + "eval_negation-triplets_runtime": 1.0889, + "eval_negation-triplets_samples_per_second": 117.553, + "eval_negation-triplets_steps_per_second": 1.837, + "step": 860 + }, + { + "epoch": 0.6671838634600465, + "eval_scitail-pairs-pos_loss": 0.0998324453830719, + "eval_scitail-pairs-pos_runtime": 1.2829, + "eval_scitail-pairs-pos_samples_per_second": 99.774, + "eval_scitail-pairs-pos_steps_per_second": 1.559, + "step": 860 + }, + { + "epoch": 0.6671838634600465, + "eval_scitail-pairs-qa_loss": 0.012504334561526775, + "eval_scitail-pairs-qa_runtime": 0.7999, + "eval_scitail-pairs-qa_samples_per_second": 160.011, + "eval_scitail-pairs-qa_steps_per_second": 2.5, + "step": 860 + }, + { + "epoch": 0.6671838634600465, + "eval_xsum-pairs_loss": 0.42498546838760376, + "eval_xsum-pairs_runtime": 3.9316, + "eval_xsum-pairs_samples_per_second": 32.556, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 860 + }, + { + "epoch": 0.6671838634600465, + "eval_sciq_pairs_loss": 0.055727776139974594, + "eval_sciq_pairs_runtime": 6.014, + "eval_sciq_pairs_samples_per_second": 21.284, + "eval_sciq_pairs_steps_per_second": 0.333, + "step": 860 + }, + { + "epoch": 0.6671838634600465, + "eval_qasc_pairs_loss": 0.5652861595153809, + "eval_qasc_pairs_runtime": 0.9279, + "eval_qasc_pairs_samples_per_second": 137.953, + "eval_qasc_pairs_steps_per_second": 2.156, + "step": 860 + }, + { + "epoch": 0.6671838634600465, + "eval_openbookqa_pairs_loss": 1.3354136943817139, + "eval_openbookqa_pairs_runtime": 0.8737, + "eval_openbookqa_pairs_samples_per_second": 146.504, + "eval_openbookqa_pairs_steps_per_second": 2.289, + "step": 860 + }, + { + "epoch": 0.6671838634600465, + "eval_msmarco_pairs_loss": 0.6380911469459534, + "eval_msmarco_pairs_runtime": 1.977, + "eval_msmarco_pairs_samples_per_second": 64.745, + "eval_msmarco_pairs_steps_per_second": 1.012, + "step": 860 + }, + { + "epoch": 0.6671838634600465, + "eval_nq_pairs_loss": 1.2092591524124146, + "eval_nq_pairs_runtime": 3.3357, + "eval_nq_pairs_samples_per_second": 38.372, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 860 + }, + { + "epoch": 0.6671838634600465, + "eval_trivia_pairs_loss": 1.2707291841506958, + "eval_trivia_pairs_runtime": 4.4851, + "eval_trivia_pairs_samples_per_second": 28.539, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 860 + }, + { + "epoch": 0.6671838634600465, + "eval_gooaq_pairs_loss": 0.9791943430900574, + "eval_gooaq_pairs_runtime": 1.4633, + "eval_gooaq_pairs_samples_per_second": 87.475, + "eval_gooaq_pairs_steps_per_second": 1.367, + "step": 860 + }, + { + "epoch": 0.6671838634600465, + "eval_paws-pos_loss": 0.04471004381775856, + "eval_paws-pos_runtime": 1.0598, + "eval_paws-pos_samples_per_second": 120.777, + "eval_paws-pos_steps_per_second": 1.887, + "step": 860 + }, + { + "epoch": 0.6671838634600465, + "eval_global_dataset_loss": 0.4852733016014099, + "eval_global_dataset_runtime": 19.4459, + "eval_global_dataset_samples_per_second": 21.393, + "eval_global_dataset_steps_per_second": 0.36, + "step": 860 + }, + { + "epoch": 0.6679596586501164, + "grad_norm": 8.364562034606934, + "learning_rate": 3.499872310306529e-05, + "loss": 0.4189, + "step": 861 + }, + { + "epoch": 0.6687354538401862, + "grad_norm": 14.270318031311035, + "learning_rate": 3.499833222314721e-05, + "loss": 1.168, + "step": 862 + }, + { + "epoch": 0.669511249030256, + "grad_norm": 6.3441853523254395, + "learning_rate": 3.499788922981593e-05, + "loss": 0.3288, + "step": 863 + }, + { + "epoch": 0.6702870442203258, + "grad_norm": 12.241477966308594, + "learning_rate": 3.499739412453738e-05, + "loss": 1.2634, + "step": 864 + }, + { + "epoch": 0.6710628394103957, + "grad_norm": 11.538920402526855, + "learning_rate": 3.499684690894993e-05, + "loss": 0.9304, + "step": 865 + }, + { + "epoch": 0.6718386346004654, + "grad_norm": 7.381165027618408, + "learning_rate": 3.499624758486438e-05, + "loss": 0.3868, + "step": 866 + }, + { + "epoch": 0.6726144297905353, + "grad_norm": 9.375253677368164, + "learning_rate": 3.4995596154263995e-05, + "loss": 0.9083, + "step": 867 + }, + { + "epoch": 0.6733902249806051, + "grad_norm": 8.622047424316406, + "learning_rate": 3.4994892619304444e-05, + "loss": 0.4259, + "step": 868 + }, + { + "epoch": 0.674166020170675, + "grad_norm": 8.653176307678223, + "learning_rate": 3.499413698231381e-05, + "loss": 0.5275, + "step": 869 + }, + { + "epoch": 0.6749418153607447, + "grad_norm": 15.130743026733398, + "learning_rate": 3.499332924579261e-05, + "loss": 1.7172, + "step": 870 + }, + { + "epoch": 0.6757176105508146, + "grad_norm": 8.044624328613281, + "learning_rate": 3.4992469412413765e-05, + "loss": 0.5603, + "step": 871 + }, + { + "epoch": 0.6764934057408845, + "grad_norm": 11.914979934692383, + "learning_rate": 3.4991557485022566e-05, + "loss": 0.8937, + "step": 872 + }, + { + "epoch": 0.6772692009309542, + "grad_norm": 7.6670989990234375, + "learning_rate": 3.499059346663671e-05, + "loss": 0.3739, + "step": 873 + }, + { + "epoch": 0.6780449961210241, + "grad_norm": 9.105637550354004, + "learning_rate": 3.498957736044628e-05, + "loss": 0.4901, + "step": 874 + }, + { + "epoch": 0.6788207913110939, + "grad_norm": 11.809444427490234, + "learning_rate": 3.49885091698137e-05, + "loss": 0.9293, + "step": 875 + }, + { + "epoch": 0.6795965865011637, + "grad_norm": 11.217194557189941, + "learning_rate": 3.498738889827376e-05, + "loss": 1.1831, + "step": 876 + }, + { + "epoch": 0.6803723816912335, + "grad_norm": 11.778733253479004, + "learning_rate": 3.49862165495336e-05, + "loss": 0.9599, + "step": 877 + }, + { + "epoch": 0.6811481768813034, + "grad_norm": 7.655272483825684, + "learning_rate": 3.498499212747269e-05, + "loss": 0.4518, + "step": 878 + }, + { + "epoch": 0.6819239720713731, + "grad_norm": 9.153346061706543, + "learning_rate": 3.49837156361428e-05, + "loss": 0.5798, + "step": 879 + }, + { + "epoch": 0.682699767261443, + "grad_norm": 9.612966537475586, + "learning_rate": 3.4982387079768024e-05, + "loss": 0.6668, + "step": 880 + }, + { + "epoch": 0.682699767261443, + "eval_Qnli-dev_cosine_accuracy": 0.658203125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8739014863967896, + "eval_Qnli-dev_cosine_ap": 0.6832396250223273, + "eval_Qnli-dev_cosine_f1": 0.6656534954407294, + "eval_Qnli-dev_cosine_f1_threshold": 0.7337082624435425, + "eval_Qnli-dev_cosine_precision": 0.518957345971564, + "eval_Qnli-dev_cosine_recall": 0.9279661016949152, + "eval_Qnli-dev_dot_accuracy": 0.615234375, + "eval_Qnli-dev_dot_accuracy_threshold": 483.6365966796875, + "eval_Qnli-dev_dot_ap": 0.5811670879278963, + "eval_Qnli-dev_dot_f1": 0.6567164179104478, + "eval_Qnli-dev_dot_f1_threshold": 356.31573486328125, + "eval_Qnli-dev_dot_precision": 0.5069124423963134, + "eval_Qnli-dev_dot_recall": 0.9322033898305084, + "eval_Qnli-dev_euclidean_accuracy": 0.66015625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.706809043884277, + "eval_Qnli-dev_euclidean_ap": 0.6875489517288256, + "eval_Qnli-dev_euclidean_f1": 0.6709886547811994, + "eval_Qnli-dev_euclidean_f1_threshold": 15.960616111755371, + "eval_Qnli-dev_euclidean_precision": 0.5433070866141733, + "eval_Qnli-dev_euclidean_recall": 0.8771186440677966, + "eval_Qnli-dev_manhattan_accuracy": 0.666015625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 214.6758575439453, + "eval_Qnli-dev_manhattan_ap": 0.6933964909228749, + "eval_Qnli-dev_manhattan_f1": 0.6697965571205008, + "eval_Qnli-dev_manhattan_f1_threshold": 291.2758483886719, + "eval_Qnli-dev_manhattan_precision": 0.5310173697270472, + "eval_Qnli-dev_manhattan_recall": 0.9067796610169492, + "eval_Qnli-dev_max_accuracy": 0.666015625, + "eval_Qnli-dev_max_accuracy_threshold": 483.6365966796875, + "eval_Qnli-dev_max_ap": 0.6933964909228749, + "eval_Qnli-dev_max_f1": 0.6709886547811994, + "eval_Qnli-dev_max_f1_threshold": 356.31573486328125, + "eval_Qnli-dev_max_precision": 0.5433070866141733, + "eval_Qnli-dev_max_recall": 0.9322033898305084, + "eval_allNLI-dev_cosine_accuracy": 0.7109375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9285610318183899, + "eval_allNLI-dev_cosine_ap": 0.5798671366823452, + "eval_allNLI-dev_cosine_f1": 0.584493041749503, + "eval_allNLI-dev_cosine_f1_threshold": 0.7967034578323364, + "eval_allNLI-dev_cosine_precision": 0.44545454545454544, + "eval_allNLI-dev_cosine_recall": 0.8497109826589595, + "eval_allNLI-dev_dot_accuracy": 0.67578125, + "eval_allNLI-dev_dot_accuracy_threshold": 472.23895263671875, + "eval_allNLI-dev_dot_ap": 0.47875534648608414, + "eval_allNLI-dev_dot_f1": 0.562753036437247, + "eval_allNLI-dev_dot_f1_threshold": 382.5263671875, + "eval_allNLI-dev_dot_precision": 0.43302180685358255, + "eval_allNLI-dev_dot_recall": 0.8034682080924855, + "eval_allNLI-dev_euclidean_accuracy": 0.716796875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 9.449457168579102, + "eval_allNLI-dev_euclidean_ap": 0.5816487832325188, + "eval_allNLI-dev_euclidean_f1": 0.5894736842105264, + "eval_allNLI-dev_euclidean_f1_threshold": 13.506660461425781, + "eval_allNLI-dev_euclidean_precision": 0.46357615894039733, + "eval_allNLI-dev_euclidean_recall": 0.8092485549132948, + "eval_allNLI-dev_manhattan_accuracy": 0.716796875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 182.58740234375, + "eval_allNLI-dev_manhattan_ap": 0.5893710904383558, + "eval_allNLI-dev_manhattan_f1": 0.6041666666666666, + "eval_allNLI-dev_manhattan_f1_threshold": 242.03350830078125, + "eval_allNLI-dev_manhattan_precision": 0.4723127035830619, + "eval_allNLI-dev_manhattan_recall": 0.838150289017341, + "eval_allNLI-dev_max_accuracy": 0.716796875, + "eval_allNLI-dev_max_accuracy_threshold": 472.23895263671875, + "eval_allNLI-dev_max_ap": 0.5893710904383558, + "eval_allNLI-dev_max_f1": 0.6041666666666666, + "eval_allNLI-dev_max_f1_threshold": 382.5263671875, + "eval_allNLI-dev_max_precision": 0.4723127035830619, + "eval_allNLI-dev_max_recall": 0.8497109826589595, + "eval_sequential_score": 0.6933964909228749, + "eval_sts-test_pearson_cosine": 0.811044707920606, + "eval_sts-test_pearson_dot": 0.7419008243975282, + "eval_sts-test_pearson_euclidean": 0.8420267935379353, + "eval_sts-test_pearson_manhattan": 0.8394854650030468, + "eval_sts-test_pearson_max": 0.8420267935379353, + "eval_sts-test_spearman_cosine": 0.8440954282838484, + "eval_sts-test_spearman_dot": 0.7264572952004014, + "eval_sts-test_spearman_euclidean": 0.83790896915784, + "eval_sts-test_spearman_manhattan": 0.835837075257941, + "eval_sts-test_spearman_max": 0.8440954282838484, + "eval_vitaminc-pairs_loss": 3.650113821029663, + "eval_vitaminc-pairs_runtime": 5.4944, + "eval_vitaminc-pairs_samples_per_second": 23.296, + "eval_vitaminc-pairs_steps_per_second": 0.364, + "step": 880 + }, + { + "epoch": 0.682699767261443, + "eval_negation-triplets_loss": 1.7521673440933228, + "eval_negation-triplets_runtime": 1.0917, + "eval_negation-triplets_samples_per_second": 117.248, + "eval_negation-triplets_steps_per_second": 1.832, + "step": 880 + }, + { + "epoch": 0.682699767261443, + "eval_scitail-pairs-pos_loss": 0.15876206755638123, + "eval_scitail-pairs-pos_runtime": 1.2972, + "eval_scitail-pairs-pos_samples_per_second": 98.674, + "eval_scitail-pairs-pos_steps_per_second": 1.542, + "step": 880 + }, + { + "epoch": 0.682699767261443, + "eval_scitail-pairs-qa_loss": 0.03474993258714676, + "eval_scitail-pairs-qa_runtime": 0.7908, + "eval_scitail-pairs-qa_samples_per_second": 161.857, + "eval_scitail-pairs-qa_steps_per_second": 2.529, + "step": 880 + }, + { + "epoch": 0.682699767261443, + "eval_xsum-pairs_loss": 0.403542160987854, + "eval_xsum-pairs_runtime": 3.9373, + "eval_xsum-pairs_samples_per_second": 32.509, + "eval_xsum-pairs_steps_per_second": 0.508, + "step": 880 + }, + { + "epoch": 0.682699767261443, + "eval_sciq_pairs_loss": 0.059477634727954865, + "eval_sciq_pairs_runtime": 6.0628, + "eval_sciq_pairs_samples_per_second": 21.112, + "eval_sciq_pairs_steps_per_second": 0.33, + "step": 880 + }, + { + "epoch": 0.682699767261443, + "eval_qasc_pairs_loss": 0.7362544536590576, + "eval_qasc_pairs_runtime": 0.9322, + "eval_qasc_pairs_samples_per_second": 137.307, + "eval_qasc_pairs_steps_per_second": 2.145, + "step": 880 + }, + { + "epoch": 0.682699767261443, + "eval_openbookqa_pairs_loss": 1.4471515417099, + "eval_openbookqa_pairs_runtime": 0.885, + "eval_openbookqa_pairs_samples_per_second": 144.626, + "eval_openbookqa_pairs_steps_per_second": 2.26, + "step": 880 + }, + { + "epoch": 0.682699767261443, + "eval_msmarco_pairs_loss": 0.6078033447265625, + "eval_msmarco_pairs_runtime": 1.9891, + "eval_msmarco_pairs_samples_per_second": 64.351, + "eval_msmarco_pairs_steps_per_second": 1.005, + "step": 880 + }, + { + "epoch": 0.682699767261443, + "eval_nq_pairs_loss": 1.0600759983062744, + "eval_nq_pairs_runtime": 3.3403, + "eval_nq_pairs_samples_per_second": 38.319, + "eval_nq_pairs_steps_per_second": 0.599, + "step": 880 + }, + { + "epoch": 0.682699767261443, + "eval_trivia_pairs_loss": 1.1039668321609497, + "eval_trivia_pairs_runtime": 4.4847, + "eval_trivia_pairs_samples_per_second": 28.541, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 880 + }, + { + "epoch": 0.682699767261443, + "eval_gooaq_pairs_loss": 0.7967941761016846, + "eval_gooaq_pairs_runtime": 1.4628, + "eval_gooaq_pairs_samples_per_second": 87.502, + "eval_gooaq_pairs_steps_per_second": 1.367, + "step": 880 + }, + { + "epoch": 0.682699767261443, + "eval_paws-pos_loss": 0.04289524629712105, + "eval_paws-pos_runtime": 1.0689, + "eval_paws-pos_samples_per_second": 119.744, + "eval_paws-pos_steps_per_second": 1.871, + "step": 880 + }, + { + "epoch": 0.682699767261443, + "eval_global_dataset_loss": 0.47498950362205505, + "eval_global_dataset_runtime": 19.4168, + "eval_global_dataset_samples_per_second": 21.425, + "eval_global_dataset_steps_per_second": 0.361, + "step": 880 + }, + { + "epoch": 0.6834755624515128, + "grad_norm": 8.79952621459961, + "learning_rate": 3.498100646274474e-05, + "loss": 0.4126, + "step": 881 + }, + { + "epoch": 0.6842513576415826, + "grad_norm": 6.2995781898498535, + "learning_rate": 3.4979573789641585e-05, + "loss": 0.2988, + "step": 882 + }, + { + "epoch": 0.6850271528316525, + "grad_norm": 4.987382411956787, + "learning_rate": 3.4978089065199493e-05, + "loss": 0.3833, + "step": 883 + }, + { + "epoch": 0.6858029480217223, + "grad_norm": 10.842425346374512, + "learning_rate": 3.497655229433162e-05, + "loss": 1.1883, + "step": 884 + }, + { + "epoch": 0.6865787432117921, + "grad_norm": 9.746257781982422, + "learning_rate": 3.497496348212334e-05, + "loss": 0.5292, + "step": 885 + }, + { + "epoch": 0.6873545384018619, + "grad_norm": 12.727143287658691, + "learning_rate": 3.497332263383227e-05, + "loss": 1.2004, + "step": 886 + }, + { + "epoch": 0.6881303335919318, + "grad_norm": 14.975640296936035, + "learning_rate": 3.4971629754888185e-05, + "loss": 1.2384, + "step": 887 + }, + { + "epoch": 0.6889061287820015, + "grad_norm": 9.211393356323242, + "learning_rate": 3.4969884850893075e-05, + "loss": 0.9863, + "step": 888 + }, + { + "epoch": 0.6896819239720714, + "grad_norm": 8.209000587463379, + "learning_rate": 3.4968087927621064e-05, + "loss": 0.4061, + "step": 889 + }, + { + "epoch": 0.6904577191621412, + "grad_norm": 8.929417610168457, + "learning_rate": 3.496623899101841e-05, + "loss": 0.5534, + "step": 890 + }, + { + "epoch": 0.691233514352211, + "grad_norm": 18.9887638092041, + "learning_rate": 3.496433804720351e-05, + "loss": 2.9585, + "step": 891 + }, + { + "epoch": 0.6920093095422808, + "grad_norm": 13.784379959106445, + "learning_rate": 3.4962385102466844e-05, + "loss": 1.2675, + "step": 892 + }, + { + "epoch": 0.6927851047323507, + "grad_norm": 1.0738762617111206, + "learning_rate": 3.496038016327099e-05, + "loss": 0.0425, + "step": 893 + }, + { + "epoch": 0.6935608999224204, + "grad_norm": 8.122675895690918, + "learning_rate": 3.495832323625056e-05, + "loss": 0.5381, + "step": 894 + }, + { + "epoch": 0.6943366951124903, + "grad_norm": 10.782946586608887, + "learning_rate": 3.49562143282122e-05, + "loss": 1.0622, + "step": 895 + }, + { + "epoch": 0.6951124903025602, + "grad_norm": 10.662599563598633, + "learning_rate": 3.49540534461346e-05, + "loss": 0.9062, + "step": 896 + }, + { + "epoch": 0.6958882854926299, + "grad_norm": 7.868858337402344, + "learning_rate": 3.49518405971684e-05, + "loss": 0.3983, + "step": 897 + }, + { + "epoch": 0.6966640806826998, + "grad_norm": 9.266629219055176, + "learning_rate": 3.494957578863623e-05, + "loss": 0.5936, + "step": 898 + }, + { + "epoch": 0.6974398758727696, + "grad_norm": 12.59846305847168, + "learning_rate": 3.494725902803267e-05, + "loss": 1.2099, + "step": 899 + }, + { + "epoch": 0.6982156710628394, + "grad_norm": 13.477904319763184, + "learning_rate": 3.494489032302418e-05, + "loss": 1.3554, + "step": 900 + }, + { + "epoch": 0.6982156710628394, + "eval_Qnli-dev_cosine_accuracy": 0.677734375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8567171096801758, + "eval_Qnli-dev_cosine_ap": 0.7094826708729062, + "eval_Qnli-dev_cosine_f1": 0.6752411575562701, + "eval_Qnli-dev_cosine_f1_threshold": 0.764941930770874, + "eval_Qnli-dev_cosine_precision": 0.5440414507772021, + "eval_Qnli-dev_cosine_recall": 0.8898305084745762, + "eval_Qnli-dev_dot_accuracy": 0.630859375, + "eval_Qnli-dev_dot_accuracy_threshold": 486.6195068359375, + "eval_Qnli-dev_dot_ap": 0.5995735525894889, + "eval_Qnli-dev_dot_f1": 0.6495176848874598, + "eval_Qnli-dev_dot_f1_threshold": 397.50433349609375, + "eval_Qnli-dev_dot_precision": 0.5233160621761658, + "eval_Qnli-dev_dot_recall": 0.8559322033898306, + "eval_Qnli-dev_euclidean_accuracy": 0.673828125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.704582214355469, + "eval_Qnli-dev_euclidean_ap": 0.7076896940503825, + "eval_Qnli-dev_euclidean_f1": 0.6833333333333335, + "eval_Qnli-dev_euclidean_f1_threshold": 15.329047203063965, + "eval_Qnli-dev_euclidean_precision": 0.5631868131868132, + "eval_Qnli-dev_euclidean_recall": 0.8686440677966102, + "eval_Qnli-dev_manhattan_accuracy": 0.671875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 224.72830200195312, + "eval_Qnli-dev_manhattan_ap": 0.7062101616210243, + "eval_Qnli-dev_manhattan_f1": 0.6829268292682927, + "eval_Qnli-dev_manhattan_f1_threshold": 276.5677795410156, + "eval_Qnli-dev_manhattan_precision": 0.554089709762533, + "eval_Qnli-dev_manhattan_recall": 0.8898305084745762, + "eval_Qnli-dev_max_accuracy": 0.677734375, + "eval_Qnli-dev_max_accuracy_threshold": 486.6195068359375, + "eval_Qnli-dev_max_ap": 0.7094826708729062, + "eval_Qnli-dev_max_f1": 0.6833333333333335, + "eval_Qnli-dev_max_f1_threshold": 397.50433349609375, + "eval_Qnli-dev_max_precision": 0.5631868131868132, + "eval_Qnli-dev_max_recall": 0.8898305084745762, + "eval_allNLI-dev_cosine_accuracy": 0.71484375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8934484720230103, + "eval_allNLI-dev_cosine_ap": 0.5832941968478464, + "eval_allNLI-dev_cosine_f1": 0.5944700460829493, + "eval_allNLI-dev_cosine_f1_threshold": 0.8264994621276855, + "eval_allNLI-dev_cosine_precision": 0.4942528735632184, + "eval_allNLI-dev_cosine_recall": 0.7456647398843931, + "eval_allNLI-dev_dot_accuracy": 0.67578125, + "eval_allNLI-dev_dot_accuracy_threshold": 515.9781494140625, + "eval_allNLI-dev_dot_ap": 0.47975114675727787, + "eval_allNLI-dev_dot_f1": 0.5664739884393063, + "eval_allNLI-dev_dot_f1_threshold": 363.66107177734375, + "eval_allNLI-dev_dot_precision": 0.42485549132947975, + "eval_allNLI-dev_dot_recall": 0.8497109826589595, + "eval_allNLI-dev_euclidean_accuracy": 0.712890625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.550495147705078, + "eval_allNLI-dev_euclidean_ap": 0.5838831241824369, + "eval_allNLI-dev_euclidean_f1": 0.5986696230598669, + "eval_allNLI-dev_euclidean_f1_threshold": 13.402144432067871, + "eval_allNLI-dev_euclidean_precision": 0.4856115107913669, + "eval_allNLI-dev_euclidean_recall": 0.7803468208092486, + "eval_allNLI-dev_manhattan_accuracy": 0.720703125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 186.66268920898438, + "eval_allNLI-dev_manhattan_ap": 0.5907220466661656, + "eval_allNLI-dev_manhattan_f1": 0.6060606060606061, + "eval_allNLI-dev_manhattan_f1_threshold": 230.32687377929688, + "eval_allNLI-dev_manhattan_precision": 0.5078125, + "eval_allNLI-dev_manhattan_recall": 0.7514450867052023, + "eval_allNLI-dev_max_accuracy": 0.720703125, + "eval_allNLI-dev_max_accuracy_threshold": 515.9781494140625, + "eval_allNLI-dev_max_ap": 0.5907220466661656, + "eval_allNLI-dev_max_f1": 0.6060606060606061, + "eval_allNLI-dev_max_f1_threshold": 363.66107177734375, + "eval_allNLI-dev_max_precision": 0.5078125, + "eval_allNLI-dev_max_recall": 0.8497109826589595, + "eval_sequential_score": 0.7094826708729062, + "eval_sts-test_pearson_cosine": 0.8204430236461935, + "eval_sts-test_pearson_dot": 0.7577759507611783, + "eval_sts-test_pearson_euclidean": 0.8432313462757036, + "eval_sts-test_pearson_manhattan": 0.840958244747654, + "eval_sts-test_pearson_max": 0.8432313462757036, + "eval_sts-test_spearman_cosine": 0.8474270254406541, + "eval_sts-test_spearman_dot": 0.7465720474193461, + "eval_sts-test_spearman_euclidean": 0.8376924625552284, + "eval_sts-test_spearman_manhattan": 0.836714602742565, + "eval_sts-test_spearman_max": 0.8474270254406541, + "eval_vitaminc-pairs_loss": 3.7364420890808105, + "eval_vitaminc-pairs_runtime": 5.5123, + "eval_vitaminc-pairs_samples_per_second": 23.221, + "eval_vitaminc-pairs_steps_per_second": 0.363, + "step": 900 + }, + { + "epoch": 0.6982156710628394, + "eval_negation-triplets_loss": 1.7152256965637207, + "eval_negation-triplets_runtime": 1.0868, + "eval_negation-triplets_samples_per_second": 117.773, + "eval_negation-triplets_steps_per_second": 1.84, + "step": 900 + }, + { + "epoch": 0.6982156710628394, + "eval_scitail-pairs-pos_loss": 0.1542925089597702, + "eval_scitail-pairs-pos_runtime": 1.3113, + "eval_scitail-pairs-pos_samples_per_second": 97.616, + "eval_scitail-pairs-pos_steps_per_second": 1.525, + "step": 900 + }, + { + "epoch": 0.6982156710628394, + "eval_scitail-pairs-qa_loss": 0.031740859150886536, + "eval_scitail-pairs-qa_runtime": 0.78, + "eval_scitail-pairs-qa_samples_per_second": 164.105, + "eval_scitail-pairs-qa_steps_per_second": 2.564, + "step": 900 + }, + { + "epoch": 0.6982156710628394, + "eval_xsum-pairs_loss": 0.33082854747772217, + "eval_xsum-pairs_runtime": 3.9293, + "eval_xsum-pairs_samples_per_second": 32.575, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 900 + }, + { + "epoch": 0.6982156710628394, + "eval_sciq_pairs_loss": 0.05822315067052841, + "eval_sciq_pairs_runtime": 6.0813, + "eval_sciq_pairs_samples_per_second": 21.048, + "eval_sciq_pairs_steps_per_second": 0.329, + "step": 900 + }, + { + "epoch": 0.6982156710628394, + "eval_qasc_pairs_loss": 0.5999854207038879, + "eval_qasc_pairs_runtime": 0.9353, + "eval_qasc_pairs_samples_per_second": 136.853, + "eval_qasc_pairs_steps_per_second": 2.138, + "step": 900 + }, + { + "epoch": 0.6982156710628394, + "eval_openbookqa_pairs_loss": 1.3743079900741577, + "eval_openbookqa_pairs_runtime": 0.8899, + "eval_openbookqa_pairs_samples_per_second": 143.831, + "eval_openbookqa_pairs_steps_per_second": 2.247, + "step": 900 + }, + { + "epoch": 0.6982156710628394, + "eval_msmarco_pairs_loss": 0.5703839659690857, + "eval_msmarco_pairs_runtime": 1.9802, + "eval_msmarco_pairs_samples_per_second": 64.641, + "eval_msmarco_pairs_steps_per_second": 1.01, + "step": 900 + }, + { + "epoch": 0.6982156710628394, + "eval_nq_pairs_loss": 1.066963791847229, + "eval_nq_pairs_runtime": 3.3495, + "eval_nq_pairs_samples_per_second": 38.215, + "eval_nq_pairs_steps_per_second": 0.597, + "step": 900 + }, + { + "epoch": 0.6982156710628394, + "eval_trivia_pairs_loss": 0.995316743850708, + "eval_trivia_pairs_runtime": 4.4938, + "eval_trivia_pairs_samples_per_second": 28.484, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 900 + }, + { + "epoch": 0.6982156710628394, + "eval_gooaq_pairs_loss": 0.8502556085586548, + "eval_gooaq_pairs_runtime": 1.4839, + "eval_gooaq_pairs_samples_per_second": 86.261, + "eval_gooaq_pairs_steps_per_second": 1.348, + "step": 900 + }, + { + "epoch": 0.6982156710628394, + "eval_paws-pos_loss": 0.042505402117967606, + "eval_paws-pos_runtime": 1.048, + "eval_paws-pos_samples_per_second": 122.132, + "eval_paws-pos_steps_per_second": 1.908, + "step": 900 + }, + { + "epoch": 0.6982156710628394, + "eval_global_dataset_loss": 0.4473607838153839, + "eval_global_dataset_runtime": 19.4498, + "eval_global_dataset_samples_per_second": 21.388, + "eval_global_dataset_steps_per_second": 0.36, + "step": 900 + }, + { + "epoch": 0.6989914662529092, + "grad_norm": 4.887351036071777, + "learning_rate": 3.494246968144915e-05, + "loss": 0.1761, + "step": 901 + }, + { + "epoch": 0.6997672614429791, + "grad_norm": 7.084521770477295, + "learning_rate": 3.4939997111317794e-05, + "loss": 0.3289, + "step": 902 + }, + { + "epoch": 0.7005430566330488, + "grad_norm": 9.108835220336914, + "learning_rate": 3.493747262081222e-05, + "loss": 0.64, + "step": 903 + }, + { + "epoch": 0.7013188518231187, + "grad_norm": 8.509134292602539, + "learning_rate": 3.493489621828629e-05, + "loss": 0.4796, + "step": 904 + }, + { + "epoch": 0.7020946470131885, + "grad_norm": 7.357263088226318, + "learning_rate": 3.49322679122657e-05, + "loss": 0.3145, + "step": 905 + }, + { + "epoch": 0.7028704422032583, + "grad_norm": 7.948583126068115, + "learning_rate": 3.4929587711447865e-05, + "loss": 0.3716, + "step": 906 + }, + { + "epoch": 0.7036462373933282, + "grad_norm": 11.190608024597168, + "learning_rate": 3.492685562470194e-05, + "loss": 0.9676, + "step": 907 + }, + { + "epoch": 0.704422032583398, + "grad_norm": 8.347633361816406, + "learning_rate": 3.492407166106878e-05, + "loss": 0.6769, + "step": 908 + }, + { + "epoch": 0.7051978277734678, + "grad_norm": 11.00367546081543, + "learning_rate": 3.492123582976092e-05, + "loss": 0.6006, + "step": 909 + }, + { + "epoch": 0.7059736229635376, + "grad_norm": 9.012520790100098, + "learning_rate": 3.491834814016249e-05, + "loss": 0.3612, + "step": 910 + }, + { + "epoch": 0.7067494181536075, + "grad_norm": 9.403765678405762, + "learning_rate": 3.4915408601829286e-05, + "loss": 0.7262, + "step": 911 + }, + { + "epoch": 0.7075252133436772, + "grad_norm": 9.686113357543945, + "learning_rate": 3.491241722448863e-05, + "loss": 0.4605, + "step": 912 + }, + { + "epoch": 0.7083010085337471, + "grad_norm": 11.69682502746582, + "learning_rate": 3.4909374018039414e-05, + "loss": 1.0258, + "step": 913 + }, + { + "epoch": 0.7090768037238169, + "grad_norm": 8.281636238098145, + "learning_rate": 3.490627899255201e-05, + "loss": 0.3937, + "step": 914 + }, + { + "epoch": 0.7098525989138867, + "grad_norm": 12.542149543762207, + "learning_rate": 3.4903132158268305e-05, + "loss": 1.1565, + "step": 915 + }, + { + "epoch": 0.7106283941039565, + "grad_norm": 14.342009544372559, + "learning_rate": 3.489993352560161e-05, + "loss": 1.2102, + "step": 916 + }, + { + "epoch": 0.7114041892940264, + "grad_norm": 15.103821754455566, + "learning_rate": 3.489668310513663e-05, + "loss": 1.3208, + "step": 917 + }, + { + "epoch": 0.7121799844840963, + "grad_norm": 8.335806846618652, + "learning_rate": 3.4893380907629466e-05, + "loss": 0.4204, + "step": 918 + }, + { + "epoch": 0.712955779674166, + "grad_norm": 7.120835781097412, + "learning_rate": 3.489002694400755e-05, + "loss": 0.3666, + "step": 919 + }, + { + "epoch": 0.7137315748642359, + "grad_norm": 12.769718170166016, + "learning_rate": 3.4886621225369605e-05, + "loss": 1.2682, + "step": 920 + }, + { + "epoch": 0.7137315748642359, + "eval_Qnli-dev_cosine_accuracy": 0.673828125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8663387298583984, + "eval_Qnli-dev_cosine_ap": 0.7011581034275928, + "eval_Qnli-dev_cosine_f1": 0.6677316293929711, + "eval_Qnli-dev_cosine_f1_threshold": 0.7536342740058899, + "eval_Qnli-dev_cosine_precision": 0.5358974358974359, + "eval_Qnli-dev_cosine_recall": 0.885593220338983, + "eval_Qnli-dev_dot_accuracy": 0.64453125, + "eval_Qnli-dev_dot_accuracy_threshold": 442.28472900390625, + "eval_Qnli-dev_dot_ap": 0.6160452811759833, + "eval_Qnli-dev_dot_f1": 0.6559766763848397, + "eval_Qnli-dev_dot_f1_threshold": 343.661376953125, + "eval_Qnli-dev_dot_precision": 0.5, + "eval_Qnli-dev_dot_recall": 0.9533898305084746, + "eval_Qnli-dev_euclidean_accuracy": 0.662109375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.209592819213867, + "eval_Qnli-dev_euclidean_ap": 0.7014673664853852, + "eval_Qnli-dev_euclidean_f1": 0.6763754045307444, + "eval_Qnli-dev_euclidean_f1_threshold": 15.89691162109375, + "eval_Qnli-dev_euclidean_precision": 0.5471204188481675, + "eval_Qnli-dev_euclidean_recall": 0.885593220338983, + "eval_Qnli-dev_manhattan_accuracy": 0.66796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 217.30770874023438, + "eval_Qnli-dev_manhattan_ap": 0.6991581369945994, + "eval_Qnli-dev_manhattan_f1": 0.6798029556650246, + "eval_Qnli-dev_manhattan_f1_threshold": 281.1316833496094, + "eval_Qnli-dev_manhattan_precision": 0.5549597855227882, + "eval_Qnli-dev_manhattan_recall": 0.8771186440677966, + "eval_Qnli-dev_max_accuracy": 0.673828125, + "eval_Qnli-dev_max_accuracy_threshold": 442.28472900390625, + "eval_Qnli-dev_max_ap": 0.7014673664853852, + "eval_Qnli-dev_max_f1": 0.6798029556650246, + "eval_Qnli-dev_max_f1_threshold": 343.661376953125, + "eval_Qnli-dev_max_precision": 0.5549597855227882, + "eval_Qnli-dev_max_recall": 0.9533898305084746, + "eval_allNLI-dev_cosine_accuracy": 0.71484375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8784551620483398, + "eval_allNLI-dev_cosine_ap": 0.5848850964198085, + "eval_allNLI-dev_cosine_f1": 0.5923076923076924, + "eval_allNLI-dev_cosine_f1_threshold": 0.767518162727356, + "eval_allNLI-dev_cosine_precision": 0.4438040345821326, + "eval_allNLI-dev_cosine_recall": 0.8901734104046243, + "eval_allNLI-dev_dot_accuracy": 0.673828125, + "eval_allNLI-dev_dot_accuracy_threshold": 462.3367004394531, + "eval_allNLI-dev_dot_ap": 0.4753203626546576, + "eval_allNLI-dev_dot_f1": 0.5555555555555556, + "eval_allNLI-dev_dot_f1_threshold": 336.5610046386719, + "eval_allNLI-dev_dot_precision": 0.4087193460490463, + "eval_allNLI-dev_dot_recall": 0.8670520231213873, + "eval_allNLI-dev_euclidean_accuracy": 0.712890625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.113510131835938, + "eval_allNLI-dev_euclidean_ap": 0.5863555746391488, + "eval_allNLI-dev_euclidean_f1": 0.6042553191489363, + "eval_allNLI-dev_euclidean_f1_threshold": 13.709831237792969, + "eval_allNLI-dev_euclidean_precision": 0.4781144781144781, + "eval_allNLI-dev_euclidean_recall": 0.8208092485549133, + "eval_allNLI-dev_manhattan_accuracy": 0.7109375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 171.8656005859375, + "eval_allNLI-dev_manhattan_ap": 0.5940185076677719, + "eval_allNLI-dev_manhattan_f1": 0.6059225512528473, + "eval_allNLI-dev_manhattan_f1_threshold": 236.40133666992188, + "eval_allNLI-dev_manhattan_precision": 0.5, + "eval_allNLI-dev_manhattan_recall": 0.7687861271676301, + "eval_allNLI-dev_max_accuracy": 0.71484375, + "eval_allNLI-dev_max_accuracy_threshold": 462.3367004394531, + "eval_allNLI-dev_max_ap": 0.5940185076677719, + "eval_allNLI-dev_max_f1": 0.6059225512528473, + "eval_allNLI-dev_max_f1_threshold": 336.5610046386719, + "eval_allNLI-dev_max_precision": 0.5, + "eval_allNLI-dev_max_recall": 0.8901734104046243, + "eval_sequential_score": 0.7014673664853852, + "eval_sts-test_pearson_cosine": 0.8255116029658587, + "eval_sts-test_pearson_dot": 0.7879340794313661, + "eval_sts-test_pearson_euclidean": 0.8490714081771065, + "eval_sts-test_pearson_manhattan": 0.8455351279667078, + "eval_sts-test_pearson_max": 0.8490714081771065, + "eval_sts-test_spearman_cosine": 0.8540002697188684, + "eval_sts-test_spearman_dot": 0.7757822250085247, + "eval_sts-test_spearman_euclidean": 0.843399411915428, + "eval_sts-test_spearman_manhattan": 0.8418668154665568, + "eval_sts-test_spearman_max": 0.8540002697188684, + "eval_vitaminc-pairs_loss": 3.759916305541992, + "eval_vitaminc-pairs_runtime": 5.5184, + "eval_vitaminc-pairs_samples_per_second": 23.195, + "eval_vitaminc-pairs_steps_per_second": 0.362, + "step": 920 + }, + { + "epoch": 0.7137315748642359, + "eval_negation-triplets_loss": 1.7734837532043457, + "eval_negation-triplets_runtime": 1.1074, + "eval_negation-triplets_samples_per_second": 115.586, + "eval_negation-triplets_steps_per_second": 1.806, + "step": 920 + }, + { + "epoch": 0.7137315748642359, + "eval_scitail-pairs-pos_loss": 0.1350197196006775, + "eval_scitail-pairs-pos_runtime": 1.3296, + "eval_scitail-pairs-pos_samples_per_second": 96.271, + "eval_scitail-pairs-pos_steps_per_second": 1.504, + "step": 920 + }, + { + "epoch": 0.7137315748642359, + "eval_scitail-pairs-qa_loss": 0.03005460649728775, + "eval_scitail-pairs-qa_runtime": 0.7829, + "eval_scitail-pairs-qa_samples_per_second": 163.486, + "eval_scitail-pairs-qa_steps_per_second": 2.554, + "step": 920 + }, + { + "epoch": 0.7137315748642359, + "eval_xsum-pairs_loss": 0.2775123119354248, + "eval_xsum-pairs_runtime": 3.9263, + "eval_xsum-pairs_samples_per_second": 32.601, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 920 + }, + { + "epoch": 0.7137315748642359, + "eval_sciq_pairs_loss": 0.05456567928195, + "eval_sciq_pairs_runtime": 5.9974, + "eval_sciq_pairs_samples_per_second": 21.342, + "eval_sciq_pairs_steps_per_second": 0.333, + "step": 920 + }, + { + "epoch": 0.7137315748642359, + "eval_qasc_pairs_loss": 0.5647321343421936, + "eval_qasc_pairs_runtime": 0.9092, + "eval_qasc_pairs_samples_per_second": 140.786, + "eval_qasc_pairs_steps_per_second": 2.2, + "step": 920 + }, + { + "epoch": 0.7137315748642359, + "eval_openbookqa_pairs_loss": 1.345017671585083, + "eval_openbookqa_pairs_runtime": 0.8787, + "eval_openbookqa_pairs_samples_per_second": 145.662, + "eval_openbookqa_pairs_steps_per_second": 2.276, + "step": 920 + }, + { + "epoch": 0.7137315748642359, + "eval_msmarco_pairs_loss": 0.5561577677726746, + "eval_msmarco_pairs_runtime": 1.9887, + "eval_msmarco_pairs_samples_per_second": 64.365, + "eval_msmarco_pairs_steps_per_second": 1.006, + "step": 920 + }, + { + "epoch": 0.7137315748642359, + "eval_nq_pairs_loss": 1.0891139507293701, + "eval_nq_pairs_runtime": 3.3527, + "eval_nq_pairs_samples_per_second": 38.178, + "eval_nq_pairs_steps_per_second": 0.597, + "step": 920 + }, + { + "epoch": 0.7137315748642359, + "eval_trivia_pairs_loss": 1.0408532619476318, + "eval_trivia_pairs_runtime": 4.492, + "eval_trivia_pairs_samples_per_second": 28.495, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 920 + }, + { + "epoch": 0.7137315748642359, + "eval_gooaq_pairs_loss": 0.8888065814971924, + "eval_gooaq_pairs_runtime": 1.478, + "eval_gooaq_pairs_samples_per_second": 86.603, + "eval_gooaq_pairs_steps_per_second": 1.353, + "step": 920 + }, + { + "epoch": 0.7137315748642359, + "eval_paws-pos_loss": 0.04470743238925934, + "eval_paws-pos_runtime": 1.0293, + "eval_paws-pos_samples_per_second": 124.362, + "eval_paws-pos_steps_per_second": 1.943, + "step": 920 + }, + { + "epoch": 0.7137315748642359, + "eval_global_dataset_loss": 0.4433729648590088, + "eval_global_dataset_runtime": 19.439, + "eval_global_dataset_samples_per_second": 21.4, + "eval_global_dataset_steps_per_second": 0.36, + "step": 920 + }, + { + "epoch": 0.7145073700543056, + "grad_norm": 10.645776748657227, + "learning_rate": 3.488316376298563e-05, + "loss": 0.732, + "step": 921 + }, + { + "epoch": 0.7152831652443755, + "grad_norm": 11.578245162963867, + "learning_rate": 3.4879654568296846e-05, + "loss": 1.0932, + "step": 922 + }, + { + "epoch": 0.7160589604344453, + "grad_norm": 13.087963104248047, + "learning_rate": 3.4876093652915666e-05, + "loss": 1.0546, + "step": 923 + }, + { + "epoch": 0.7168347556245152, + "grad_norm": 6.977993488311768, + "learning_rate": 3.487248102862565e-05, + "loss": 0.4105, + "step": 924 + }, + { + "epoch": 0.7176105508145849, + "grad_norm": 7.872547626495361, + "learning_rate": 3.486881670738147e-05, + "loss": 0.4764, + "step": 925 + }, + { + "epoch": 0.7183863460046548, + "grad_norm": 2.0418319702148438, + "learning_rate": 3.4865100701308877e-05, + "loss": 0.0551, + "step": 926 + }, + { + "epoch": 0.7191621411947245, + "grad_norm": 10.137189865112305, + "learning_rate": 3.486133302270465e-05, + "loss": 0.6324, + "step": 927 + }, + { + "epoch": 0.7199379363847944, + "grad_norm": 13.254183769226074, + "learning_rate": 3.485751368403656e-05, + "loss": 1.1658, + "step": 928 + }, + { + "epoch": 0.7207137315748642, + "grad_norm": 8.744871139526367, + "learning_rate": 3.485364269794333e-05, + "loss": 0.3725, + "step": 929 + }, + { + "epoch": 0.721489526764934, + "grad_norm": 12.860345840454102, + "learning_rate": 3.484972007723457e-05, + "loss": 1.0303, + "step": 930 + }, + { + "epoch": 0.7222653219550039, + "grad_norm": 0.6916890740394592, + "learning_rate": 3.48457458348908e-05, + "loss": 0.0143, + "step": 931 + }, + { + "epoch": 0.7230411171450737, + "grad_norm": 8.768937110900879, + "learning_rate": 3.484171998406332e-05, + "loss": 0.6967, + "step": 932 + }, + { + "epoch": 0.7238169123351436, + "grad_norm": 12.096138000488281, + "learning_rate": 3.483764253807425e-05, + "loss": 1.0323, + "step": 933 + }, + { + "epoch": 0.7245927075252133, + "grad_norm": 11.769198417663574, + "learning_rate": 3.48335135104164e-05, + "loss": 0.5981, + "step": 934 + }, + { + "epoch": 0.7253685027152832, + "grad_norm": 7.922912120819092, + "learning_rate": 3.4829332914753316e-05, + "loss": 0.4261, + "step": 935 + }, + { + "epoch": 0.726144297905353, + "grad_norm": 17.935937881469727, + "learning_rate": 3.4825100764919154e-05, + "loss": 1.6686, + "step": 936 + }, + { + "epoch": 0.7269200930954228, + "grad_norm": 12.311088562011719, + "learning_rate": 3.4820817074918685e-05, + "loss": 0.7292, + "step": 937 + }, + { + "epoch": 0.7276958882854926, + "grad_norm": 7.077203273773193, + "learning_rate": 3.4816481858927244e-05, + "loss": 0.5332, + "step": 938 + }, + { + "epoch": 0.7284716834755625, + "grad_norm": 2.056474208831787, + "learning_rate": 3.481209513129064e-05, + "loss": 0.0532, + "step": 939 + }, + { + "epoch": 0.7292474786656322, + "grad_norm": 10.370635032653809, + "learning_rate": 3.480765690652518e-05, + "loss": 0.7376, + "step": 940 + }, + { + "epoch": 0.7292474786656322, + "eval_Qnli-dev_cosine_accuracy": 0.6796875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8280144929885864, + "eval_Qnli-dev_cosine_ap": 0.7058509042895263, + "eval_Qnli-dev_cosine_f1": 0.6778625954198474, + "eval_Qnli-dev_cosine_f1_threshold": 0.7085098028182983, + "eval_Qnli-dev_cosine_precision": 0.5298329355608592, + "eval_Qnli-dev_cosine_recall": 0.940677966101695, + "eval_Qnli-dev_dot_accuracy": 0.640625, + "eval_Qnli-dev_dot_accuracy_threshold": 443.646484375, + "eval_Qnli-dev_dot_ap": 0.6044636556615453, + "eval_Qnli-dev_dot_f1": 0.663677130044843, + "eval_Qnli-dev_dot_f1_threshold": 345.868896484375, + "eval_Qnli-dev_dot_precision": 0.5127020785219399, + "eval_Qnli-dev_dot_recall": 0.940677966101695, + "eval_Qnli-dev_euclidean_accuracy": 0.669921875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.699701309204102, + "eval_Qnli-dev_euclidean_ap": 0.7061007571201711, + "eval_Qnli-dev_euclidean_f1": 0.6782334384858044, + "eval_Qnli-dev_euclidean_f1_threshold": 17.033733367919922, + "eval_Qnli-dev_euclidean_precision": 0.5402010050251256, + "eval_Qnli-dev_euclidean_recall": 0.9110169491525424, + "eval_Qnli-dev_manhattan_accuracy": 0.666015625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 226.12686157226562, + "eval_Qnli-dev_manhattan_ap": 0.7025844391350389, + "eval_Qnli-dev_manhattan_f1": 0.6740157480314961, + "eval_Qnli-dev_manhattan_f1_threshold": 303.73370361328125, + "eval_Qnli-dev_manhattan_precision": 0.5363408521303258, + "eval_Qnli-dev_manhattan_recall": 0.9067796610169492, + "eval_Qnli-dev_max_accuracy": 0.6796875, + "eval_Qnli-dev_max_accuracy_threshold": 443.646484375, + "eval_Qnli-dev_max_ap": 0.7061007571201711, + "eval_Qnli-dev_max_f1": 0.6782334384858044, + "eval_Qnli-dev_max_f1_threshold": 345.868896484375, + "eval_Qnli-dev_max_precision": 0.5402010050251256, + "eval_Qnli-dev_max_recall": 0.940677966101695, + "eval_allNLI-dev_cosine_accuracy": 0.712890625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9065077304840088, + "eval_allNLI-dev_cosine_ap": 0.57775990915654, + "eval_allNLI-dev_cosine_f1": 0.5841121495327103, + "eval_allNLI-dev_cosine_f1_threshold": 0.8144516348838806, + "eval_allNLI-dev_cosine_precision": 0.49019607843137253, + "eval_allNLI-dev_cosine_recall": 0.7225433526011561, + "eval_allNLI-dev_dot_accuracy": 0.671875, + "eval_allNLI-dev_dot_accuracy_threshold": 449.16064453125, + "eval_allNLI-dev_dot_ap": 0.45729689491160724, + "eval_allNLI-dev_dot_f1": 0.5467128027681661, + "eval_allNLI-dev_dot_f1_threshold": 313.206787109375, + "eval_allNLI-dev_dot_precision": 0.39012345679012345, + "eval_allNLI-dev_dot_recall": 0.9132947976878613, + "eval_allNLI-dev_euclidean_accuracy": 0.71875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.180174827575684, + "eval_allNLI-dev_euclidean_ap": 0.5803810871730595, + "eval_allNLI-dev_euclidean_f1": 0.5927505330490405, + "eval_allNLI-dev_euclidean_f1_threshold": 14.028618812561035, + "eval_allNLI-dev_euclidean_precision": 0.46959459459459457, + "eval_allNLI-dev_euclidean_recall": 0.8034682080924855, + "eval_allNLI-dev_manhattan_accuracy": 0.71484375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 167.8479766845703, + "eval_allNLI-dev_manhattan_ap": 0.5844999705183276, + "eval_allNLI-dev_manhattan_f1": 0.6053811659192826, + "eval_allNLI-dev_manhattan_f1_threshold": 238.28770446777344, + "eval_allNLI-dev_manhattan_precision": 0.4945054945054945, + "eval_allNLI-dev_manhattan_recall": 0.7803468208092486, + "eval_allNLI-dev_max_accuracy": 0.71875, + "eval_allNLI-dev_max_accuracy_threshold": 449.16064453125, + "eval_allNLI-dev_max_ap": 0.5844999705183276, + "eval_allNLI-dev_max_f1": 0.6053811659192826, + "eval_allNLI-dev_max_f1_threshold": 313.206787109375, + "eval_allNLI-dev_max_precision": 0.4945054945054945, + "eval_allNLI-dev_max_recall": 0.9132947976878613, + "eval_sequential_score": 0.7061007571201711, + "eval_sts-test_pearson_cosine": 0.8258559801510045, + "eval_sts-test_pearson_dot": 0.7768522247641632, + "eval_sts-test_pearson_euclidean": 0.8539592579225929, + "eval_sts-test_pearson_manhattan": 0.8523939672474321, + "eval_sts-test_pearson_max": 0.8539592579225929, + "eval_sts-test_spearman_cosine": 0.8550871883555428, + "eval_sts-test_spearman_dot": 0.7621918934416161, + "eval_sts-test_spearman_euclidean": 0.8479704400077298, + "eval_sts-test_spearman_manhattan": 0.8472401441730836, + "eval_sts-test_spearman_max": 0.8550871883555428, + "eval_vitaminc-pairs_loss": 3.7480885982513428, + "eval_vitaminc-pairs_runtime": 5.4843, + "eval_vitaminc-pairs_samples_per_second": 23.339, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 940 + }, + { + "epoch": 0.7292474786656322, + "eval_negation-triplets_loss": 1.8641772270202637, + "eval_negation-triplets_runtime": 1.0931, + "eval_negation-triplets_samples_per_second": 117.096, + "eval_negation-triplets_steps_per_second": 1.83, + "step": 940 + }, + { + "epoch": 0.7292474786656322, + "eval_scitail-pairs-pos_loss": 0.17913472652435303, + "eval_scitail-pairs-pos_runtime": 1.3209, + "eval_scitail-pairs-pos_samples_per_second": 96.905, + "eval_scitail-pairs-pos_steps_per_second": 1.514, + "step": 940 + }, + { + "epoch": 0.7292474786656322, + "eval_scitail-pairs-qa_loss": 0.03495565056800842, + "eval_scitail-pairs-qa_runtime": 0.7864, + "eval_scitail-pairs-qa_samples_per_second": 162.763, + "eval_scitail-pairs-qa_steps_per_second": 2.543, + "step": 940 + }, + { + "epoch": 0.7292474786656322, + "eval_xsum-pairs_loss": 0.28637024760246277, + "eval_xsum-pairs_runtime": 3.9477, + "eval_xsum-pairs_samples_per_second": 32.424, + "eval_xsum-pairs_steps_per_second": 0.507, + "step": 940 + }, + { + "epoch": 0.7292474786656322, + "eval_sciq_pairs_loss": 0.07482024282217026, + "eval_sciq_pairs_runtime": 5.997, + "eval_sciq_pairs_samples_per_second": 21.344, + "eval_sciq_pairs_steps_per_second": 0.333, + "step": 940 + }, + { + "epoch": 0.7292474786656322, + "eval_qasc_pairs_loss": 0.542219877243042, + "eval_qasc_pairs_runtime": 0.9198, + "eval_qasc_pairs_samples_per_second": 139.162, + "eval_qasc_pairs_steps_per_second": 2.174, + "step": 940 + }, + { + "epoch": 0.7292474786656322, + "eval_openbookqa_pairs_loss": 1.573500156402588, + "eval_openbookqa_pairs_runtime": 0.8794, + "eval_openbookqa_pairs_samples_per_second": 145.557, + "eval_openbookqa_pairs_steps_per_second": 2.274, + "step": 940 + }, + { + "epoch": 0.7292474786656322, + "eval_msmarco_pairs_loss": 0.5651401281356812, + "eval_msmarco_pairs_runtime": 1.9828, + "eval_msmarco_pairs_samples_per_second": 64.556, + "eval_msmarco_pairs_steps_per_second": 1.009, + "step": 940 + }, + { + "epoch": 0.7292474786656322, + "eval_nq_pairs_loss": 0.9907878637313843, + "eval_nq_pairs_runtime": 3.3483, + "eval_nq_pairs_samples_per_second": 38.228, + "eval_nq_pairs_steps_per_second": 0.597, + "step": 940 + }, + { + "epoch": 0.7292474786656322, + "eval_trivia_pairs_loss": 1.112267255783081, + "eval_trivia_pairs_runtime": 4.4813, + "eval_trivia_pairs_samples_per_second": 28.563, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 940 + }, + { + "epoch": 0.7292474786656322, + "eval_gooaq_pairs_loss": 0.874494194984436, + "eval_gooaq_pairs_runtime": 1.4752, + "eval_gooaq_pairs_samples_per_second": 86.77, + "eval_gooaq_pairs_steps_per_second": 1.356, + "step": 940 + }, + { + "epoch": 0.7292474786656322, + "eval_paws-pos_loss": 0.04339267686009407, + "eval_paws-pos_runtime": 1.041, + "eval_paws-pos_samples_per_second": 122.963, + "eval_paws-pos_steps_per_second": 1.921, + "step": 940 + }, + { + "epoch": 0.7292474786656322, + "eval_global_dataset_loss": 0.45819559693336487, + "eval_global_dataset_runtime": 19.4273, + "eval_global_dataset_samples_per_second": 21.413, + "eval_global_dataset_steps_per_second": 0.36, + "step": 940 + }, + { + "epoch": 0.7300232738557021, + "grad_norm": 15.412274360656738, + "learning_rate": 3.480316719931757e-05, + "loss": 1.225, + "step": 941 + }, + { + "epoch": 0.730799069045772, + "grad_norm": 12.429174423217773, + "learning_rate": 3.479862602452486e-05, + "loss": 1.3277, + "step": 942 + }, + { + "epoch": 0.7315748642358417, + "grad_norm": 13.888239860534668, + "learning_rate": 3.479403339717442e-05, + "loss": 1.3168, + "step": 943 + }, + { + "epoch": 0.7323506594259116, + "grad_norm": 7.709323406219482, + "learning_rate": 3.47893893324639e-05, + "loss": 0.4183, + "step": 944 + }, + { + "epoch": 0.7331264546159814, + "grad_norm": 8.998419761657715, + "learning_rate": 3.4784693845761165e-05, + "loss": 0.3764, + "step": 945 + }, + { + "epoch": 0.7339022498060512, + "grad_norm": 9.600420951843262, + "learning_rate": 3.4779946952604215e-05, + "loss": 0.5638, + "step": 946 + }, + { + "epoch": 0.734678044996121, + "grad_norm": 6.363062381744385, + "learning_rate": 3.477514866870118e-05, + "loss": 0.3658, + "step": 947 + }, + { + "epoch": 0.7354538401861909, + "grad_norm": 7.470354080200195, + "learning_rate": 3.477029900993025e-05, + "loss": 0.3867, + "step": 948 + }, + { + "epoch": 0.7362296353762606, + "grad_norm": 8.215919494628906, + "learning_rate": 3.4765397992339616e-05, + "loss": 0.5428, + "step": 949 + }, + { + "epoch": 0.7370054305663305, + "grad_norm": 1.1038960218429565, + "learning_rate": 3.476044563214742e-05, + "loss": 0.0486, + "step": 950 + }, + { + "epoch": 0.7377812257564003, + "grad_norm": 17.344226837158203, + "learning_rate": 3.475544194574171e-05, + "loss": 2.6881, + "step": 951 + }, + { + "epoch": 0.7385570209464701, + "grad_norm": 6.872907638549805, + "learning_rate": 3.4750386949680384e-05, + "loss": 0.314, + "step": 952 + }, + { + "epoch": 0.73933281613654, + "grad_norm": 6.5277605056762695, + "learning_rate": 3.474528066069112e-05, + "loss": 0.2931, + "step": 953 + }, + { + "epoch": 0.7401086113266098, + "grad_norm": 7.285345077514648, + "learning_rate": 3.474012309567134e-05, + "loss": 0.3684, + "step": 954 + }, + { + "epoch": 0.7408844065166796, + "grad_norm": 9.644484519958496, + "learning_rate": 3.4734914271688125e-05, + "loss": 0.6858, + "step": 955 + }, + { + "epoch": 0.7416602017067494, + "grad_norm": 8.192893981933594, + "learning_rate": 3.4729654205978214e-05, + "loss": 0.425, + "step": 956 + }, + { + "epoch": 0.7424359968968193, + "grad_norm": 6.8188910484313965, + "learning_rate": 3.472434291594789e-05, + "loss": 0.4366, + "step": 957 + }, + { + "epoch": 0.743211792086889, + "grad_norm": 7.57077693939209, + "learning_rate": 3.471898041917294e-05, + "loss": 0.3495, + "step": 958 + }, + { + "epoch": 0.7439875872769589, + "grad_norm": 1.092852234840393, + "learning_rate": 3.471356673339861e-05, + "loss": 0.0577, + "step": 959 + }, + { + "epoch": 0.7447633824670287, + "grad_norm": 6.401102066040039, + "learning_rate": 3.470810187653955e-05, + "loss": 0.4148, + "step": 960 + }, + { + "epoch": 0.7447633824670287, + "eval_Qnli-dev_cosine_accuracy": 0.685546875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8656002283096313, + "eval_Qnli-dev_cosine_ap": 0.7255235372298487, + "eval_Qnli-dev_cosine_f1": 0.674496644295302, + "eval_Qnli-dev_cosine_f1_threshold": 0.7862218022346497, + "eval_Qnli-dev_cosine_precision": 0.5583333333333333, + "eval_Qnli-dev_cosine_recall": 0.8516949152542372, + "eval_Qnli-dev_dot_accuracy": 0.642578125, + "eval_Qnli-dev_dot_accuracy_threshold": 477.6988525390625, + "eval_Qnli-dev_dot_ap": 0.6173507160125875, + "eval_Qnli-dev_dot_f1": 0.6688102893890675, + "eval_Qnli-dev_dot_f1_threshold": 387.66192626953125, + "eval_Qnli-dev_dot_precision": 0.538860103626943, + "eval_Qnli-dev_dot_recall": 0.8813559322033898, + "eval_Qnli-dev_euclidean_accuracy": 0.6796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.54013442993164, + "eval_Qnli-dev_euclidean_ap": 0.7226780626222544, + "eval_Qnli-dev_euclidean_f1": 0.6774774774774774, + "eval_Qnli-dev_euclidean_f1_threshold": 14.36434555053711, + "eval_Qnli-dev_euclidean_precision": 0.5893416927899686, + "eval_Qnli-dev_euclidean_recall": 0.7966101694915254, + "eval_Qnli-dev_manhattan_accuracy": 0.673828125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 212.01638793945312, + "eval_Qnli-dev_manhattan_ap": 0.7174460312641439, + "eval_Qnli-dev_manhattan_f1": 0.6766355140186916, + "eval_Qnli-dev_manhattan_f1_threshold": 251.17071533203125, + "eval_Qnli-dev_manhattan_precision": 0.6053511705685619, + "eval_Qnli-dev_manhattan_recall": 0.7669491525423728, + "eval_Qnli-dev_max_accuracy": 0.685546875, + "eval_Qnli-dev_max_accuracy_threshold": 477.6988525390625, + "eval_Qnli-dev_max_ap": 0.7255235372298487, + "eval_Qnli-dev_max_f1": 0.6774774774774774, + "eval_Qnli-dev_max_f1_threshold": 387.66192626953125, + "eval_Qnli-dev_max_precision": 0.6053511705685619, + "eval_Qnli-dev_max_recall": 0.8813559322033898, + "eval_allNLI-dev_cosine_accuracy": 0.71484375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9092354774475098, + "eval_allNLI-dev_cosine_ap": 0.5890921194908372, + "eval_allNLI-dev_cosine_f1": 0.6150341685649203, + "eval_allNLI-dev_cosine_f1_threshold": 0.8473508358001709, + "eval_allNLI-dev_cosine_precision": 0.5075187969924813, + "eval_allNLI-dev_cosine_recall": 0.7803468208092486, + "eval_allNLI-dev_dot_accuracy": 0.6796875, + "eval_allNLI-dev_dot_accuracy_threshold": 500.93719482421875, + "eval_allNLI-dev_dot_ap": 0.47197649007439835, + "eval_allNLI-dev_dot_f1": 0.5467625899280575, + "eval_allNLI-dev_dot_f1_threshold": 384.30609130859375, + "eval_allNLI-dev_dot_precision": 0.3968668407310705, + "eval_allNLI-dev_dot_recall": 0.8786127167630058, + "eval_allNLI-dev_euclidean_accuracy": 0.71484375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 9.310173034667969, + "eval_allNLI-dev_euclidean_ap": 0.5891813531307, + "eval_allNLI-dev_euclidean_f1": 0.6146993318485523, + "eval_allNLI-dev_euclidean_f1_threshold": 12.670351028442383, + "eval_allNLI-dev_euclidean_precision": 0.5, + "eval_allNLI-dev_euclidean_recall": 0.7976878612716763, + "eval_allNLI-dev_manhattan_accuracy": 0.71484375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 178.8214569091797, + "eval_allNLI-dev_manhattan_ap": 0.5960774566459738, + "eval_allNLI-dev_manhattan_f1": 0.6272727272727273, + "eval_allNLI-dev_manhattan_f1_threshold": 223.1829833984375, + "eval_allNLI-dev_manhattan_precision": 0.5168539325842697, + "eval_allNLI-dev_manhattan_recall": 0.7976878612716763, + "eval_allNLI-dev_max_accuracy": 0.71484375, + "eval_allNLI-dev_max_accuracy_threshold": 500.93719482421875, + "eval_allNLI-dev_max_ap": 0.5960774566459738, + "eval_allNLI-dev_max_f1": 0.6272727272727273, + "eval_allNLI-dev_max_f1_threshold": 384.30609130859375, + "eval_allNLI-dev_max_precision": 0.5168539325842697, + "eval_allNLI-dev_max_recall": 0.8786127167630058, + "eval_sequential_score": 0.7255235372298487, + "eval_sts-test_pearson_cosine": 0.818045462157344, + "eval_sts-test_pearson_dot": 0.7472446957354837, + "eval_sts-test_pearson_euclidean": 0.8461722565508376, + "eval_sts-test_pearson_manhattan": 0.8433768937096828, + "eval_sts-test_pearson_max": 0.8461722565508376, + "eval_sts-test_spearman_cosine": 0.8506462074288443, + "eval_sts-test_spearman_dot": 0.7355198283935123, + "eval_sts-test_spearman_euclidean": 0.8425965295348049, + "eval_sts-test_spearman_manhattan": 0.841473630883972, + "eval_sts-test_spearman_max": 0.8506462074288443, + "eval_vitaminc-pairs_loss": 4.06262731552124, + "eval_vitaminc-pairs_runtime": 5.4867, + "eval_vitaminc-pairs_samples_per_second": 23.329, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 960 + }, + { + "epoch": 0.7447633824670287, + "eval_negation-triplets_loss": 1.7792047262191772, + "eval_negation-triplets_runtime": 1.0664, + "eval_negation-triplets_samples_per_second": 120.033, + "eval_negation-triplets_steps_per_second": 1.876, + "step": 960 + }, + { + "epoch": 0.7447633824670287, + "eval_scitail-pairs-pos_loss": 0.17724700272083282, + "eval_scitail-pairs-pos_runtime": 1.2551, + "eval_scitail-pairs-pos_samples_per_second": 101.986, + "eval_scitail-pairs-pos_steps_per_second": 1.594, + "step": 960 + }, + { + "epoch": 0.7447633824670287, + "eval_scitail-pairs-qa_loss": 0.03478042036294937, + "eval_scitail-pairs-qa_runtime": 0.7709, + "eval_scitail-pairs-qa_samples_per_second": 166.034, + "eval_scitail-pairs-qa_steps_per_second": 2.594, + "step": 960 + }, + { + "epoch": 0.7447633824670287, + "eval_xsum-pairs_loss": 0.29500454664230347, + "eval_xsum-pairs_runtime": 3.9374, + "eval_xsum-pairs_samples_per_second": 32.508, + "eval_xsum-pairs_steps_per_second": 0.508, + "step": 960 + }, + { + "epoch": 0.7447633824670287, + "eval_sciq_pairs_loss": 0.061557844281196594, + "eval_sciq_pairs_runtime": 6.0558, + "eval_sciq_pairs_samples_per_second": 21.137, + "eval_sciq_pairs_steps_per_second": 0.33, + "step": 960 + }, + { + "epoch": 0.7447633824670287, + "eval_qasc_pairs_loss": 0.5219850540161133, + "eval_qasc_pairs_runtime": 0.9254, + "eval_qasc_pairs_samples_per_second": 138.314, + "eval_qasc_pairs_steps_per_second": 2.161, + "step": 960 + }, + { + "epoch": 0.7447633824670287, + "eval_openbookqa_pairs_loss": 1.3571869134902954, + "eval_openbookqa_pairs_runtime": 0.8718, + "eval_openbookqa_pairs_samples_per_second": 146.816, + "eval_openbookqa_pairs_steps_per_second": 2.294, + "step": 960 + }, + { + "epoch": 0.7447633824670287, + "eval_msmarco_pairs_loss": 0.5442098379135132, + "eval_msmarco_pairs_runtime": 1.9906, + "eval_msmarco_pairs_samples_per_second": 64.301, + "eval_msmarco_pairs_steps_per_second": 1.005, + "step": 960 + }, + { + "epoch": 0.7447633824670287, + "eval_nq_pairs_loss": 1.0508705377578735, + "eval_nq_pairs_runtime": 3.3401, + "eval_nq_pairs_samples_per_second": 38.322, + "eval_nq_pairs_steps_per_second": 0.599, + "step": 960 + }, + { + "epoch": 0.7447633824670287, + "eval_trivia_pairs_loss": 1.0006734132766724, + "eval_trivia_pairs_runtime": 4.4933, + "eval_trivia_pairs_samples_per_second": 28.487, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 960 + }, + { + "epoch": 0.7447633824670287, + "eval_gooaq_pairs_loss": 0.7926682829856873, + "eval_gooaq_pairs_runtime": 1.4732, + "eval_gooaq_pairs_samples_per_second": 86.888, + "eval_gooaq_pairs_steps_per_second": 1.358, + "step": 960 + }, + { + "epoch": 0.7447633824670287, + "eval_paws-pos_loss": 0.046479757875204086, + "eval_paws-pos_runtime": 1.045, + "eval_paws-pos_samples_per_second": 122.487, + "eval_paws-pos_steps_per_second": 1.914, + "step": 960 + }, + { + "epoch": 0.7447633824670287, + "eval_global_dataset_loss": 0.4857000410556793, + "eval_global_dataset_runtime": 19.4377, + "eval_global_dataset_samples_per_second": 21.402, + "eval_global_dataset_steps_per_second": 0.36, + "step": 960 + }, + { + "epoch": 0.7455391776570985, + "grad_norm": 11.5498628616333, + "learning_rate": 3.47025858666797e-05, + "loss": 1.0169, + "step": 961 + }, + { + "epoch": 0.7463149728471683, + "grad_norm": 13.27639389038086, + "learning_rate": 3.469701872207233e-05, + "loss": 1.6831, + "step": 962 + }, + { + "epoch": 0.7470907680372382, + "grad_norm": 12.058984756469727, + "learning_rate": 3.469140046113987e-05, + "loss": 1.1021, + "step": 963 + }, + { + "epoch": 0.747866563227308, + "grad_norm": 10.550564765930176, + "learning_rate": 3.468573110247395e-05, + "loss": 1.0278, + "step": 964 + }, + { + "epoch": 0.7486423584173778, + "grad_norm": 11.664640426635742, + "learning_rate": 3.468001066483524e-05, + "loss": 0.9023, + "step": 965 + }, + { + "epoch": 0.7494181536074477, + "grad_norm": 7.251794338226318, + "learning_rate": 3.467423916715347e-05, + "loss": 0.3945, + "step": 966 + }, + { + "epoch": 0.7501939487975174, + "grad_norm": 14.326492309570312, + "learning_rate": 3.466841662852732e-05, + "loss": 2.3776, + "step": 967 + }, + { + "epoch": 0.7509697439875873, + "grad_norm": 7.800229549407959, + "learning_rate": 3.4662543068224384e-05, + "loss": 0.2596, + "step": 968 + }, + { + "epoch": 0.7517455391776571, + "grad_norm": 9.339960098266602, + "learning_rate": 3.465661850568109e-05, + "loss": 0.8079, + "step": 969 + }, + { + "epoch": 0.752521334367727, + "grad_norm": 11.21225643157959, + "learning_rate": 3.4650642960502624e-05, + "loss": 1.156, + "step": 970 + }, + { + "epoch": 0.7532971295577967, + "grad_norm": 8.088404655456543, + "learning_rate": 3.4644616452462904e-05, + "loss": 0.6678, + "step": 971 + }, + { + "epoch": 0.7540729247478666, + "grad_norm": 10.209814071655273, + "learning_rate": 3.463853900150447e-05, + "loss": 0.7599, + "step": 972 + }, + { + "epoch": 0.7548487199379363, + "grad_norm": 9.69403076171875, + "learning_rate": 3.463241062773845e-05, + "loss": 0.6048, + "step": 973 + }, + { + "epoch": 0.7556245151280062, + "grad_norm": 15.4074125289917, + "learning_rate": 3.4626231351444495e-05, + "loss": 1.2722, + "step": 974 + }, + { + "epoch": 0.756400310318076, + "grad_norm": 14.900464057922363, + "learning_rate": 3.462000119307068e-05, + "loss": 1.3419, + "step": 975 + }, + { + "epoch": 0.7571761055081458, + "grad_norm": 17.61496353149414, + "learning_rate": 3.4613720173233456e-05, + "loss": 2.6248, + "step": 976 + }, + { + "epoch": 0.7579519006982157, + "grad_norm": 14.07415771484375, + "learning_rate": 3.460738831271759e-05, + "loss": 1.2239, + "step": 977 + }, + { + "epoch": 0.7587276958882855, + "grad_norm": 4.216908931732178, + "learning_rate": 3.460100563247609e-05, + "loss": 0.2392, + "step": 978 + }, + { + "epoch": 0.7595034910783554, + "grad_norm": 11.05616569519043, + "learning_rate": 3.4594572153630126e-05, + "loss": 1.0528, + "step": 979 + }, + { + "epoch": 0.7602792862684251, + "grad_norm": 11.93682861328125, + "learning_rate": 3.4588087897468975e-05, + "loss": 1.3056, + "step": 980 + }, + { + "epoch": 0.7602792862684251, + "eval_Qnli-dev_cosine_accuracy": 0.66015625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8404420614242554, + "eval_Qnli-dev_cosine_ap": 0.6941312390718141, + "eval_Qnli-dev_cosine_f1": 0.6666666666666666, + "eval_Qnli-dev_cosine_f1_threshold": 0.7989790439605713, + "eval_Qnli-dev_cosine_precision": 0.5568181818181818, + "eval_Qnli-dev_cosine_recall": 0.8305084745762712, + "eval_Qnli-dev_dot_accuracy": 0.61328125, + "eval_Qnli-dev_dot_accuracy_threshold": 444.3583984375, + "eval_Qnli-dev_dot_ap": 0.5618087057330504, + "eval_Qnli-dev_dot_f1": 0.6618075801749271, + "eval_Qnli-dev_dot_f1_threshold": 359.6221923828125, + "eval_Qnli-dev_dot_precision": 0.5044444444444445, + "eval_Qnli-dev_dot_recall": 0.961864406779661, + "eval_Qnli-dev_euclidean_accuracy": 0.6640625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.234148025512695, + "eval_Qnli-dev_euclidean_ap": 0.6923964385935938, + "eval_Qnli-dev_euclidean_f1": 0.6655791190864601, + "eval_Qnli-dev_euclidean_f1_threshold": 15.007804870605469, + "eval_Qnli-dev_euclidean_precision": 0.5411140583554377, + "eval_Qnli-dev_euclidean_recall": 0.864406779661017, + "eval_Qnli-dev_manhattan_accuracy": 0.654296875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 211.71136474609375, + "eval_Qnli-dev_manhattan_ap": 0.6901108974358061, + "eval_Qnli-dev_manhattan_f1": 0.6688741721854304, + "eval_Qnli-dev_manhattan_f1_threshold": 263.3021240234375, + "eval_Qnli-dev_manhattan_precision": 0.5489130434782609, + "eval_Qnli-dev_manhattan_recall": 0.8559322033898306, + "eval_Qnli-dev_max_accuracy": 0.6640625, + "eval_Qnli-dev_max_accuracy_threshold": 444.3583984375, + "eval_Qnli-dev_max_ap": 0.6941312390718141, + "eval_Qnli-dev_max_f1": 0.6688741721854304, + "eval_Qnli-dev_max_f1_threshold": 359.6221923828125, + "eval_Qnli-dev_max_precision": 0.5568181818181818, + "eval_Qnli-dev_max_recall": 0.961864406779661, + "eval_allNLI-dev_cosine_accuracy": 0.712890625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9104657173156738, + "eval_allNLI-dev_cosine_ap": 0.5779393990460262, + "eval_allNLI-dev_cosine_f1": 0.5919661733615224, + "eval_allNLI-dev_cosine_f1_threshold": 0.841463565826416, + "eval_allNLI-dev_cosine_precision": 0.4666666666666667, + "eval_allNLI-dev_cosine_recall": 0.8092485549132948, + "eval_allNLI-dev_dot_accuracy": 0.669921875, + "eval_allNLI-dev_dot_accuracy_threshold": 516.3771362304688, + "eval_allNLI-dev_dot_ap": 0.45016307963080104, + "eval_allNLI-dev_dot_f1": 0.5376712328767123, + "eval_allNLI-dev_dot_f1_threshold": 374.87420654296875, + "eval_allNLI-dev_dot_precision": 0.3819951338199513, + "eval_allNLI-dev_dot_recall": 0.9075144508670521, + "eval_allNLI-dev_euclidean_accuracy": 0.708984375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 9.567673683166504, + "eval_allNLI-dev_euclidean_ap": 0.5792851643715327, + "eval_allNLI-dev_euclidean_f1": 0.6013071895424836, + "eval_allNLI-dev_euclidean_f1_threshold": 12.404951095581055, + "eval_allNLI-dev_euclidean_precision": 0.4825174825174825, + "eval_allNLI-dev_euclidean_recall": 0.7976878612716763, + "eval_allNLI-dev_manhattan_accuracy": 0.71484375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 176.902099609375, + "eval_allNLI-dev_manhattan_ap": 0.5874153676959651, + "eval_allNLI-dev_manhattan_f1": 0.6117647058823529, + "eval_allNLI-dev_manhattan_f1_threshold": 209.28500366210938, + "eval_allNLI-dev_manhattan_precision": 0.5158730158730159, + "eval_allNLI-dev_manhattan_recall": 0.7514450867052023, + "eval_allNLI-dev_max_accuracy": 0.71484375, + "eval_allNLI-dev_max_accuracy_threshold": 516.3771362304688, + "eval_allNLI-dev_max_ap": 0.5874153676959651, + "eval_allNLI-dev_max_f1": 0.6117647058823529, + "eval_allNLI-dev_max_f1_threshold": 374.87420654296875, + "eval_allNLI-dev_max_precision": 0.5158730158730159, + "eval_allNLI-dev_max_recall": 0.9075144508670521, + "eval_sequential_score": 0.6941312390718141, + "eval_sts-test_pearson_cosine": 0.8177811310172752, + "eval_sts-test_pearson_dot": 0.7335788831179144, + "eval_sts-test_pearson_euclidean": 0.850014522826015, + "eval_sts-test_pearson_manhattan": 0.848001330484057, + "eval_sts-test_pearson_max": 0.850014522826015, + "eval_sts-test_spearman_cosine": 0.8537648333326882, + "eval_sts-test_spearman_dot": 0.7230402678855996, + "eval_sts-test_spearman_euclidean": 0.8465784520221504, + "eval_sts-test_spearman_manhattan": 0.8451639094008296, + "eval_sts-test_spearman_max": 0.8537648333326882, + "eval_vitaminc-pairs_loss": 3.625579357147217, + "eval_vitaminc-pairs_runtime": 5.5175, + "eval_vitaminc-pairs_samples_per_second": 23.199, + "eval_vitaminc-pairs_steps_per_second": 0.362, + "step": 980 + }, + { + "epoch": 0.7602792862684251, + "eval_negation-triplets_loss": 1.7572097778320312, + "eval_negation-triplets_runtime": 1.1154, + "eval_negation-triplets_samples_per_second": 114.752, + "eval_negation-triplets_steps_per_second": 1.793, + "step": 980 + }, + { + "epoch": 0.7602792862684251, + "eval_scitail-pairs-pos_loss": 0.16017776727676392, + "eval_scitail-pairs-pos_runtime": 1.3365, + "eval_scitail-pairs-pos_samples_per_second": 95.774, + "eval_scitail-pairs-pos_steps_per_second": 1.496, + "step": 980 + }, + { + "epoch": 0.7602792862684251, + "eval_scitail-pairs-qa_loss": 0.04132052883505821, + "eval_scitail-pairs-qa_runtime": 0.7902, + "eval_scitail-pairs-qa_samples_per_second": 161.98, + "eval_scitail-pairs-qa_steps_per_second": 2.531, + "step": 980 + }, + { + "epoch": 0.7602792862684251, + "eval_xsum-pairs_loss": 0.23874011635780334, + "eval_xsum-pairs_runtime": 3.9414, + "eval_xsum-pairs_samples_per_second": 32.476, + "eval_xsum-pairs_steps_per_second": 0.507, + "step": 980 + }, + { + "epoch": 0.7602792862684251, + "eval_sciq_pairs_loss": 0.055444370955228806, + "eval_sciq_pairs_runtime": 5.9941, + "eval_sciq_pairs_samples_per_second": 21.354, + "eval_sciq_pairs_steps_per_second": 0.334, + "step": 980 + }, + { + "epoch": 0.7602792862684251, + "eval_qasc_pairs_loss": 0.5690393447875977, + "eval_qasc_pairs_runtime": 0.9289, + "eval_qasc_pairs_samples_per_second": 137.793, + "eval_qasc_pairs_steps_per_second": 2.153, + "step": 980 + }, + { + "epoch": 0.7602792862684251, + "eval_openbookqa_pairs_loss": 1.3320801258087158, + "eval_openbookqa_pairs_runtime": 0.8665, + "eval_openbookqa_pairs_samples_per_second": 147.721, + "eval_openbookqa_pairs_steps_per_second": 2.308, + "step": 980 + }, + { + "epoch": 0.7602792862684251, + "eval_msmarco_pairs_loss": 0.5957273244857788, + "eval_msmarco_pairs_runtime": 1.978, + "eval_msmarco_pairs_samples_per_second": 64.713, + "eval_msmarco_pairs_steps_per_second": 1.011, + "step": 980 + }, + { + "epoch": 0.7602792862684251, + "eval_nq_pairs_loss": 1.0263999700546265, + "eval_nq_pairs_runtime": 3.3353, + "eval_nq_pairs_samples_per_second": 38.377, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 980 + }, + { + "epoch": 0.7602792862684251, + "eval_trivia_pairs_loss": 0.9815043807029724, + "eval_trivia_pairs_runtime": 4.4807, + "eval_trivia_pairs_samples_per_second": 28.567, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 980 + }, + { + "epoch": 0.7602792862684251, + "eval_gooaq_pairs_loss": 0.8182553648948669, + "eval_gooaq_pairs_runtime": 1.4737, + "eval_gooaq_pairs_samples_per_second": 86.857, + "eval_gooaq_pairs_steps_per_second": 1.357, + "step": 980 + }, + { + "epoch": 0.7602792862684251, + "eval_paws-pos_loss": 0.03923580050468445, + "eval_paws-pos_runtime": 1.0333, + "eval_paws-pos_samples_per_second": 123.876, + "eval_paws-pos_steps_per_second": 1.936, + "step": 980 + }, + { + "epoch": 0.7602792862684251, + "eval_global_dataset_loss": 0.4378804564476013, + "eval_global_dataset_runtime": 19.4523, + "eval_global_dataset_samples_per_second": 21.386, + "eval_global_dataset_steps_per_second": 0.36, + "step": 980 + }, + { + "epoch": 0.761055081458495, + "grad_norm": 9.621814727783203, + "learning_rate": 3.4581552885449926e-05, + "loss": 0.8806, + "step": 981 + }, + { + "epoch": 0.7618308766485647, + "grad_norm": 8.521477699279785, + "learning_rate": 3.457496713919825e-05, + "loss": 0.6343, + "step": 982 + }, + { + "epoch": 0.7626066718386346, + "grad_norm": 7.55567741394043, + "learning_rate": 3.4568330680507104e-05, + "loss": 0.9935, + "step": 983 + }, + { + "epoch": 0.7633824670287044, + "grad_norm": 7.183990478515625, + "learning_rate": 3.456164353133743e-05, + "loss": 0.5072, + "step": 984 + }, + { + "epoch": 0.7641582622187743, + "grad_norm": 12.03846263885498, + "learning_rate": 3.455490571381794e-05, + "loss": 1.1407, + "step": 985 + }, + { + "epoch": 0.764934057408844, + "grad_norm": 9.04784870147705, + "learning_rate": 3.4548117250245e-05, + "loss": 0.9667, + "step": 986 + }, + { + "epoch": 0.7657098525989139, + "grad_norm": 14.017303466796875, + "learning_rate": 3.454127816308259e-05, + "loss": 1.3694, + "step": 987 + }, + { + "epoch": 0.7664856477889838, + "grad_norm": 7.459906578063965, + "learning_rate": 3.453438847496218e-05, + "loss": 0.4272, + "step": 988 + }, + { + "epoch": 0.7672614429790535, + "grad_norm": 5.984421730041504, + "learning_rate": 3.452744820868271e-05, + "loss": 0.3278, + "step": 989 + }, + { + "epoch": 0.7680372381691234, + "grad_norm": 5.240382194519043, + "learning_rate": 3.452045738721049e-05, + "loss": 0.3453, + "step": 990 + }, + { + "epoch": 0.7688130333591932, + "grad_norm": 8.625643730163574, + "learning_rate": 3.4513416033679105e-05, + "loss": 0.7828, + "step": 991 + }, + { + "epoch": 0.769588828549263, + "grad_norm": 5.903829574584961, + "learning_rate": 3.4506324171389376e-05, + "loss": 0.3041, + "step": 992 + }, + { + "epoch": 0.7703646237393328, + "grad_norm": 6.317503929138184, + "learning_rate": 3.449918182380925e-05, + "loss": 0.3975, + "step": 993 + }, + { + "epoch": 0.7711404189294027, + "grad_norm": 11.507726669311523, + "learning_rate": 3.449198901457376e-05, + "loss": 0.9688, + "step": 994 + }, + { + "epoch": 0.7719162141194724, + "grad_norm": 11.324596405029297, + "learning_rate": 3.44847457674849e-05, + "loss": 1.1401, + "step": 995 + }, + { + "epoch": 0.7726920093095423, + "grad_norm": 6.348566055297852, + "learning_rate": 3.447745210651157e-05, + "loss": 0.2577, + "step": 996 + }, + { + "epoch": 0.773467804499612, + "grad_norm": 7.979206562042236, + "learning_rate": 3.447010805578952e-05, + "loss": 0.4082, + "step": 997 + }, + { + "epoch": 0.7742435996896819, + "grad_norm": 11.176900863647461, + "learning_rate": 3.4462713639621216e-05, + "loss": 1.1123, + "step": 998 + }, + { + "epoch": 0.7750193948797518, + "grad_norm": 11.583893775939941, + "learning_rate": 3.445526888247582e-05, + "loss": 1.1666, + "step": 999 + }, + { + "epoch": 0.7757951900698216, + "grad_norm": 8.623787879943848, + "learning_rate": 3.444777380898906e-05, + "loss": 0.3604, + "step": 1000 + }, + { + "epoch": 0.7757951900698216, + "eval_Qnli-dev_cosine_accuracy": 0.6640625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8585967421531677, + "eval_Qnli-dev_cosine_ap": 0.7072261339706888, + "eval_Qnli-dev_cosine_f1": 0.6784, + "eval_Qnli-dev_cosine_f1_threshold": 0.7310217022895813, + "eval_Qnli-dev_cosine_precision": 0.5449871465295629, + "eval_Qnli-dev_cosine_recall": 0.8983050847457628, + "eval_Qnli-dev_dot_accuracy": 0.66015625, + "eval_Qnli-dev_dot_accuracy_threshold": 438.5562438964844, + "eval_Qnli-dev_dot_ap": 0.6181116448278678, + "eval_Qnli-dev_dot_f1": 0.6862068965517241, + "eval_Qnli-dev_dot_f1_threshold": 387.94915771484375, + "eval_Qnli-dev_dot_precision": 0.5784883720930233, + "eval_Qnli-dev_dot_recall": 0.8432203389830508, + "eval_Qnli-dev_euclidean_accuracy": 0.662109375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.772767066955566, + "eval_Qnli-dev_euclidean_ap": 0.7042470542356045, + "eval_Qnli-dev_euclidean_f1": 0.67430441898527, + "eval_Qnli-dev_euclidean_f1_threshold": 16.53150177001953, + "eval_Qnli-dev_euclidean_precision": 0.5493333333333333, + "eval_Qnli-dev_euclidean_recall": 0.8728813559322034, + "eval_Qnli-dev_manhattan_accuracy": 0.662109375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 233.6343536376953, + "eval_Qnli-dev_manhattan_ap": 0.7049491900641356, + "eval_Qnli-dev_manhattan_f1": 0.6740157480314961, + "eval_Qnli-dev_manhattan_f1_threshold": 310.9897155761719, + "eval_Qnli-dev_manhattan_precision": 0.5363408521303258, + "eval_Qnli-dev_manhattan_recall": 0.9067796610169492, + "eval_Qnli-dev_max_accuracy": 0.6640625, + "eval_Qnli-dev_max_accuracy_threshold": 438.5562438964844, + "eval_Qnli-dev_max_ap": 0.7072261339706888, + "eval_Qnli-dev_max_f1": 0.6862068965517241, + "eval_Qnli-dev_max_f1_threshold": 387.94915771484375, + "eval_Qnli-dev_max_precision": 0.5784883720930233, + "eval_Qnli-dev_max_recall": 0.9067796610169492, + "eval_allNLI-dev_cosine_accuracy": 0.71875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8951315879821777, + "eval_allNLI-dev_cosine_ap": 0.5922299489907299, + "eval_allNLI-dev_cosine_f1": 0.6082949308755761, + "eval_allNLI-dev_cosine_f1_threshold": 0.8141263723373413, + "eval_allNLI-dev_cosine_precision": 0.5057471264367817, + "eval_allNLI-dev_cosine_recall": 0.7630057803468208, + "eval_allNLI-dev_dot_accuracy": 0.681640625, + "eval_allNLI-dev_dot_accuracy_threshold": 446.1053466796875, + "eval_allNLI-dev_dot_ap": 0.48246622052927113, + "eval_allNLI-dev_dot_f1": 0.5590062111801242, + "eval_allNLI-dev_dot_f1_threshold": 374.9225769042969, + "eval_allNLI-dev_dot_precision": 0.43548387096774194, + "eval_allNLI-dev_dot_recall": 0.7803468208092486, + "eval_allNLI-dev_euclidean_accuracy": 0.720703125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.087106704711914, + "eval_allNLI-dev_euclidean_ap": 0.5941553994592871, + "eval_allNLI-dev_euclidean_f1": 0.6103896103896104, + "eval_allNLI-dev_euclidean_f1_threshold": 14.128737449645996, + "eval_allNLI-dev_euclidean_precision": 0.48788927335640137, + "eval_allNLI-dev_euclidean_recall": 0.815028901734104, + "eval_allNLI-dev_manhattan_accuracy": 0.72265625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 187.36474609375, + "eval_allNLI-dev_manhattan_ap": 0.5980100739317054, + "eval_allNLI-dev_manhattan_f1": 0.6112359550561798, + "eval_allNLI-dev_manhattan_f1_threshold": 250.49891662597656, + "eval_allNLI-dev_manhattan_precision": 0.5, + "eval_allNLI-dev_manhattan_recall": 0.7861271676300579, + "eval_allNLI-dev_max_accuracy": 0.72265625, + "eval_allNLI-dev_max_accuracy_threshold": 446.1053466796875, + "eval_allNLI-dev_max_ap": 0.5980100739317054, + "eval_allNLI-dev_max_f1": 0.6112359550561798, + "eval_allNLI-dev_max_f1_threshold": 374.9225769042969, + "eval_allNLI-dev_max_precision": 0.5057471264367817, + "eval_allNLI-dev_max_recall": 0.815028901734104, + "eval_sequential_score": 0.7072261339706888, + "eval_sts-test_pearson_cosine": 0.8240460299579315, + "eval_sts-test_pearson_dot": 0.7760484561339727, + "eval_sts-test_pearson_euclidean": 0.8490394139563605, + "eval_sts-test_pearson_manhattan": 0.8452740837154643, + "eval_sts-test_pearson_max": 0.8490394139563605, + "eval_sts-test_spearman_cosine": 0.8523662221604228, + "eval_sts-test_spearman_dot": 0.7650894481141866, + "eval_sts-test_spearman_euclidean": 0.8438821885233185, + "eval_sts-test_spearman_manhattan": 0.8415195456781596, + "eval_sts-test_spearman_max": 0.8523662221604228, + "eval_vitaminc-pairs_loss": 3.8202943801879883, + "eval_vitaminc-pairs_runtime": 5.5018, + "eval_vitaminc-pairs_samples_per_second": 23.265, + "eval_vitaminc-pairs_steps_per_second": 0.364, + "step": 1000 + }, + { + "epoch": 0.7757951900698216, + "eval_negation-triplets_loss": 1.8671759366989136, + "eval_negation-triplets_runtime": 1.0857, + "eval_negation-triplets_samples_per_second": 117.901, + "eval_negation-triplets_steps_per_second": 1.842, + "step": 1000 + }, + { + "epoch": 0.7757951900698216, + "eval_scitail-pairs-pos_loss": 0.11585887521505356, + "eval_scitail-pairs-pos_runtime": 1.3118, + "eval_scitail-pairs-pos_samples_per_second": 97.574, + "eval_scitail-pairs-pos_steps_per_second": 1.525, + "step": 1000 + }, + { + "epoch": 0.7757951900698216, + "eval_scitail-pairs-qa_loss": 0.007170978933572769, + "eval_scitail-pairs-qa_runtime": 0.7957, + "eval_scitail-pairs-qa_samples_per_second": 160.866, + "eval_scitail-pairs-qa_steps_per_second": 2.514, + "step": 1000 + }, + { + "epoch": 0.7757951900698216, + "eval_xsum-pairs_loss": 0.21687336266040802, + "eval_xsum-pairs_runtime": 3.9429, + "eval_xsum-pairs_samples_per_second": 32.464, + "eval_xsum-pairs_steps_per_second": 0.507, + "step": 1000 + }, + { + "epoch": 0.7757951900698216, + "eval_sciq_pairs_loss": 0.05961114913225174, + "eval_sciq_pairs_runtime": 6.0342, + "eval_sciq_pairs_samples_per_second": 21.212, + "eval_sciq_pairs_steps_per_second": 0.331, + "step": 1000 + }, + { + "epoch": 0.7757951900698216, + "eval_qasc_pairs_loss": 0.5340542197227478, + "eval_qasc_pairs_runtime": 0.9281, + "eval_qasc_pairs_samples_per_second": 137.921, + "eval_qasc_pairs_steps_per_second": 2.155, + "step": 1000 + }, + { + "epoch": 0.7757951900698216, + "eval_openbookqa_pairs_loss": 1.3652724027633667, + "eval_openbookqa_pairs_runtime": 0.8824, + "eval_openbookqa_pairs_samples_per_second": 145.055, + "eval_openbookqa_pairs_steps_per_second": 2.266, + "step": 1000 + }, + { + "epoch": 0.7757951900698216, + "eval_msmarco_pairs_loss": 0.5528172254562378, + "eval_msmarco_pairs_runtime": 1.9796, + "eval_msmarco_pairs_samples_per_second": 64.659, + "eval_msmarco_pairs_steps_per_second": 1.01, + "step": 1000 + }, + { + "epoch": 0.7757951900698216, + "eval_nq_pairs_loss": 0.9207841753959656, + "eval_nq_pairs_runtime": 3.3449, + "eval_nq_pairs_samples_per_second": 38.267, + "eval_nq_pairs_steps_per_second": 0.598, + "step": 1000 + }, + { + "epoch": 0.7757951900698216, + "eval_trivia_pairs_loss": 1.1129192113876343, + "eval_trivia_pairs_runtime": 4.4712, + "eval_trivia_pairs_samples_per_second": 28.627, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 1000 + }, + { + "epoch": 0.7757951900698216, + "eval_gooaq_pairs_loss": 0.7599539160728455, + "eval_gooaq_pairs_runtime": 1.4703, + "eval_gooaq_pairs_samples_per_second": 87.056, + "eval_gooaq_pairs_steps_per_second": 1.36, + "step": 1000 + }, + { + "epoch": 0.7757951900698216, + "eval_paws-pos_loss": 0.03802761435508728, + "eval_paws-pos_runtime": 1.0313, + "eval_paws-pos_samples_per_second": 124.114, + "eval_paws-pos_steps_per_second": 1.939, + "step": 1000 + }, + { + "epoch": 0.7757951900698216, + "eval_global_dataset_loss": 0.46472543478012085, + "eval_global_dataset_runtime": 19.4347, + "eval_global_dataset_samples_per_second": 21.405, + "eval_global_dataset_steps_per_second": 0.36, + "step": 1000 + }, + { + "epoch": 0.7765709852598914, + "grad_norm": 9.618617057800293, + "learning_rate": 3.444022844396316e-05, + "loss": 0.5992, + "step": 1001 + }, + { + "epoch": 0.7773467804499612, + "grad_norm": 11.019386291503906, + "learning_rate": 3.443263281236679e-05, + "loss": 0.9816, + "step": 1002 + }, + { + "epoch": 0.7781225756400311, + "grad_norm": 6.277504920959473, + "learning_rate": 3.442498693933495e-05, + "loss": 0.2457, + "step": 1003 + }, + { + "epoch": 0.7788983708301008, + "grad_norm": 9.241889953613281, + "learning_rate": 3.4417290850168886e-05, + "loss": 0.4026, + "step": 1004 + }, + { + "epoch": 0.7796741660201707, + "grad_norm": 11.455988883972168, + "learning_rate": 3.4409544570336025e-05, + "loss": 1.0561, + "step": 1005 + }, + { + "epoch": 0.7804499612102405, + "grad_norm": 6.769739151000977, + "learning_rate": 3.440174812546989e-05, + "loss": 0.3819, + "step": 1006 + }, + { + "epoch": 0.7812257564003103, + "grad_norm": 9.21048641204834, + "learning_rate": 3.439390154136999e-05, + "loss": 0.7407, + "step": 1007 + }, + { + "epoch": 0.7820015515903801, + "grad_norm": 0.6326342225074768, + "learning_rate": 3.4386004844001745e-05, + "loss": 0.0148, + "step": 1008 + }, + { + "epoch": 0.78277734678045, + "grad_norm": 16.204675674438477, + "learning_rate": 3.437805805949644e-05, + "loss": 1.5554, + "step": 1009 + }, + { + "epoch": 0.7835531419705197, + "grad_norm": 13.959929466247559, + "learning_rate": 3.437006121415108e-05, + "loss": 1.3148, + "step": 1010 + }, + { + "epoch": 0.7843289371605896, + "grad_norm": 13.183578491210938, + "learning_rate": 3.436201433442833e-05, + "loss": 1.1629, + "step": 1011 + }, + { + "epoch": 0.7851047323506595, + "grad_norm": 11.557097434997559, + "learning_rate": 3.435391744695643e-05, + "loss": 0.9787, + "step": 1012 + }, + { + "epoch": 0.7858805275407292, + "grad_norm": 6.3334197998046875, + "learning_rate": 3.434577057852912e-05, + "loss": 0.2663, + "step": 1013 + }, + { + "epoch": 0.7866563227307991, + "grad_norm": 7.671533584594727, + "learning_rate": 3.433757375610549e-05, + "loss": 0.4504, + "step": 1014 + }, + { + "epoch": 0.7874321179208689, + "grad_norm": 2.7367801666259766, + "learning_rate": 3.432932700680998e-05, + "loss": 0.088, + "step": 1015 + }, + { + "epoch": 0.7882079131109387, + "grad_norm": 7.571774959564209, + "learning_rate": 3.432103035793221e-05, + "loss": 0.4763, + "step": 1016 + }, + { + "epoch": 0.7889837083010085, + "grad_norm": 11.234426498413086, + "learning_rate": 3.4312683836926964e-05, + "loss": 1.1033, + "step": 1017 + }, + { + "epoch": 0.7897595034910784, + "grad_norm": 9.548408508300781, + "learning_rate": 3.4304287471414005e-05, + "loss": 0.8237, + "step": 1018 + }, + { + "epoch": 0.7905352986811481, + "grad_norm": 7.409108638763428, + "learning_rate": 3.42958412891781e-05, + "loss": 0.6254, + "step": 1019 + }, + { + "epoch": 0.791311093871218, + "grad_norm": 6.861879825592041, + "learning_rate": 3.428734531816882e-05, + "loss": 0.3469, + "step": 1020 + }, + { + "epoch": 0.791311093871218, + "eval_Qnli-dev_cosine_accuracy": 0.666015625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8681036233901978, + "eval_Qnli-dev_cosine_ap": 0.7094084030823454, + "eval_Qnli-dev_cosine_f1": 0.6764705882352942, + "eval_Qnli-dev_cosine_f1_threshold": 0.765945315361023, + "eval_Qnli-dev_cosine_precision": 0.550531914893617, + "eval_Qnli-dev_cosine_recall": 0.8771186440677966, + "eval_Qnli-dev_dot_accuracy": 0.65234375, + "eval_Qnli-dev_dot_accuracy_threshold": 457.37042236328125, + "eval_Qnli-dev_dot_ap": 0.6358214479317308, + "eval_Qnli-dev_dot_f1": 0.6646433990895295, + "eval_Qnli-dev_dot_f1_threshold": 364.1715087890625, + "eval_Qnli-dev_dot_precision": 0.5177304964539007, + "eval_Qnli-dev_dot_recall": 0.9279661016949152, + "eval_Qnli-dev_euclidean_accuracy": 0.666015625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.555442810058594, + "eval_Qnli-dev_euclidean_ap": 0.703264483424873, + "eval_Qnli-dev_euclidean_f1": 0.6811352253756261, + "eval_Qnli-dev_euclidean_f1_threshold": 15.256208419799805, + "eval_Qnli-dev_euclidean_precision": 0.5619834710743802, + "eval_Qnli-dev_euclidean_recall": 0.864406779661017, + "eval_Qnli-dev_manhattan_accuracy": 0.6640625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 209.95579528808594, + "eval_Qnli-dev_manhattan_ap": 0.7010388632047662, + "eval_Qnli-dev_manhattan_f1": 0.678688524590164, + "eval_Qnli-dev_manhattan_f1_threshold": 279.05859375, + "eval_Qnli-dev_manhattan_precision": 0.553475935828877, + "eval_Qnli-dev_manhattan_recall": 0.8771186440677966, + "eval_Qnli-dev_max_accuracy": 0.666015625, + "eval_Qnli-dev_max_accuracy_threshold": 457.37042236328125, + "eval_Qnli-dev_max_ap": 0.7094084030823454, + "eval_Qnli-dev_max_f1": 0.6811352253756261, + "eval_Qnli-dev_max_f1_threshold": 364.1715087890625, + "eval_Qnli-dev_max_precision": 0.5619834710743802, + "eval_Qnli-dev_max_recall": 0.9279661016949152, + "eval_allNLI-dev_cosine_accuracy": 0.724609375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9020903706550598, + "eval_allNLI-dev_cosine_ap": 0.5908777095252007, + "eval_allNLI-dev_cosine_f1": 0.6072186836518046, + "eval_allNLI-dev_cosine_f1_threshold": 0.8101892471313477, + "eval_allNLI-dev_cosine_precision": 0.4798657718120805, + "eval_allNLI-dev_cosine_recall": 0.8265895953757225, + "eval_allNLI-dev_dot_accuracy": 0.673828125, + "eval_allNLI-dev_dot_accuracy_threshold": 445.72357177734375, + "eval_allNLI-dev_dot_ap": 0.4811334688217761, + "eval_allNLI-dev_dot_f1": 0.5601577909270217, + "eval_allNLI-dev_dot_f1_threshold": 373.24920654296875, + "eval_allNLI-dev_dot_precision": 0.4251497005988024, + "eval_allNLI-dev_dot_recall": 0.8208092485549133, + "eval_allNLI-dev_euclidean_accuracy": 0.724609375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 9.453763961791992, + "eval_allNLI-dev_euclidean_ap": 0.5979380797387099, + "eval_allNLI-dev_euclidean_f1": 0.6030368763557483, + "eval_allNLI-dev_euclidean_f1_threshold": 13.23131275177002, + "eval_allNLI-dev_euclidean_precision": 0.4826388888888889, + "eval_allNLI-dev_euclidean_recall": 0.8034682080924855, + "eval_allNLI-dev_manhattan_accuracy": 0.73046875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 175.73252868652344, + "eval_allNLI-dev_manhattan_ap": 0.6015885881415546, + "eval_allNLI-dev_manhattan_f1": 0.6041666666666666, + "eval_allNLI-dev_manhattan_f1_threshold": 246.5721435546875, + "eval_allNLI-dev_manhattan_precision": 0.4723127035830619, + "eval_allNLI-dev_manhattan_recall": 0.838150289017341, + "eval_allNLI-dev_max_accuracy": 0.73046875, + "eval_allNLI-dev_max_accuracy_threshold": 445.72357177734375, + "eval_allNLI-dev_max_ap": 0.6015885881415546, + "eval_allNLI-dev_max_f1": 0.6072186836518046, + "eval_allNLI-dev_max_f1_threshold": 373.24920654296875, + "eval_allNLI-dev_max_precision": 0.4826388888888889, + "eval_allNLI-dev_max_recall": 0.838150289017341, + "eval_sequential_score": 0.7094084030823454, + "eval_sts-test_pearson_cosine": 0.8171211771924991, + "eval_sts-test_pearson_dot": 0.7826872705119017, + "eval_sts-test_pearson_euclidean": 0.8449734475782972, + "eval_sts-test_pearson_manhattan": 0.8394242105464415, + "eval_sts-test_pearson_max": 0.8449734475782972, + "eval_sts-test_spearman_cosine": 0.8478566375500115, + "eval_sts-test_spearman_dot": 0.7765593754138477, + "eval_sts-test_spearman_euclidean": 0.8410154227185854, + "eval_sts-test_spearman_manhattan": 0.8379763198062341, + "eval_sts-test_spearman_max": 0.8478566375500115, + "eval_vitaminc-pairs_loss": 3.7102108001708984, + "eval_vitaminc-pairs_runtime": 5.4983, + "eval_vitaminc-pairs_samples_per_second": 23.28, + "eval_vitaminc-pairs_steps_per_second": 0.364, + "step": 1020 + }, + { + "epoch": 0.791311093871218, + "eval_negation-triplets_loss": 1.8503952026367188, + "eval_negation-triplets_runtime": 1.0838, + "eval_negation-triplets_samples_per_second": 118.098, + "eval_negation-triplets_steps_per_second": 1.845, + "step": 1020 + }, + { + "epoch": 0.791311093871218, + "eval_scitail-pairs-pos_loss": 0.11608420312404633, + "eval_scitail-pairs-pos_runtime": 1.3048, + "eval_scitail-pairs-pos_samples_per_second": 98.098, + "eval_scitail-pairs-pos_steps_per_second": 1.533, + "step": 1020 + }, + { + "epoch": 0.791311093871218, + "eval_scitail-pairs-qa_loss": 0.015517156571149826, + "eval_scitail-pairs-qa_runtime": 0.7864, + "eval_scitail-pairs-qa_samples_per_second": 162.767, + "eval_scitail-pairs-qa_steps_per_second": 2.543, + "step": 1020 + }, + { + "epoch": 0.791311093871218, + "eval_xsum-pairs_loss": 0.2630694508552551, + "eval_xsum-pairs_runtime": 3.9337, + "eval_xsum-pairs_samples_per_second": 32.539, + "eval_xsum-pairs_steps_per_second": 0.508, + "step": 1020 + }, + { + "epoch": 0.791311093871218, + "eval_sciq_pairs_loss": 0.05809640884399414, + "eval_sciq_pairs_runtime": 6.0331, + "eval_sciq_pairs_samples_per_second": 21.216, + "eval_sciq_pairs_steps_per_second": 0.332, + "step": 1020 + }, + { + "epoch": 0.791311093871218, + "eval_qasc_pairs_loss": 0.4373838007450104, + "eval_qasc_pairs_runtime": 0.9312, + "eval_qasc_pairs_samples_per_second": 137.457, + "eval_qasc_pairs_steps_per_second": 2.148, + "step": 1020 + }, + { + "epoch": 0.791311093871218, + "eval_openbookqa_pairs_loss": 1.2920573949813843, + "eval_openbookqa_pairs_runtime": 0.878, + "eval_openbookqa_pairs_samples_per_second": 145.79, + "eval_openbookqa_pairs_steps_per_second": 2.278, + "step": 1020 + }, + { + "epoch": 0.791311093871218, + "eval_msmarco_pairs_loss": 0.4513913691043854, + "eval_msmarco_pairs_runtime": 1.9774, + "eval_msmarco_pairs_samples_per_second": 64.732, + "eval_msmarco_pairs_steps_per_second": 1.011, + "step": 1020 + }, + { + "epoch": 0.791311093871218, + "eval_nq_pairs_loss": 0.9309043884277344, + "eval_nq_pairs_runtime": 3.3634, + "eval_nq_pairs_samples_per_second": 38.057, + "eval_nq_pairs_steps_per_second": 0.595, + "step": 1020 + }, + { + "epoch": 0.791311093871218, + "eval_trivia_pairs_loss": 1.0333608388900757, + "eval_trivia_pairs_runtime": 4.5041, + "eval_trivia_pairs_samples_per_second": 28.418, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 1020 + }, + { + "epoch": 0.791311093871218, + "eval_gooaq_pairs_loss": 0.6852787733078003, + "eval_gooaq_pairs_runtime": 1.4614, + "eval_gooaq_pairs_samples_per_second": 87.585, + "eval_gooaq_pairs_steps_per_second": 1.369, + "step": 1020 + }, + { + "epoch": 0.791311093871218, + "eval_paws-pos_loss": 0.0355130210518837, + "eval_paws-pos_runtime": 1.0428, + "eval_paws-pos_samples_per_second": 122.744, + "eval_paws-pos_steps_per_second": 1.918, + "step": 1020 + }, + { + "epoch": 0.791311093871218, + "eval_global_dataset_loss": 0.44458168745040894, + "eval_global_dataset_runtime": 19.4069, + "eval_global_dataset_samples_per_second": 21.436, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1020 + }, + { + "epoch": 0.7920868890612878, + "grad_norm": 7.659545421600342, + "learning_rate": 3.427879958650051e-05, + "loss": 0.5202, + "step": 1021 + }, + { + "epoch": 0.7928626842513576, + "grad_norm": 10.905531883239746, + "learning_rate": 3.4270204122452186e-05, + "loss": 1.2528, + "step": 1022 + }, + { + "epoch": 0.7936384794414275, + "grad_norm": 6.264749050140381, + "learning_rate": 3.4261558954467424e-05, + "loss": 0.3869, + "step": 1023 + }, + { + "epoch": 0.7944142746314973, + "grad_norm": 0.9429741501808167, + "learning_rate": 3.425286411115429e-05, + "loss": 0.0338, + "step": 1024 + }, + { + "epoch": 0.7951900698215671, + "grad_norm": 9.22243595123291, + "learning_rate": 3.4244119621285214e-05, + "loss": 0.8394, + "step": 1025 + }, + { + "epoch": 0.7959658650116369, + "grad_norm": 6.922791481018066, + "learning_rate": 3.4235325513796934e-05, + "loss": 0.2863, + "step": 1026 + }, + { + "epoch": 0.7967416602017068, + "grad_norm": 8.849251747131348, + "learning_rate": 3.422648181779037e-05, + "loss": 0.4839, + "step": 1027 + }, + { + "epoch": 0.7975174553917765, + "grad_norm": 10.697627067565918, + "learning_rate": 3.4217588562530535e-05, + "loss": 0.7893, + "step": 1028 + }, + { + "epoch": 0.7982932505818464, + "grad_norm": 14.224536895751953, + "learning_rate": 3.420864577744644e-05, + "loss": 1.2165, + "step": 1029 + }, + { + "epoch": 0.7990690457719162, + "grad_norm": 12.774467468261719, + "learning_rate": 3.4199653492131e-05, + "loss": 1.0787, + "step": 1030 + }, + { + "epoch": 0.799844840961986, + "grad_norm": 7.543542861938477, + "learning_rate": 3.4190611736340934e-05, + "loss": 0.3367, + "step": 1031 + }, + { + "epoch": 0.8006206361520558, + "grad_norm": 7.519282817840576, + "learning_rate": 3.418152053999667e-05, + "loss": 0.5754, + "step": 1032 + }, + { + "epoch": 0.8013964313421257, + "grad_norm": 11.25484848022461, + "learning_rate": 3.417237993318223e-05, + "loss": 1.3409, + "step": 1033 + }, + { + "epoch": 0.8021722265321956, + "grad_norm": 5.334971904754639, + "learning_rate": 3.4163189946145155e-05, + "loss": 0.2152, + "step": 1034 + }, + { + "epoch": 0.8029480217222653, + "grad_norm": 8.074272155761719, + "learning_rate": 3.415395060929637e-05, + "loss": 0.9073, + "step": 1035 + }, + { + "epoch": 0.8037238169123352, + "grad_norm": 0.739331066608429, + "learning_rate": 3.414466195321014e-05, + "loss": 0.0427, + "step": 1036 + }, + { + "epoch": 0.804499612102405, + "grad_norm": 8.767532348632812, + "learning_rate": 3.4135324008623904e-05, + "loss": 0.6175, + "step": 1037 + }, + { + "epoch": 0.8052754072924748, + "grad_norm": 10.862985610961914, + "learning_rate": 3.412593680643821e-05, + "loss": 1.0296, + "step": 1038 + }, + { + "epoch": 0.8060512024825446, + "grad_norm": 7.909887790679932, + "learning_rate": 3.411650037771663e-05, + "loss": 0.5433, + "step": 1039 + }, + { + "epoch": 0.8068269976726145, + "grad_norm": 0.7263232469558716, + "learning_rate": 3.410701475368559e-05, + "loss": 0.0112, + "step": 1040 + }, + { + "epoch": 0.8068269976726145, + "eval_Qnli-dev_cosine_accuracy": 0.666015625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8210340738296509, + "eval_Qnli-dev_cosine_ap": 0.7125189943681585, + "eval_Qnli-dev_cosine_f1": 0.6808510638297872, + "eval_Qnli-dev_cosine_f1_threshold": 0.748835563659668, + "eval_Qnli-dev_cosine_precision": 0.5546666666666666, + "eval_Qnli-dev_cosine_recall": 0.8813559322033898, + "eval_Qnli-dev_dot_accuracy": 0.654296875, + "eval_Qnli-dev_dot_accuracy_threshold": 438.3790283203125, + "eval_Qnli-dev_dot_ap": 0.6211449599128007, + "eval_Qnli-dev_dot_f1": 0.6687116564417178, + "eval_Qnli-dev_dot_f1_threshold": 344.33233642578125, + "eval_Qnli-dev_dot_precision": 0.5240384615384616, + "eval_Qnli-dev_dot_recall": 0.923728813559322, + "eval_Qnli-dev_euclidean_accuracy": 0.671875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.90127944946289, + "eval_Qnli-dev_euclidean_ap": 0.7093914143784855, + "eval_Qnli-dev_euclidean_f1": 0.685064935064935, + "eval_Qnli-dev_euclidean_f1_threshold": 15.914340019226074, + "eval_Qnli-dev_euclidean_precision": 0.5552631578947368, + "eval_Qnli-dev_euclidean_recall": 0.8940677966101694, + "eval_Qnli-dev_manhattan_accuracy": 0.666015625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 226.31259155273438, + "eval_Qnli-dev_manhattan_ap": 0.7089358837181056, + "eval_Qnli-dev_manhattan_f1": 0.6799336650082919, + "eval_Qnli-dev_manhattan_f1_threshold": 280.6347961425781, + "eval_Qnli-dev_manhattan_precision": 0.55858310626703, + "eval_Qnli-dev_manhattan_recall": 0.8686440677966102, + "eval_Qnli-dev_max_accuracy": 0.671875, + "eval_Qnli-dev_max_accuracy_threshold": 438.3790283203125, + "eval_Qnli-dev_max_ap": 0.7125189943681585, + "eval_Qnli-dev_max_f1": 0.685064935064935, + "eval_Qnli-dev_max_f1_threshold": 344.33233642578125, + "eval_Qnli-dev_max_precision": 0.55858310626703, + "eval_Qnli-dev_max_recall": 0.923728813559322, + "eval_allNLI-dev_cosine_accuracy": 0.728515625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8840154409408569, + "eval_allNLI-dev_cosine_ap": 0.60558583122062, + "eval_allNLI-dev_cosine_f1": 0.6170212765957447, + "eval_allNLI-dev_cosine_f1_threshold": 0.7948219180107117, + "eval_allNLI-dev_cosine_precision": 0.4882154882154882, + "eval_allNLI-dev_cosine_recall": 0.838150289017341, + "eval_allNLI-dev_dot_accuracy": 0.693359375, + "eval_allNLI-dev_dot_accuracy_threshold": 427.6092529296875, + "eval_allNLI-dev_dot_ap": 0.5121882361867933, + "eval_allNLI-dev_dot_f1": 0.5795454545454545, + "eval_allNLI-dev_dot_f1_threshold": 357.69793701171875, + "eval_allNLI-dev_dot_precision": 0.4309859154929577, + "eval_allNLI-dev_dot_recall": 0.884393063583815, + "eval_allNLI-dev_euclidean_accuracy": 0.7265625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 9.761795043945312, + "eval_allNLI-dev_euclidean_ap": 0.6067515220246574, + "eval_allNLI-dev_euclidean_f1": 0.6150234741784038, + "eval_allNLI-dev_euclidean_f1_threshold": 13.159971237182617, + "eval_allNLI-dev_euclidean_precision": 0.5177865612648221, + "eval_allNLI-dev_euclidean_recall": 0.7572254335260116, + "eval_allNLI-dev_manhattan_accuracy": 0.73046875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 182.9873809814453, + "eval_allNLI-dev_manhattan_ap": 0.6110602554360581, + "eval_allNLI-dev_manhattan_f1": 0.6271929824561404, + "eval_allNLI-dev_manhattan_f1_threshold": 250.41961669921875, + "eval_allNLI-dev_manhattan_precision": 0.5053003533568905, + "eval_allNLI-dev_manhattan_recall": 0.8265895953757225, + "eval_allNLI-dev_max_accuracy": 0.73046875, + "eval_allNLI-dev_max_accuracy_threshold": 427.6092529296875, + "eval_allNLI-dev_max_ap": 0.6110602554360581, + "eval_allNLI-dev_max_f1": 0.6271929824561404, + "eval_allNLI-dev_max_f1_threshold": 357.69793701171875, + "eval_allNLI-dev_max_precision": 0.5177865612648221, + "eval_allNLI-dev_max_recall": 0.884393063583815, + "eval_sequential_score": 0.7125189943681585, + "eval_sts-test_pearson_cosine": 0.828586001539912, + "eval_sts-test_pearson_dot": 0.8102505522217422, + "eval_sts-test_pearson_euclidean": 0.8572352361257441, + "eval_sts-test_pearson_manhattan": 0.8528714501696302, + "eval_sts-test_pearson_max": 0.8572352361257441, + "eval_sts-test_spearman_cosine": 0.8609469185221326, + "eval_sts-test_spearman_dot": 0.8089330196707842, + "eval_sts-test_spearman_euclidean": 0.8537058511701744, + "eval_sts-test_spearman_manhattan": 0.8516830046527604, + "eval_sts-test_spearman_max": 0.8609469185221326, + "eval_vitaminc-pairs_loss": 3.7627058029174805, + "eval_vitaminc-pairs_runtime": 5.4682, + "eval_vitaminc-pairs_samples_per_second": 23.408, + "eval_vitaminc-pairs_steps_per_second": 0.366, + "step": 1040 + }, + { + "epoch": 0.8068269976726145, + "eval_negation-triplets_loss": 1.8618046045303345, + "eval_negation-triplets_runtime": 1.1302, + "eval_negation-triplets_samples_per_second": 113.25, + "eval_negation-triplets_steps_per_second": 1.77, + "step": 1040 + }, + { + "epoch": 0.8068269976726145, + "eval_scitail-pairs-pos_loss": 0.1311870664358139, + "eval_scitail-pairs-pos_runtime": 1.3315, + "eval_scitail-pairs-pos_samples_per_second": 96.134, + "eval_scitail-pairs-pos_steps_per_second": 1.502, + "step": 1040 + }, + { + "epoch": 0.8068269976726145, + "eval_scitail-pairs-qa_loss": 0.02023470588028431, + "eval_scitail-pairs-qa_runtime": 0.82, + "eval_scitail-pairs-qa_samples_per_second": 156.092, + "eval_scitail-pairs-qa_steps_per_second": 2.439, + "step": 1040 + }, + { + "epoch": 0.8068269976726145, + "eval_xsum-pairs_loss": 0.2975301444530487, + "eval_xsum-pairs_runtime": 3.9314, + "eval_xsum-pairs_samples_per_second": 32.558, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 1040 + }, + { + "epoch": 0.8068269976726145, + "eval_sciq_pairs_loss": 0.05644915625452995, + "eval_sciq_pairs_runtime": 6.0552, + "eval_sciq_pairs_samples_per_second": 21.139, + "eval_sciq_pairs_steps_per_second": 0.33, + "step": 1040 + }, + { + "epoch": 0.8068269976726145, + "eval_qasc_pairs_loss": 0.39358311891555786, + "eval_qasc_pairs_runtime": 0.9707, + "eval_qasc_pairs_samples_per_second": 131.86, + "eval_qasc_pairs_steps_per_second": 2.06, + "step": 1040 + }, + { + "epoch": 0.8068269976726145, + "eval_openbookqa_pairs_loss": 1.296079397201538, + "eval_openbookqa_pairs_runtime": 0.894, + "eval_openbookqa_pairs_samples_per_second": 143.169, + "eval_openbookqa_pairs_steps_per_second": 2.237, + "step": 1040 + }, + { + "epoch": 0.8068269976726145, + "eval_msmarco_pairs_loss": 0.461254358291626, + "eval_msmarco_pairs_runtime": 1.9982, + "eval_msmarco_pairs_samples_per_second": 64.057, + "eval_msmarco_pairs_steps_per_second": 1.001, + "step": 1040 + }, + { + "epoch": 0.8068269976726145, + "eval_nq_pairs_loss": 1.0291471481323242, + "eval_nq_pairs_runtime": 3.3771, + "eval_nq_pairs_samples_per_second": 37.902, + "eval_nq_pairs_steps_per_second": 0.592, + "step": 1040 + }, + { + "epoch": 0.8068269976726145, + "eval_trivia_pairs_loss": 1.0602896213531494, + "eval_trivia_pairs_runtime": 4.5433, + "eval_trivia_pairs_samples_per_second": 28.174, + "eval_trivia_pairs_steps_per_second": 0.44, + "step": 1040 + }, + { + "epoch": 0.8068269976726145, + "eval_gooaq_pairs_loss": 0.6396035552024841, + "eval_gooaq_pairs_runtime": 1.4538, + "eval_gooaq_pairs_samples_per_second": 88.044, + "eval_gooaq_pairs_steps_per_second": 1.376, + "step": 1040 + }, + { + "epoch": 0.8068269976726145, + "eval_paws-pos_loss": 0.03374774754047394, + "eval_paws-pos_runtime": 1.0196, + "eval_paws-pos_samples_per_second": 125.538, + "eval_paws-pos_steps_per_second": 1.962, + "step": 1040 + }, + { + "epoch": 0.8068269976726145, + "eval_global_dataset_loss": 0.46478256583213806, + "eval_global_dataset_runtime": 19.4103, + "eval_global_dataset_samples_per_second": 21.432, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1040 + }, + { + "epoch": 0.8076027928626842, + "grad_norm": 0.8158115744590759, + "learning_rate": 3.409747996573435e-05, + "loss": 0.0366, + "step": 1041 + }, + { + "epoch": 0.8083785880527541, + "grad_norm": 6.480457305908203, + "learning_rate": 3.408789604541484e-05, + "loss": 0.3327, + "step": 1042 + }, + { + "epoch": 0.8091543832428238, + "grad_norm": 5.6955342292785645, + "learning_rate": 3.407826302444158e-05, + "loss": 0.2304, + "step": 1043 + }, + { + "epoch": 0.8099301784328937, + "grad_norm": 14.485492706298828, + "learning_rate": 3.406858093469157e-05, + "loss": 2.3118, + "step": 1044 + }, + { + "epoch": 0.8107059736229635, + "grad_norm": 10.123367309570312, + "learning_rate": 3.4058849808204185e-05, + "loss": 0.6781, + "step": 1045 + }, + { + "epoch": 0.8114817688130334, + "grad_norm": 0.4824964106082916, + "learning_rate": 3.404906967718108e-05, + "loss": 0.0058, + "step": 1046 + }, + { + "epoch": 0.8122575640031032, + "grad_norm": 1.9418292045593262, + "learning_rate": 3.403924057398605e-05, + "loss": 0.0584, + "step": 1047 + }, + { + "epoch": 0.813033359193173, + "grad_norm": 10.041891098022461, + "learning_rate": 3.4029362531144965e-05, + "loss": 0.8155, + "step": 1048 + }, + { + "epoch": 0.8138091543832429, + "grad_norm": 11.220827102661133, + "learning_rate": 3.401943558134563e-05, + "loss": 0.6515, + "step": 1049 + }, + { + "epoch": 0.8145849495733126, + "grad_norm": 13.521173477172852, + "learning_rate": 3.400945975743772e-05, + "loss": 0.9968, + "step": 1050 + }, + { + "epoch": 0.8153607447633825, + "grad_norm": 7.716485500335693, + "learning_rate": 3.39994350924326e-05, + "loss": 0.3533, + "step": 1051 + }, + { + "epoch": 0.8161365399534523, + "grad_norm": 9.911101341247559, + "learning_rate": 3.398936161950328e-05, + "loss": 0.4776, + "step": 1052 + }, + { + "epoch": 0.8169123351435221, + "grad_norm": 6.120659351348877, + "learning_rate": 3.397923937198427e-05, + "loss": 0.2374, + "step": 1053 + }, + { + "epoch": 0.8176881303335919, + "grad_norm": 11.548603057861328, + "learning_rate": 3.396906838337151e-05, + "loss": 0.804, + "step": 1054 + }, + { + "epoch": 0.8184639255236618, + "grad_norm": 9.334627151489258, + "learning_rate": 3.395884868732218e-05, + "loss": 0.4868, + "step": 1055 + }, + { + "epoch": 0.8192397207137315, + "grad_norm": 13.305825233459473, + "learning_rate": 3.3948580317654697e-05, + "loss": 0.984, + "step": 1056 + }, + { + "epoch": 0.8200155159038014, + "grad_norm": 8.298413276672363, + "learning_rate": 3.3938263308348506e-05, + "loss": 0.3604, + "step": 1057 + }, + { + "epoch": 0.8207913110938713, + "grad_norm": 8.242761611938477, + "learning_rate": 3.3927897693544014e-05, + "loss": 0.4089, + "step": 1058 + }, + { + "epoch": 0.821567106283941, + "grad_norm": 6.78032112121582, + "learning_rate": 3.39174835075425e-05, + "loss": 0.2856, + "step": 1059 + }, + { + "epoch": 0.8223429014740109, + "grad_norm": 2.351248025894165, + "learning_rate": 3.390702078480593e-05, + "loss": 0.0571, + "step": 1060 + }, + { + "epoch": 0.8223429014740109, + "eval_Qnli-dev_cosine_accuracy": 0.685546875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8572807312011719, + "eval_Qnli-dev_cosine_ap": 0.7260879427899244, + "eval_Qnli-dev_cosine_f1": 0.6775244299674267, + "eval_Qnli-dev_cosine_f1_threshold": 0.7407022714614868, + "eval_Qnli-dev_cosine_precision": 0.5502645502645502, + "eval_Qnli-dev_cosine_recall": 0.8813559322033898, + "eval_Qnli-dev_dot_accuracy": 0.63671875, + "eval_Qnli-dev_dot_accuracy_threshold": 457.49664306640625, + "eval_Qnli-dev_dot_ap": 0.6286501144442942, + "eval_Qnli-dev_dot_f1": 0.6731078904991948, + "eval_Qnli-dev_dot_f1_threshold": 377.1715087890625, + "eval_Qnli-dev_dot_precision": 0.5428571428571428, + "eval_Qnli-dev_dot_recall": 0.885593220338983, + "eval_Qnli-dev_euclidean_accuracy": 0.685546875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.524063110351562, + "eval_Qnli-dev_euclidean_ap": 0.7222909344445767, + "eval_Qnli-dev_euclidean_f1": 0.6779141104294478, + "eval_Qnli-dev_euclidean_f1_threshold": 17.603818893432617, + "eval_Qnli-dev_euclidean_precision": 0.53125, + "eval_Qnli-dev_euclidean_recall": 0.9364406779661016, + "eval_Qnli-dev_manhattan_accuracy": 0.681640625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 225.10433959960938, + "eval_Qnli-dev_manhattan_ap": 0.718828756798489, + "eval_Qnli-dev_manhattan_f1": 0.6766020864381519, + "eval_Qnli-dev_manhattan_f1_threshold": 324.259765625, + "eval_Qnli-dev_manhattan_precision": 0.5218390804597701, + "eval_Qnli-dev_manhattan_recall": 0.961864406779661, + "eval_Qnli-dev_max_accuracy": 0.685546875, + "eval_Qnli-dev_max_accuracy_threshold": 457.49664306640625, + "eval_Qnli-dev_max_ap": 0.7260879427899244, + "eval_Qnli-dev_max_f1": 0.6779141104294478, + "eval_Qnli-dev_max_f1_threshold": 377.1715087890625, + "eval_Qnli-dev_max_precision": 0.5502645502645502, + "eval_Qnli-dev_max_recall": 0.961864406779661, + "eval_allNLI-dev_cosine_accuracy": 0.720703125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8911333680152893, + "eval_allNLI-dev_cosine_ap": 0.5964489037579743, + "eval_allNLI-dev_cosine_f1": 0.5924170616113743, + "eval_allNLI-dev_cosine_f1_threshold": 0.8230544328689575, + "eval_allNLI-dev_cosine_precision": 0.5020080321285141, + "eval_allNLI-dev_cosine_recall": 0.7225433526011561, + "eval_allNLI-dev_dot_accuracy": 0.671875, + "eval_allNLI-dev_dot_accuracy_threshold": 522.8793334960938, + "eval_allNLI-dev_dot_ap": 0.4715419145035368, + "eval_allNLI-dev_dot_f1": 0.5573192239858906, + "eval_allNLI-dev_dot_f1_threshold": 358.88092041015625, + "eval_allNLI-dev_dot_precision": 0.4010152284263959, + "eval_allNLI-dev_dot_recall": 0.9132947976878613, + "eval_allNLI-dev_euclidean_accuracy": 0.720703125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 9.75027847290039, + "eval_allNLI-dev_euclidean_ap": 0.5982347605384456, + "eval_allNLI-dev_euclidean_f1": 0.6022471910112359, + "eval_allNLI-dev_euclidean_f1_threshold": 13.982678413391113, + "eval_allNLI-dev_euclidean_precision": 0.49264705882352944, + "eval_allNLI-dev_euclidean_recall": 0.7745664739884393, + "eval_allNLI-dev_manhattan_accuracy": 0.734375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 185.0028076171875, + "eval_allNLI-dev_manhattan_ap": 0.6068177025911232, + "eval_allNLI-dev_manhattan_f1": 0.6140350877192983, + "eval_allNLI-dev_manhattan_f1_threshold": 255.35751342773438, + "eval_allNLI-dev_manhattan_precision": 0.49469964664310956, + "eval_allNLI-dev_manhattan_recall": 0.8092485549132948, + "eval_allNLI-dev_max_accuracy": 0.734375, + "eval_allNLI-dev_max_accuracy_threshold": 522.8793334960938, + "eval_allNLI-dev_max_ap": 0.6068177025911232, + "eval_allNLI-dev_max_f1": 0.6140350877192983, + "eval_allNLI-dev_max_f1_threshold": 358.88092041015625, + "eval_allNLI-dev_max_precision": 0.5020080321285141, + "eval_allNLI-dev_max_recall": 0.9132947976878613, + "eval_sequential_score": 0.7260879427899244, + "eval_sts-test_pearson_cosine": 0.8262657403046215, + "eval_sts-test_pearson_dot": 0.7712637125208547, + "eval_sts-test_pearson_euclidean": 0.8577234471396207, + "eval_sts-test_pearson_manhattan": 0.8565991424934236, + "eval_sts-test_pearson_max": 0.8577234471396207, + "eval_sts-test_spearman_cosine": 0.8587271664545963, + "eval_sts-test_spearman_dot": 0.7626295160267625, + "eval_sts-test_spearman_euclidean": 0.8522862964075778, + "eval_sts-test_spearman_manhattan": 0.8523196360973144, + "eval_sts-test_spearman_max": 0.8587271664545963, + "eval_vitaminc-pairs_loss": 3.4151241779327393, + "eval_vitaminc-pairs_runtime": 5.4431, + "eval_vitaminc-pairs_samples_per_second": 23.516, + "eval_vitaminc-pairs_steps_per_second": 0.367, + "step": 1060 + }, + { + "epoch": 0.8223429014740109, + "eval_negation-triplets_loss": 1.8282644748687744, + "eval_negation-triplets_runtime": 1.0655, + "eval_negation-triplets_samples_per_second": 120.129, + "eval_negation-triplets_steps_per_second": 1.877, + "step": 1060 + }, + { + "epoch": 0.8223429014740109, + "eval_scitail-pairs-pos_loss": 0.12452009320259094, + "eval_scitail-pairs-pos_runtime": 1.2748, + "eval_scitail-pairs-pos_samples_per_second": 100.411, + "eval_scitail-pairs-pos_steps_per_second": 1.569, + "step": 1060 + }, + { + "epoch": 0.8223429014740109, + "eval_scitail-pairs-qa_loss": 0.024065986275672913, + "eval_scitail-pairs-qa_runtime": 0.7828, + "eval_scitail-pairs-qa_samples_per_second": 163.513, + "eval_scitail-pairs-qa_steps_per_second": 2.555, + "step": 1060 + }, + { + "epoch": 0.8223429014740109, + "eval_xsum-pairs_loss": 0.28645166754722595, + "eval_xsum-pairs_runtime": 3.928, + "eval_xsum-pairs_samples_per_second": 32.587, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 1060 + }, + { + "epoch": 0.8223429014740109, + "eval_sciq_pairs_loss": 0.06582558155059814, + "eval_sciq_pairs_runtime": 5.9299, + "eval_sciq_pairs_samples_per_second": 21.586, + "eval_sciq_pairs_steps_per_second": 0.337, + "step": 1060 + }, + { + "epoch": 0.8223429014740109, + "eval_qasc_pairs_loss": 0.4072370231151581, + "eval_qasc_pairs_runtime": 0.8995, + "eval_qasc_pairs_samples_per_second": 142.304, + "eval_qasc_pairs_steps_per_second": 2.223, + "step": 1060 + }, + { + "epoch": 0.8223429014740109, + "eval_openbookqa_pairs_loss": 1.319349765777588, + "eval_openbookqa_pairs_runtime": 0.8567, + "eval_openbookqa_pairs_samples_per_second": 149.416, + "eval_openbookqa_pairs_steps_per_second": 2.335, + "step": 1060 + }, + { + "epoch": 0.8223429014740109, + "eval_msmarco_pairs_loss": 0.5258224606513977, + "eval_msmarco_pairs_runtime": 1.9599, + "eval_msmarco_pairs_samples_per_second": 65.309, + "eval_msmarco_pairs_steps_per_second": 1.02, + "step": 1060 + }, + { + "epoch": 0.8223429014740109, + "eval_nq_pairs_loss": 0.9063143730163574, + "eval_nq_pairs_runtime": 3.3193, + "eval_nq_pairs_samples_per_second": 38.563, + "eval_nq_pairs_steps_per_second": 0.603, + "step": 1060 + }, + { + "epoch": 0.8223429014740109, + "eval_trivia_pairs_loss": 0.9630963206291199, + "eval_trivia_pairs_runtime": 4.467, + "eval_trivia_pairs_samples_per_second": 28.654, + "eval_trivia_pairs_steps_per_second": 0.448, + "step": 1060 + }, + { + "epoch": 0.8223429014740109, + "eval_gooaq_pairs_loss": 0.734113335609436, + "eval_gooaq_pairs_runtime": 1.4569, + "eval_gooaq_pairs_samples_per_second": 87.858, + "eval_gooaq_pairs_steps_per_second": 1.373, + "step": 1060 + }, + { + "epoch": 0.8223429014740109, + "eval_paws-pos_loss": 0.03285708278417587, + "eval_paws-pos_runtime": 1.0238, + "eval_paws-pos_samples_per_second": 125.029, + "eval_paws-pos_steps_per_second": 1.954, + "step": 1060 + }, + { + "epoch": 0.8223429014740109, + "eval_global_dataset_loss": 0.41156744956970215, + "eval_global_dataset_runtime": 19.3856, + "eval_global_dataset_samples_per_second": 21.459, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1060 + }, + { + "epoch": 0.8231186966640807, + "grad_norm": 7.60694694519043, + "learning_rate": 3.3896509559956894e-05, + "loss": 0.3969, + "step": 1061 + }, + { + "epoch": 0.8238944918541505, + "grad_norm": 9.652764320373535, + "learning_rate": 3.388594986777853e-05, + "loss": 0.7813, + "step": 1062 + }, + { + "epoch": 0.8246702870442203, + "grad_norm": 3.640479564666748, + "learning_rate": 3.387534174321427e-05, + "loss": 0.0861, + "step": 1063 + }, + { + "epoch": 0.8254460822342902, + "grad_norm": 2.244710922241211, + "learning_rate": 3.386468522136791e-05, + "loss": 0.046, + "step": 1064 + }, + { + "epoch": 0.8262218774243599, + "grad_norm": 2.5353660583496094, + "learning_rate": 3.385398033750333e-05, + "loss": 0.065, + "step": 1065 + }, + { + "epoch": 0.8269976726144298, + "grad_norm": 6.420120716094971, + "learning_rate": 3.3843227127044495e-05, + "loss": 0.3961, + "step": 1066 + }, + { + "epoch": 0.8277734678044996, + "grad_norm": 11.906704902648926, + "learning_rate": 3.3832425625575255e-05, + "loss": 0.9256, + "step": 1067 + }, + { + "epoch": 0.8285492629945694, + "grad_norm": 11.832637786865234, + "learning_rate": 3.382157586883928e-05, + "loss": 0.8869, + "step": 1068 + }, + { + "epoch": 0.8293250581846393, + "grad_norm": 6.807697296142578, + "learning_rate": 3.381067789273992e-05, + "loss": 0.2412, + "step": 1069 + }, + { + "epoch": 0.8301008533747091, + "grad_norm": 5.5007476806640625, + "learning_rate": 3.379973173334008e-05, + "loss": 0.2397, + "step": 1070 + }, + { + "epoch": 0.8308766485647789, + "grad_norm": 7.73779821395874, + "learning_rate": 3.378873742686212e-05, + "loss": 0.7689, + "step": 1071 + }, + { + "epoch": 0.8316524437548487, + "grad_norm": 8.138158798217773, + "learning_rate": 3.3777695009687726e-05, + "loss": 0.4859, + "step": 1072 + }, + { + "epoch": 0.8324282389449186, + "grad_norm": 13.420025825500488, + "learning_rate": 3.376660451835779e-05, + "loss": 1.3741, + "step": 1073 + }, + { + "epoch": 0.8332040341349883, + "grad_norm": 6.765268802642822, + "learning_rate": 3.375546598957228e-05, + "loss": 0.2996, + "step": 1074 + }, + { + "epoch": 0.8339798293250582, + "grad_norm": 7.034034729003906, + "learning_rate": 3.374427946019013e-05, + "loss": 0.3983, + "step": 1075 + }, + { + "epoch": 0.834755624515128, + "grad_norm": 6.580236434936523, + "learning_rate": 3.3733044967229115e-05, + "loss": 0.294, + "step": 1076 + }, + { + "epoch": 0.8355314197051978, + "grad_norm": 9.310775756835938, + "learning_rate": 3.372176254786574e-05, + "loss": 0.9269, + "step": 1077 + }, + { + "epoch": 0.8363072148952676, + "grad_norm": 10.471474647521973, + "learning_rate": 3.3710432239435075e-05, + "loss": 0.9298, + "step": 1078 + }, + { + "epoch": 0.8370830100853375, + "grad_norm": 6.726208686828613, + "learning_rate": 3.3699054079430696e-05, + "loss": 0.3399, + "step": 1079 + }, + { + "epoch": 0.8378588052754073, + "grad_norm": 9.386270523071289, + "learning_rate": 3.36876281055045e-05, + "loss": 0.5445, + "step": 1080 + }, + { + "epoch": 0.8378588052754073, + "eval_Qnli-dev_cosine_accuracy": 0.673828125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.869576096534729, + "eval_Qnli-dev_cosine_ap": 0.7180215732053381, + "eval_Qnli-dev_cosine_f1": 0.6754098360655738, + "eval_Qnli-dev_cosine_f1_threshold": 0.7540361881256104, + "eval_Qnli-dev_cosine_precision": 0.5508021390374331, + "eval_Qnli-dev_cosine_recall": 0.8728813559322034, + "eval_Qnli-dev_dot_accuracy": 0.63671875, + "eval_Qnli-dev_dot_accuracy_threshold": 398.443603515625, + "eval_Qnli-dev_dot_ap": 0.6220324524023767, + "eval_Qnli-dev_dot_f1": 0.6676557863501484, + "eval_Qnli-dev_dot_f1_threshold": 319.01409912109375, + "eval_Qnli-dev_dot_precision": 0.5136986301369864, + "eval_Qnli-dev_dot_recall": 0.9533898305084746, + "eval_Qnli-dev_euclidean_accuracy": 0.671875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.089141845703125, + "eval_Qnli-dev_euclidean_ap": 0.7160892733788555, + "eval_Qnli-dev_euclidean_f1": 0.6742671009771988, + "eval_Qnli-dev_euclidean_f1_threshold": 15.264583587646484, + "eval_Qnli-dev_euclidean_precision": 0.5476190476190477, + "eval_Qnli-dev_euclidean_recall": 0.8771186440677966, + "eval_Qnli-dev_manhattan_accuracy": 0.673828125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 218.53875732421875, + "eval_Qnli-dev_manhattan_ap": 0.7135863815437172, + "eval_Qnli-dev_manhattan_f1": 0.6793650793650794, + "eval_Qnli-dev_manhattan_f1_threshold": 282.01385498046875, + "eval_Qnli-dev_manhattan_precision": 0.5431472081218274, + "eval_Qnli-dev_manhattan_recall": 0.9067796610169492, + "eval_Qnli-dev_max_accuracy": 0.673828125, + "eval_Qnli-dev_max_accuracy_threshold": 398.443603515625, + "eval_Qnli-dev_max_ap": 0.7180215732053381, + "eval_Qnli-dev_max_f1": 0.6793650793650794, + "eval_Qnli-dev_max_f1_threshold": 319.01409912109375, + "eval_Qnli-dev_max_precision": 0.5508021390374331, + "eval_Qnli-dev_max_recall": 0.9533898305084746, + "eval_allNLI-dev_cosine_accuracy": 0.724609375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8663487434387207, + "eval_allNLI-dev_cosine_ap": 0.599379397865133, + "eval_allNLI-dev_cosine_f1": 0.6136865342163355, + "eval_allNLI-dev_cosine_f1_threshold": 0.8112990856170654, + "eval_allNLI-dev_cosine_precision": 0.49642857142857144, + "eval_allNLI-dev_cosine_recall": 0.8034682080924855, + "eval_allNLI-dev_dot_accuracy": 0.681640625, + "eval_allNLI-dev_dot_accuracy_threshold": 426.448486328125, + "eval_allNLI-dev_dot_ap": 0.4955331820931158, + "eval_allNLI-dev_dot_f1": 0.5661764705882353, + "eval_allNLI-dev_dot_f1_threshold": 330.9703369140625, + "eval_allNLI-dev_dot_precision": 0.41509433962264153, + "eval_allNLI-dev_dot_recall": 0.8901734104046243, + "eval_allNLI-dev_euclidean_accuracy": 0.73046875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.493402481079102, + "eval_allNLI-dev_euclidean_ap": 0.6022417820623912, + "eval_allNLI-dev_euclidean_f1": 0.6163793103448275, + "eval_allNLI-dev_euclidean_f1_threshold": 13.069513320922852, + "eval_allNLI-dev_euclidean_precision": 0.49140893470790376, + "eval_allNLI-dev_euclidean_recall": 0.8265895953757225, + "eval_allNLI-dev_manhattan_accuracy": 0.734375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 184.45361328125, + "eval_allNLI-dev_manhattan_ap": 0.6113726981673174, + "eval_allNLI-dev_manhattan_f1": 0.6238938053097345, + "eval_allNLI-dev_manhattan_f1_threshold": 233.15341186523438, + "eval_allNLI-dev_manhattan_precision": 0.5053763440860215, + "eval_allNLI-dev_manhattan_recall": 0.815028901734104, + "eval_allNLI-dev_max_accuracy": 0.734375, + "eval_allNLI-dev_max_accuracy_threshold": 426.448486328125, + "eval_allNLI-dev_max_ap": 0.6113726981673174, + "eval_allNLI-dev_max_f1": 0.6238938053097345, + "eval_allNLI-dev_max_f1_threshold": 330.9703369140625, + "eval_allNLI-dev_max_precision": 0.5053763440860215, + "eval_allNLI-dev_max_recall": 0.8901734104046243, + "eval_sequential_score": 0.7180215732053381, + "eval_sts-test_pearson_cosine": 0.8238103736964475, + "eval_sts-test_pearson_dot": 0.7665972384478967, + "eval_sts-test_pearson_euclidean": 0.8573278603705008, + "eval_sts-test_pearson_manhattan": 0.8538372489899912, + "eval_sts-test_pearson_max": 0.8573278603705008, + "eval_sts-test_spearman_cosine": 0.854271909875514, + "eval_sts-test_spearman_dot": 0.7511187755767125, + "eval_sts-test_spearman_euclidean": 0.8533233173878007, + "eval_sts-test_spearman_manhattan": 0.8510165241168582, + "eval_sts-test_spearman_max": 0.854271909875514, + "eval_vitaminc-pairs_loss": 3.536039352416992, + "eval_vitaminc-pairs_runtime": 5.432, + "eval_vitaminc-pairs_samples_per_second": 23.564, + "eval_vitaminc-pairs_steps_per_second": 0.368, + "step": 1080 + }, + { + "epoch": 0.8378588052754073, + "eval_negation-triplets_loss": 1.7508964538574219, + "eval_negation-triplets_runtime": 1.0813, + "eval_negation-triplets_samples_per_second": 118.372, + "eval_negation-triplets_steps_per_second": 1.85, + "step": 1080 + }, + { + "epoch": 0.8378588052754073, + "eval_scitail-pairs-pos_loss": 0.12378459423780441, + "eval_scitail-pairs-pos_runtime": 1.4821, + "eval_scitail-pairs-pos_samples_per_second": 86.363, + "eval_scitail-pairs-pos_steps_per_second": 1.349, + "step": 1080 + }, + { + "epoch": 0.8378588052754073, + "eval_scitail-pairs-qa_loss": 0.01887408457696438, + "eval_scitail-pairs-qa_runtime": 0.8006, + "eval_scitail-pairs-qa_samples_per_second": 159.882, + "eval_scitail-pairs-qa_steps_per_second": 2.498, + "step": 1080 + }, + { + "epoch": 0.8378588052754073, + "eval_xsum-pairs_loss": 0.2877923250198364, + "eval_xsum-pairs_runtime": 3.9235, + "eval_xsum-pairs_samples_per_second": 32.624, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 1080 + }, + { + "epoch": 0.8378588052754073, + "eval_sciq_pairs_loss": 0.055665355175733566, + "eval_sciq_pairs_runtime": 5.9547, + "eval_sciq_pairs_samples_per_second": 21.496, + "eval_sciq_pairs_steps_per_second": 0.336, + "step": 1080 + }, + { + "epoch": 0.8378588052754073, + "eval_qasc_pairs_loss": 0.3689579367637634, + "eval_qasc_pairs_runtime": 0.9125, + "eval_qasc_pairs_samples_per_second": 140.281, + "eval_qasc_pairs_steps_per_second": 2.192, + "step": 1080 + }, + { + "epoch": 0.8378588052754073, + "eval_openbookqa_pairs_loss": 1.2368128299713135, + "eval_openbookqa_pairs_runtime": 0.8648, + "eval_openbookqa_pairs_samples_per_second": 148.008, + "eval_openbookqa_pairs_steps_per_second": 2.313, + "step": 1080 + }, + { + "epoch": 0.8378588052754073, + "eval_msmarco_pairs_loss": 0.4975939989089966, + "eval_msmarco_pairs_runtime": 1.96, + "eval_msmarco_pairs_samples_per_second": 65.305, + "eval_msmarco_pairs_steps_per_second": 1.02, + "step": 1080 + }, + { + "epoch": 0.8378588052754073, + "eval_nq_pairs_loss": 0.8287155032157898, + "eval_nq_pairs_runtime": 3.3301, + "eval_nq_pairs_samples_per_second": 38.437, + "eval_nq_pairs_steps_per_second": 0.601, + "step": 1080 + }, + { + "epoch": 0.8378588052754073, + "eval_trivia_pairs_loss": 0.8384897708892822, + "eval_trivia_pairs_runtime": 4.4719, + "eval_trivia_pairs_samples_per_second": 28.623, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 1080 + }, + { + "epoch": 0.8378588052754073, + "eval_gooaq_pairs_loss": 0.7341352701187134, + "eval_gooaq_pairs_runtime": 1.4637, + "eval_gooaq_pairs_samples_per_second": 87.45, + "eval_gooaq_pairs_steps_per_second": 1.366, + "step": 1080 + }, + { + "epoch": 0.8378588052754073, + "eval_paws-pos_loss": 0.034420281648635864, + "eval_paws-pos_runtime": 1.0225, + "eval_paws-pos_samples_per_second": 125.179, + "eval_paws-pos_steps_per_second": 1.956, + "step": 1080 + }, + { + "epoch": 0.8378588052754073, + "eval_global_dataset_loss": 0.3913795053958893, + "eval_global_dataset_runtime": 19.3816, + "eval_global_dataset_samples_per_second": 21.464, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1080 + }, + { + "epoch": 0.8386346004654771, + "grad_norm": 5.978991985321045, + "learning_rate": 3.367615435546662e-05, + "loss": 0.3299, + "step": 1081 + }, + { + "epoch": 0.839410395655547, + "grad_norm": 5.057347297668457, + "learning_rate": 3.366463286728529e-05, + "loss": 0.2673, + "step": 1082 + }, + { + "epoch": 0.8401861908456167, + "grad_norm": 7.968233585357666, + "learning_rate": 3.3653063679086714e-05, + "loss": 0.324, + "step": 1083 + }, + { + "epoch": 0.8409619860356866, + "grad_norm": 9.063628196716309, + "learning_rate": 3.3641446829154926e-05, + "loss": 0.5237, + "step": 1084 + }, + { + "epoch": 0.8417377812257564, + "grad_norm": 13.249706268310547, + "learning_rate": 3.36297823559317e-05, + "loss": 1.196, + "step": 1085 + }, + { + "epoch": 0.8425135764158262, + "grad_norm": 8.703137397766113, + "learning_rate": 3.361807029801639e-05, + "loss": 0.4114, + "step": 1086 + }, + { + "epoch": 0.843289371605896, + "grad_norm": 7.457031726837158, + "learning_rate": 3.3606310694165826e-05, + "loss": 0.2476, + "step": 1087 + }, + { + "epoch": 0.8440651667959659, + "grad_norm": 5.185329437255859, + "learning_rate": 3.359450358329416e-05, + "loss": 0.1737, + "step": 1088 + }, + { + "epoch": 0.8448409619860356, + "grad_norm": 9.03635025024414, + "learning_rate": 3.3582649004472754e-05, + "loss": 0.4646, + "step": 1089 + }, + { + "epoch": 0.8456167571761055, + "grad_norm": 11.819198608398438, + "learning_rate": 3.3570746996930063e-05, + "loss": 0.8031, + "step": 1090 + }, + { + "epoch": 0.8463925523661753, + "grad_norm": 9.829190254211426, + "learning_rate": 3.355879760005147e-05, + "loss": 0.5742, + "step": 1091 + }, + { + "epoch": 0.8471683475562451, + "grad_norm": 9.514701843261719, + "learning_rate": 3.3546800853379184e-05, + "loss": 0.6093, + "step": 1092 + }, + { + "epoch": 0.847944142746315, + "grad_norm": 9.587857246398926, + "learning_rate": 3.35347567966121e-05, + "loss": 0.5605, + "step": 1093 + }, + { + "epoch": 0.8487199379363848, + "grad_norm": 7.565169811248779, + "learning_rate": 3.3522665469605685e-05, + "loss": 0.4818, + "step": 1094 + }, + { + "epoch": 0.8494957331264547, + "grad_norm": 18.12430191040039, + "learning_rate": 3.35105269123718e-05, + "loss": 1.5096, + "step": 1095 + }, + { + "epoch": 0.8502715283165244, + "grad_norm": 10.062978744506836, + "learning_rate": 3.349834116507863e-05, + "loss": 0.8714, + "step": 1096 + }, + { + "epoch": 0.8510473235065943, + "grad_norm": 16.100385665893555, + "learning_rate": 3.3486108268050474e-05, + "loss": 1.456, + "step": 1097 + }, + { + "epoch": 0.851823118696664, + "grad_norm": 10.831403732299805, + "learning_rate": 3.347382826176771e-05, + "loss": 0.8546, + "step": 1098 + }, + { + "epoch": 0.8525989138867339, + "grad_norm": 6.318869590759277, + "learning_rate": 3.3461501186866555e-05, + "loss": 0.1871, + "step": 1099 + }, + { + "epoch": 0.8533747090768037, + "grad_norm": 6.361988544464111, + "learning_rate": 3.3449127084139034e-05, + "loss": 0.3433, + "step": 1100 + }, + { + "epoch": 0.8533747090768037, + "eval_Qnli-dev_cosine_accuracy": 0.666015625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.9060459136962891, + "eval_Qnli-dev_cosine_ap": 0.7150552289234171, + "eval_Qnli-dev_cosine_f1": 0.6805111821086263, + "eval_Qnli-dev_cosine_f1_threshold": 0.7571378946304321, + "eval_Qnli-dev_cosine_precision": 0.5461538461538461, + "eval_Qnli-dev_cosine_recall": 0.902542372881356, + "eval_Qnli-dev_dot_accuracy": 0.63671875, + "eval_Qnli-dev_dot_accuracy_threshold": 424.96337890625, + "eval_Qnli-dev_dot_ap": 0.6066700689789146, + "eval_Qnli-dev_dot_f1": 0.6686746987951808, + "eval_Qnli-dev_dot_f1_threshold": 350.5804443359375, + "eval_Qnli-dev_dot_precision": 0.5186915887850467, + "eval_Qnli-dev_dot_recall": 0.940677966101695, + "eval_Qnli-dev_euclidean_accuracy": 0.669921875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.542449951171875, + "eval_Qnli-dev_euclidean_ap": 0.7139861428547295, + "eval_Qnli-dev_euclidean_f1": 0.6758620689655173, + "eval_Qnli-dev_euclidean_f1_threshold": 14.56290340423584, + "eval_Qnli-dev_euclidean_precision": 0.5697674418604651, + "eval_Qnli-dev_euclidean_recall": 0.8305084745762712, + "eval_Qnli-dev_manhattan_accuracy": 0.671875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 206.10891723632812, + "eval_Qnli-dev_manhattan_ap": 0.7185312701941471, + "eval_Qnli-dev_manhattan_f1": 0.674457429048414, + "eval_Qnli-dev_manhattan_f1_threshold": 269.8748474121094, + "eval_Qnli-dev_manhattan_precision": 0.5564738292011019, + "eval_Qnli-dev_manhattan_recall": 0.8559322033898306, + "eval_Qnli-dev_max_accuracy": 0.671875, + "eval_Qnli-dev_max_accuracy_threshold": 424.96337890625, + "eval_Qnli-dev_max_ap": 0.7185312701941471, + "eval_Qnli-dev_max_f1": 0.6805111821086263, + "eval_Qnli-dev_max_f1_threshold": 350.5804443359375, + "eval_Qnli-dev_max_precision": 0.5697674418604651, + "eval_Qnli-dev_max_recall": 0.940677966101695, + "eval_allNLI-dev_cosine_accuracy": 0.7265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8948431015014648, + "eval_allNLI-dev_cosine_ap": 0.5855770107338941, + "eval_allNLI-dev_cosine_f1": 0.6116071428571429, + "eval_allNLI-dev_cosine_f1_threshold": 0.8248633146286011, + "eval_allNLI-dev_cosine_precision": 0.49818181818181817, + "eval_allNLI-dev_cosine_recall": 0.791907514450867, + "eval_allNLI-dev_dot_accuracy": 0.6796875, + "eval_allNLI-dev_dot_accuracy_threshold": 495.8543701171875, + "eval_allNLI-dev_dot_ap": 0.4990258380621768, + "eval_allNLI-dev_dot_f1": 0.5782178217821782, + "eval_allNLI-dev_dot_f1_threshold": 382.0567626953125, + "eval_allNLI-dev_dot_precision": 0.4397590361445783, + "eval_allNLI-dev_dot_recall": 0.8439306358381503, + "eval_allNLI-dev_euclidean_accuracy": 0.724609375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.160776138305664, + "eval_allNLI-dev_euclidean_ap": 0.5876760698421187, + "eval_allNLI-dev_euclidean_f1": 0.6124197002141327, + "eval_allNLI-dev_euclidean_f1_threshold": 13.415067672729492, + "eval_allNLI-dev_euclidean_precision": 0.48639455782312924, + "eval_allNLI-dev_euclidean_recall": 0.8265895953757225, + "eval_allNLI-dev_manhattan_accuracy": 0.732421875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 177.86550903320312, + "eval_allNLI-dev_manhattan_ap": 0.593752511755563, + "eval_allNLI-dev_manhattan_f1": 0.614406779661017, + "eval_allNLI-dev_manhattan_f1_threshold": 244.60800170898438, + "eval_allNLI-dev_manhattan_precision": 0.48494983277591974, + "eval_allNLI-dev_manhattan_recall": 0.838150289017341, + "eval_allNLI-dev_max_accuracy": 0.732421875, + "eval_allNLI-dev_max_accuracy_threshold": 495.8543701171875, + "eval_allNLI-dev_max_ap": 0.593752511755563, + "eval_allNLI-dev_max_f1": 0.614406779661017, + "eval_allNLI-dev_max_f1_threshold": 382.0567626953125, + "eval_allNLI-dev_max_precision": 0.49818181818181817, + "eval_allNLI-dev_max_recall": 0.8439306358381503, + "eval_sequential_score": 0.7185312701941471, + "eval_sts-test_pearson_cosine": 0.8233318839841255, + "eval_sts-test_pearson_dot": 0.7756995878481125, + "eval_sts-test_pearson_euclidean": 0.8595467372349506, + "eval_sts-test_pearson_manhattan": 0.8576196098552569, + "eval_sts-test_pearson_max": 0.8595467372349506, + "eval_sts-test_spearman_cosine": 0.8586427655756123, + "eval_sts-test_spearman_dot": 0.762784713401268, + "eval_sts-test_spearman_euclidean": 0.8551757511018362, + "eval_sts-test_spearman_manhattan": 0.8543319213170402, + "eval_sts-test_spearman_max": 0.8586427655756123, + "eval_vitaminc-pairs_loss": 3.771732807159424, + "eval_vitaminc-pairs_runtime": 5.4551, + "eval_vitaminc-pairs_samples_per_second": 23.464, + "eval_vitaminc-pairs_steps_per_second": 0.367, + "step": 1100 + }, + { + "epoch": 0.8533747090768037, + "eval_negation-triplets_loss": 1.755806565284729, + "eval_negation-triplets_runtime": 1.0841, + "eval_negation-triplets_samples_per_second": 118.07, + "eval_negation-triplets_steps_per_second": 1.845, + "step": 1100 + }, + { + "epoch": 0.8533747090768037, + "eval_scitail-pairs-pos_loss": 0.12076473236083984, + "eval_scitail-pairs-pos_runtime": 1.2506, + "eval_scitail-pairs-pos_samples_per_second": 102.348, + "eval_scitail-pairs-pos_steps_per_second": 1.599, + "step": 1100 + }, + { + "epoch": 0.8533747090768037, + "eval_scitail-pairs-qa_loss": 0.022578276693820953, + "eval_scitail-pairs-qa_runtime": 0.7677, + "eval_scitail-pairs-qa_samples_per_second": 166.725, + "eval_scitail-pairs-qa_steps_per_second": 2.605, + "step": 1100 + }, + { + "epoch": 0.8533747090768037, + "eval_xsum-pairs_loss": 0.19214841723442078, + "eval_xsum-pairs_runtime": 3.9052, + "eval_xsum-pairs_samples_per_second": 32.777, + "eval_xsum-pairs_steps_per_second": 0.512, + "step": 1100 + }, + { + "epoch": 0.8533747090768037, + "eval_sciq_pairs_loss": 0.06406810134649277, + "eval_sciq_pairs_runtime": 5.9252, + "eval_sciq_pairs_samples_per_second": 21.603, + "eval_sciq_pairs_steps_per_second": 0.338, + "step": 1100 + }, + { + "epoch": 0.8533747090768037, + "eval_qasc_pairs_loss": 0.3945878744125366, + "eval_qasc_pairs_runtime": 0.9044, + "eval_qasc_pairs_samples_per_second": 141.537, + "eval_qasc_pairs_steps_per_second": 2.212, + "step": 1100 + }, + { + "epoch": 0.8533747090768037, + "eval_openbookqa_pairs_loss": 1.2332226037979126, + "eval_openbookqa_pairs_runtime": 0.8629, + "eval_openbookqa_pairs_samples_per_second": 148.332, + "eval_openbookqa_pairs_steps_per_second": 2.318, + "step": 1100 + }, + { + "epoch": 0.8533747090768037, + "eval_msmarco_pairs_loss": 0.440255343914032, + "eval_msmarco_pairs_runtime": 1.9608, + "eval_msmarco_pairs_samples_per_second": 65.278, + "eval_msmarco_pairs_steps_per_second": 1.02, + "step": 1100 + }, + { + "epoch": 0.8533747090768037, + "eval_nq_pairs_loss": 0.8072230815887451, + "eval_nq_pairs_runtime": 3.3186, + "eval_nq_pairs_samples_per_second": 38.571, + "eval_nq_pairs_steps_per_second": 0.603, + "step": 1100 + }, + { + "epoch": 0.8533747090768037, + "eval_trivia_pairs_loss": 0.8540432453155518, + "eval_trivia_pairs_runtime": 4.4616, + "eval_trivia_pairs_samples_per_second": 28.69, + "eval_trivia_pairs_steps_per_second": 0.448, + "step": 1100 + }, + { + "epoch": 0.8533747090768037, + "eval_gooaq_pairs_loss": 0.7283532619476318, + "eval_gooaq_pairs_runtime": 1.4568, + "eval_gooaq_pairs_samples_per_second": 87.863, + "eval_gooaq_pairs_steps_per_second": 1.373, + "step": 1100 + }, + { + "epoch": 0.8533747090768037, + "eval_paws-pos_loss": 0.03369026631116867, + "eval_paws-pos_runtime": 1.025, + "eval_paws-pos_samples_per_second": 124.873, + "eval_paws-pos_steps_per_second": 1.951, + "step": 1100 + }, + { + "epoch": 0.8533747090768037, + "eval_global_dataset_loss": 0.46180859208106995, + "eval_global_dataset_runtime": 19.3932, + "eval_global_dataset_samples_per_second": 21.451, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1100 + }, + { + "epoch": 0.8541505042668736, + "grad_norm": 6.8964433670043945, + "learning_rate": 3.343670599453275e-05, + "loss": 0.3694, + "step": 1101 + }, + { + "epoch": 0.8549262994569433, + "grad_norm": 9.510549545288086, + "learning_rate": 3.3424237959150814e-05, + "loss": 0.7678, + "step": 1102 + }, + { + "epoch": 0.8557020946470132, + "grad_norm": 6.011091709136963, + "learning_rate": 3.3411723019251684e-05, + "loss": 0.3401, + "step": 1103 + }, + { + "epoch": 0.8564778898370831, + "grad_norm": 8.616118431091309, + "learning_rate": 3.3399161216249026e-05, + "loss": 0.52, + "step": 1104 + }, + { + "epoch": 0.8572536850271528, + "grad_norm": 6.648637294769287, + "learning_rate": 3.3386552591711594e-05, + "loss": 0.3925, + "step": 1105 + }, + { + "epoch": 0.8580294802172227, + "grad_norm": 4.9895339012146, + "learning_rate": 3.3373897187363065e-05, + "loss": 0.3324, + "step": 1106 + }, + { + "epoch": 0.8588052754072925, + "grad_norm": 7.069686412811279, + "learning_rate": 3.3361195045081934e-05, + "loss": 0.4209, + "step": 1107 + }, + { + "epoch": 0.8595810705973623, + "grad_norm": 7.353325366973877, + "learning_rate": 3.3348446206901336e-05, + "loss": 0.4282, + "step": 1108 + }, + { + "epoch": 0.8603568657874321, + "grad_norm": 12.576475143432617, + "learning_rate": 3.3335650715008954e-05, + "loss": 1.0367, + "step": 1109 + }, + { + "epoch": 0.861132660977502, + "grad_norm": 7.3023881912231445, + "learning_rate": 3.332280861174684e-05, + "loss": 0.2631, + "step": 1110 + }, + { + "epoch": 0.8619084561675717, + "grad_norm": 14.896276473999023, + "learning_rate": 3.33099199396113e-05, + "loss": 1.4839, + "step": 1111 + }, + { + "epoch": 0.8626842513576416, + "grad_norm": 9.623770713806152, + "learning_rate": 3.329698474125272e-05, + "loss": 0.6218, + "step": 1112 + }, + { + "epoch": 0.8634600465477114, + "grad_norm": 6.295792102813721, + "learning_rate": 3.3284003059475486e-05, + "loss": 0.29, + "step": 1113 + }, + { + "epoch": 0.8642358417377812, + "grad_norm": 0.40645697712898254, + "learning_rate": 3.327097493723776e-05, + "loss": 0.0079, + "step": 1114 + }, + { + "epoch": 0.8650116369278511, + "grad_norm": 7.739372253417969, + "learning_rate": 3.325790041765141e-05, + "loss": 0.4895, + "step": 1115 + }, + { + "epoch": 0.8657874321179209, + "grad_norm": 10.867382049560547, + "learning_rate": 3.3244779543981845e-05, + "loss": 0.9476, + "step": 1116 + }, + { + "epoch": 0.8665632273079907, + "grad_norm": 7.589806079864502, + "learning_rate": 3.323161235964784e-05, + "loss": 0.3517, + "step": 1117 + }, + { + "epoch": 0.8673390224980605, + "grad_norm": 11.332128524780273, + "learning_rate": 3.321839890822143e-05, + "loss": 0.8738, + "step": 1118 + }, + { + "epoch": 0.8681148176881304, + "grad_norm": 7.2625322341918945, + "learning_rate": 3.3205139233427764e-05, + "loss": 0.4498, + "step": 1119 + }, + { + "epoch": 0.8688906128782001, + "grad_norm": 1.2745305299758911, + "learning_rate": 3.3191833379144944e-05, + "loss": 0.0683, + "step": 1120 + }, + { + "epoch": 0.8688906128782001, + "eval_Qnli-dev_cosine_accuracy": 0.671875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8667693138122559, + "eval_Qnli-dev_cosine_ap": 0.7216075708845244, + "eval_Qnli-dev_cosine_f1": 0.6807131280388979, + "eval_Qnli-dev_cosine_f1_threshold": 0.7539348602294922, + "eval_Qnli-dev_cosine_precision": 0.5511811023622047, + "eval_Qnli-dev_cosine_recall": 0.8898305084745762, + "eval_Qnli-dev_dot_accuracy": 0.65234375, + "eval_Qnli-dev_dot_accuracy_threshold": 417.77288818359375, + "eval_Qnli-dev_dot_ap": 0.6314746960508899, + "eval_Qnli-dev_dot_f1": 0.6735042735042736, + "eval_Qnli-dev_dot_f1_threshold": 385.18084716796875, + "eval_Qnli-dev_dot_precision": 0.5644699140401146, + "eval_Qnli-dev_dot_recall": 0.8347457627118644, + "eval_Qnli-dev_euclidean_accuracy": 0.673828125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.011968612670898, + "eval_Qnli-dev_euclidean_ap": 0.7172904617835556, + "eval_Qnli-dev_euclidean_f1": 0.6835443037974683, + "eval_Qnli-dev_euclidean_f1_threshold": 16.50353240966797, + "eval_Qnli-dev_euclidean_precision": 0.5454545454545454, + "eval_Qnli-dev_euclidean_recall": 0.9152542372881356, + "eval_Qnli-dev_manhattan_accuracy": 0.669921875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 219.5309600830078, + "eval_Qnli-dev_manhattan_ap": 0.7157019938548889, + "eval_Qnli-dev_manhattan_f1": 0.6688741721854304, + "eval_Qnli-dev_manhattan_f1_threshold": 282.58013916015625, + "eval_Qnli-dev_manhattan_precision": 0.5489130434782609, + "eval_Qnli-dev_manhattan_recall": 0.8559322033898306, + "eval_Qnli-dev_max_accuracy": 0.673828125, + "eval_Qnli-dev_max_accuracy_threshold": 417.77288818359375, + "eval_Qnli-dev_max_ap": 0.7216075708845244, + "eval_Qnli-dev_max_f1": 0.6835443037974683, + "eval_Qnli-dev_max_f1_threshold": 385.18084716796875, + "eval_Qnli-dev_max_precision": 0.5644699140401146, + "eval_Qnli-dev_max_recall": 0.9152542372881356, + "eval_allNLI-dev_cosine_accuracy": 0.708984375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8957781791687012, + "eval_allNLI-dev_cosine_ap": 0.5764689241929717, + "eval_allNLI-dev_cosine_f1": 0.5901639344262295, + "eval_allNLI-dev_cosine_f1_threshold": 0.83022540807724, + "eval_allNLI-dev_cosine_precision": 0.49606299212598426, + "eval_allNLI-dev_cosine_recall": 0.7283236994219653, + "eval_allNLI-dev_dot_accuracy": 0.6796875, + "eval_allNLI-dev_dot_accuracy_threshold": 431.19696044921875, + "eval_allNLI-dev_dot_ap": 0.47909983434000686, + "eval_allNLI-dev_dot_f1": 0.5684647302904564, + "eval_allNLI-dev_dot_f1_threshold": 373.99139404296875, + "eval_allNLI-dev_dot_precision": 0.44336569579288027, + "eval_allNLI-dev_dot_recall": 0.791907514450867, + "eval_allNLI-dev_euclidean_accuracy": 0.7109375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.513742446899414, + "eval_allNLI-dev_euclidean_ap": 0.5790223122824858, + "eval_allNLI-dev_euclidean_f1": 0.5962441314553991, + "eval_allNLI-dev_euclidean_f1_threshold": 12.818342208862305, + "eval_allNLI-dev_euclidean_precision": 0.5019762845849802, + "eval_allNLI-dev_euclidean_recall": 0.7341040462427746, + "eval_allNLI-dev_manhattan_accuracy": 0.720703125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 169.86944580078125, + "eval_allNLI-dev_manhattan_ap": 0.58228137006283, + "eval_allNLI-dev_manhattan_f1": 0.598503740648379, + "eval_allNLI-dev_manhattan_f1_threshold": 221.64862060546875, + "eval_allNLI-dev_manhattan_precision": 0.5263157894736842, + "eval_allNLI-dev_manhattan_recall": 0.6936416184971098, + "eval_allNLI-dev_max_accuracy": 0.720703125, + "eval_allNLI-dev_max_accuracy_threshold": 431.19696044921875, + "eval_allNLI-dev_max_ap": 0.58228137006283, + "eval_allNLI-dev_max_f1": 0.598503740648379, + "eval_allNLI-dev_max_f1_threshold": 373.99139404296875, + "eval_allNLI-dev_max_precision": 0.5263157894736842, + "eval_allNLI-dev_max_recall": 0.791907514450867, + "eval_sequential_score": 0.7216075708845244, + "eval_sts-test_pearson_cosine": 0.8287680600746106, + "eval_sts-test_pearson_dot": 0.7778473667344845, + "eval_sts-test_pearson_euclidean": 0.8603328275716268, + "eval_sts-test_pearson_manhattan": 0.8588967039738422, + "eval_sts-test_pearson_max": 0.8603328275716268, + "eval_sts-test_spearman_cosine": 0.8625052021950425, + "eval_sts-test_spearman_dot": 0.7610209319360488, + "eval_sts-test_spearman_euclidean": 0.8579373514422995, + "eval_sts-test_spearman_manhattan": 0.8564934072422447, + "eval_sts-test_spearman_max": 0.8625052021950425, + "eval_vitaminc-pairs_loss": 3.571171998977661, + "eval_vitaminc-pairs_runtime": 5.4788, + "eval_vitaminc-pairs_samples_per_second": 23.363, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 1120 + }, + { + "epoch": 0.8688906128782001, + "eval_negation-triplets_loss": 1.7726892232894897, + "eval_negation-triplets_runtime": 1.0715, + "eval_negation-triplets_samples_per_second": 119.46, + "eval_negation-triplets_steps_per_second": 1.867, + "step": 1120 + }, + { + "epoch": 0.8688906128782001, + "eval_scitail-pairs-pos_loss": 0.11432746797800064, + "eval_scitail-pairs-pos_runtime": 1.3154, + "eval_scitail-pairs-pos_samples_per_second": 97.31, + "eval_scitail-pairs-pos_steps_per_second": 1.52, + "step": 1120 + }, + { + "epoch": 0.8688906128782001, + "eval_scitail-pairs-qa_loss": 0.015412604436278343, + "eval_scitail-pairs-qa_runtime": 0.7721, + "eval_scitail-pairs-qa_samples_per_second": 165.787, + "eval_scitail-pairs-qa_steps_per_second": 2.59, + "step": 1120 + }, + { + "epoch": 0.8688906128782001, + "eval_xsum-pairs_loss": 0.14450086653232574, + "eval_xsum-pairs_runtime": 3.9224, + "eval_xsum-pairs_samples_per_second": 32.633, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 1120 + }, + { + "epoch": 0.8688906128782001, + "eval_sciq_pairs_loss": 0.06374696642160416, + "eval_sciq_pairs_runtime": 5.9681, + "eval_sciq_pairs_samples_per_second": 21.447, + "eval_sciq_pairs_steps_per_second": 0.335, + "step": 1120 + }, + { + "epoch": 0.8688906128782001, + "eval_qasc_pairs_loss": 0.38392698764801025, + "eval_qasc_pairs_runtime": 0.9106, + "eval_qasc_pairs_samples_per_second": 140.563, + "eval_qasc_pairs_steps_per_second": 2.196, + "step": 1120 + }, + { + "epoch": 0.8688906128782001, + "eval_openbookqa_pairs_loss": 1.2659245729446411, + "eval_openbookqa_pairs_runtime": 0.868, + "eval_openbookqa_pairs_samples_per_second": 147.469, + "eval_openbookqa_pairs_steps_per_second": 2.304, + "step": 1120 + }, + { + "epoch": 0.8688906128782001, + "eval_msmarco_pairs_loss": 0.4464336335659027, + "eval_msmarco_pairs_runtime": 1.9641, + "eval_msmarco_pairs_samples_per_second": 65.171, + "eval_msmarco_pairs_steps_per_second": 1.018, + "step": 1120 + }, + { + "epoch": 0.8688906128782001, + "eval_nq_pairs_loss": 0.7987036108970642, + "eval_nq_pairs_runtime": 3.3432, + "eval_nq_pairs_samples_per_second": 38.287, + "eval_nq_pairs_steps_per_second": 0.598, + "step": 1120 + }, + { + "epoch": 0.8688906128782001, + "eval_trivia_pairs_loss": 0.8050705790519714, + "eval_trivia_pairs_runtime": 4.4763, + "eval_trivia_pairs_samples_per_second": 28.595, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 1120 + }, + { + "epoch": 0.8688906128782001, + "eval_gooaq_pairs_loss": 0.8165881633758545, + "eval_gooaq_pairs_runtime": 1.4648, + "eval_gooaq_pairs_samples_per_second": 87.385, + "eval_gooaq_pairs_steps_per_second": 1.365, + "step": 1120 + }, + { + "epoch": 0.8688906128782001, + "eval_paws-pos_loss": 0.03345727548003197, + "eval_paws-pos_runtime": 1.0322, + "eval_paws-pos_samples_per_second": 124.011, + "eval_paws-pos_steps_per_second": 1.938, + "step": 1120 + }, + { + "epoch": 0.8688906128782001, + "eval_global_dataset_loss": 0.40634939074516296, + "eval_global_dataset_runtime": 19.3975, + "eval_global_dataset_samples_per_second": 21.446, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1120 + }, + { + "epoch": 0.86966640806827, + "grad_norm": 7.208463191986084, + "learning_rate": 3.3178481389403884e-05, + "loss": 0.3213, + "step": 1121 + }, + { + "epoch": 0.8704422032583398, + "grad_norm": 8.684393882751465, + "learning_rate": 3.316508330838816e-05, + "loss": 0.4863, + "step": 1122 + }, + { + "epoch": 0.8712179984484096, + "grad_norm": 7.740796089172363, + "learning_rate": 3.315163918043389e-05, + "loss": 0.3312, + "step": 1123 + }, + { + "epoch": 0.8719937936384794, + "grad_norm": 0.7196974754333496, + "learning_rate": 3.3138149050029555e-05, + "loss": 0.0147, + "step": 1124 + }, + { + "epoch": 0.8727695888285493, + "grad_norm": 8.58254623413086, + "learning_rate": 3.3124612961815857e-05, + "loss": 0.4571, + "step": 1125 + }, + { + "epoch": 0.873545384018619, + "grad_norm": 7.641439437866211, + "learning_rate": 3.3111030960585585e-05, + "loss": 0.3704, + "step": 1126 + }, + { + "epoch": 0.8743211792086889, + "grad_norm": 8.70718765258789, + "learning_rate": 3.309740309128348e-05, + "loss": 0.6807, + "step": 1127 + }, + { + "epoch": 0.8750969743987588, + "grad_norm": 9.00788402557373, + "learning_rate": 3.308372939900603e-05, + "loss": 0.4869, + "step": 1128 + }, + { + "epoch": 0.8758727695888285, + "grad_norm": 12.836403846740723, + "learning_rate": 3.3070009929001394e-05, + "loss": 1.3684, + "step": 1129 + }, + { + "epoch": 0.8766485647788984, + "grad_norm": 6.269355297088623, + "learning_rate": 3.305624472666918e-05, + "loss": 0.4082, + "step": 1130 + }, + { + "epoch": 0.8774243599689682, + "grad_norm": 7.749275207519531, + "learning_rate": 3.304243383756037e-05, + "loss": 0.588, + "step": 1131 + }, + { + "epoch": 0.878200155159038, + "grad_norm": 8.381994247436523, + "learning_rate": 3.302857730737711e-05, + "loss": 0.8064, + "step": 1132 + }, + { + "epoch": 0.8789759503491078, + "grad_norm": 9.036953926086426, + "learning_rate": 3.301467518197255e-05, + "loss": 0.7681, + "step": 1133 + }, + { + "epoch": 0.8797517455391777, + "grad_norm": 6.172411918640137, + "learning_rate": 3.3000727507350785e-05, + "loss": 0.2437, + "step": 1134 + }, + { + "epoch": 0.8805275407292474, + "grad_norm": 7.784179210662842, + "learning_rate": 3.298673432966657e-05, + "loss": 0.4845, + "step": 1135 + }, + { + "epoch": 0.8813033359193173, + "grad_norm": 14.057782173156738, + "learning_rate": 3.29726956952253e-05, + "loss": 1.2761, + "step": 1136 + }, + { + "epoch": 0.8820791311093871, + "grad_norm": 10.141071319580078, + "learning_rate": 3.2958611650482736e-05, + "loss": 0.6671, + "step": 1137 + }, + { + "epoch": 0.8828549262994569, + "grad_norm": 8.106549263000488, + "learning_rate": 3.294448224204494e-05, + "loss": 0.4414, + "step": 1138 + }, + { + "epoch": 0.8836307214895268, + "grad_norm": 6.273714542388916, + "learning_rate": 3.2930307516668086e-05, + "loss": 0.3697, + "step": 1139 + }, + { + "epoch": 0.8844065166795966, + "grad_norm": 5.151251316070557, + "learning_rate": 3.29160875212583e-05, + "loss": 0.1816, + "step": 1140 + }, + { + "epoch": 0.8844065166795966, + "eval_Qnli-dev_cosine_accuracy": 0.677734375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8563957214355469, + "eval_Qnli-dev_cosine_ap": 0.7257789162753756, + "eval_Qnli-dev_cosine_f1": 0.6831683168316831, + "eval_Qnli-dev_cosine_f1_threshold": 0.7716048955917358, + "eval_Qnli-dev_cosine_precision": 0.5594594594594594, + "eval_Qnli-dev_cosine_recall": 0.8771186440677966, + "eval_Qnli-dev_dot_accuracy": 0.63671875, + "eval_Qnli-dev_dot_accuracy_threshold": 398.0260314941406, + "eval_Qnli-dev_dot_ap": 0.6223035818024821, + "eval_Qnli-dev_dot_f1": 0.6782006920415226, + "eval_Qnli-dev_dot_f1_threshold": 398.0260314941406, + "eval_Qnli-dev_dot_precision": 0.5730994152046783, + "eval_Qnli-dev_dot_recall": 0.8305084745762712, + "eval_Qnli-dev_euclidean_accuracy": 0.673828125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.425256729125977, + "eval_Qnli-dev_euclidean_ap": 0.7200441560863342, + "eval_Qnli-dev_euclidean_f1": 0.6887417218543046, + "eval_Qnli-dev_euclidean_f1_threshold": 15.509228706359863, + "eval_Qnli-dev_euclidean_precision": 0.5652173913043478, + "eval_Qnli-dev_euclidean_recall": 0.8813559322033898, + "eval_Qnli-dev_manhattan_accuracy": 0.669921875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 198.95169067382812, + "eval_Qnli-dev_manhattan_ap": 0.7194976331442562, + "eval_Qnli-dev_manhattan_f1": 0.6789297658862876, + "eval_Qnli-dev_manhattan_f1_threshold": 273.83905029296875, + "eval_Qnli-dev_manhattan_precision": 0.5607734806629834, + "eval_Qnli-dev_manhattan_recall": 0.8601694915254238, + "eval_Qnli-dev_max_accuracy": 0.677734375, + "eval_Qnli-dev_max_accuracy_threshold": 398.0260314941406, + "eval_Qnli-dev_max_ap": 0.7257789162753756, + "eval_Qnli-dev_max_f1": 0.6887417218543046, + "eval_Qnli-dev_max_f1_threshold": 398.0260314941406, + "eval_Qnli-dev_max_precision": 0.5730994152046783, + "eval_Qnli-dev_max_recall": 0.8813559322033898, + "eval_allNLI-dev_cosine_accuracy": 0.720703125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9072520136833191, + "eval_allNLI-dev_cosine_ap": 0.5906279895231493, + "eval_allNLI-dev_cosine_f1": 0.6086956521739132, + "eval_allNLI-dev_cosine_f1_threshold": 0.841861367225647, + "eval_allNLI-dev_cosine_precision": 0.5228215767634855, + "eval_allNLI-dev_cosine_recall": 0.7283236994219653, + "eval_allNLI-dev_dot_accuracy": 0.677734375, + "eval_allNLI-dev_dot_accuracy_threshold": 487.79150390625, + "eval_allNLI-dev_dot_ap": 0.4876016685384582, + "eval_allNLI-dev_dot_f1": 0.5787401574803149, + "eval_allNLI-dev_dot_f1_threshold": 385.33197021484375, + "eval_allNLI-dev_dot_precision": 0.4388059701492537, + "eval_allNLI-dev_dot_recall": 0.8497109826589595, + "eval_allNLI-dev_euclidean_accuracy": 0.71484375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.259764671325684, + "eval_allNLI-dev_euclidean_ap": 0.5922098863280456, + "eval_allNLI-dev_euclidean_f1": 0.6136363636363636, + "eval_allNLI-dev_euclidean_f1_threshold": 13.233564376831055, + "eval_allNLI-dev_euclidean_precision": 0.5056179775280899, + "eval_allNLI-dev_euclidean_recall": 0.7803468208092486, + "eval_allNLI-dev_manhattan_accuracy": 0.72265625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 181.18411254882812, + "eval_allNLI-dev_manhattan_ap": 0.5919468944839975, + "eval_allNLI-dev_manhattan_f1": 0.609271523178808, + "eval_allNLI-dev_manhattan_f1_threshold": 245.84173583984375, + "eval_allNLI-dev_manhattan_precision": 0.4928571428571429, + "eval_allNLI-dev_manhattan_recall": 0.7976878612716763, + "eval_allNLI-dev_max_accuracy": 0.72265625, + "eval_allNLI-dev_max_accuracy_threshold": 487.79150390625, + "eval_allNLI-dev_max_ap": 0.5922098863280456, + "eval_allNLI-dev_max_f1": 0.6136363636363636, + "eval_allNLI-dev_max_f1_threshold": 385.33197021484375, + "eval_allNLI-dev_max_precision": 0.5228215767634855, + "eval_allNLI-dev_max_recall": 0.8497109826589595, + "eval_sequential_score": 0.7257789162753756, + "eval_sts-test_pearson_cosine": 0.8254371960878275, + "eval_sts-test_pearson_dot": 0.734758791936589, + "eval_sts-test_pearson_euclidean": 0.8625549886936091, + "eval_sts-test_pearson_manhattan": 0.8604733593784164, + "eval_sts-test_pearson_max": 0.8625549886936091, + "eval_sts-test_spearman_cosine": 0.8639197448163632, + "eval_sts-test_spearman_dot": 0.7110987965446571, + "eval_sts-test_spearman_euclidean": 0.8601369654422095, + "eval_sts-test_spearman_manhattan": 0.8597421250629533, + "eval_sts-test_spearman_max": 0.8639197448163632, + "eval_vitaminc-pairs_loss": 3.5746660232543945, + "eval_vitaminc-pairs_runtime": 5.4705, + "eval_vitaminc-pairs_samples_per_second": 23.398, + "eval_vitaminc-pairs_steps_per_second": 0.366, + "step": 1140 + }, + { + "epoch": 0.8844065166795966, + "eval_negation-triplets_loss": 1.8934292793273926, + "eval_negation-triplets_runtime": 1.0906, + "eval_negation-triplets_samples_per_second": 117.372, + "eval_negation-triplets_steps_per_second": 1.834, + "step": 1140 + }, + { + "epoch": 0.8844065166795966, + "eval_scitail-pairs-pos_loss": 0.1276829093694687, + "eval_scitail-pairs-pos_runtime": 1.2695, + "eval_scitail-pairs-pos_samples_per_second": 100.825, + "eval_scitail-pairs-pos_steps_per_second": 1.575, + "step": 1140 + }, + { + "epoch": 0.8844065166795966, + "eval_scitail-pairs-qa_loss": 0.012894031591713428, + "eval_scitail-pairs-qa_runtime": 0.7693, + "eval_scitail-pairs-qa_samples_per_second": 166.393, + "eval_scitail-pairs-qa_steps_per_second": 2.6, + "step": 1140 + }, + { + "epoch": 0.8844065166795966, + "eval_xsum-pairs_loss": 0.17931480705738068, + "eval_xsum-pairs_runtime": 3.9136, + "eval_xsum-pairs_samples_per_second": 32.707, + "eval_xsum-pairs_steps_per_second": 0.511, + "step": 1140 + }, + { + "epoch": 0.8844065166795966, + "eval_sciq_pairs_loss": 0.05700588971376419, + "eval_sciq_pairs_runtime": 5.9428, + "eval_sciq_pairs_samples_per_second": 21.539, + "eval_sciq_pairs_steps_per_second": 0.337, + "step": 1140 + }, + { + "epoch": 0.8844065166795966, + "eval_qasc_pairs_loss": 0.30698150396347046, + "eval_qasc_pairs_runtime": 0.9032, + "eval_qasc_pairs_samples_per_second": 141.72, + "eval_qasc_pairs_steps_per_second": 2.214, + "step": 1140 + }, + { + "epoch": 0.8844065166795966, + "eval_openbookqa_pairs_loss": 1.3146690130233765, + "eval_openbookqa_pairs_runtime": 0.8637, + "eval_openbookqa_pairs_samples_per_second": 148.204, + "eval_openbookqa_pairs_steps_per_second": 2.316, + "step": 1140 + }, + { + "epoch": 0.8844065166795966, + "eval_msmarco_pairs_loss": 0.4545725882053375, + "eval_msmarco_pairs_runtime": 1.9747, + "eval_msmarco_pairs_samples_per_second": 64.819, + "eval_msmarco_pairs_steps_per_second": 1.013, + "step": 1140 + }, + { + "epoch": 0.8844065166795966, + "eval_nq_pairs_loss": 0.8171166777610779, + "eval_nq_pairs_runtime": 3.3227, + "eval_nq_pairs_samples_per_second": 38.523, + "eval_nq_pairs_steps_per_second": 0.602, + "step": 1140 + }, + { + "epoch": 0.8844065166795966, + "eval_trivia_pairs_loss": 0.8382545709609985, + "eval_trivia_pairs_runtime": 4.4737, + "eval_trivia_pairs_samples_per_second": 28.612, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 1140 + }, + { + "epoch": 0.8844065166795966, + "eval_gooaq_pairs_loss": 0.7004772424697876, + "eval_gooaq_pairs_runtime": 1.4688, + "eval_gooaq_pairs_samples_per_second": 87.147, + "eval_gooaq_pairs_steps_per_second": 1.362, + "step": 1140 + }, + { + "epoch": 0.8844065166795966, + "eval_paws-pos_loss": 0.03230363875627518, + "eval_paws-pos_runtime": 1.0323, + "eval_paws-pos_samples_per_second": 123.995, + "eval_paws-pos_steps_per_second": 1.937, + "step": 1140 + }, + { + "epoch": 0.8844065166795966, + "eval_global_dataset_loss": 0.4007374048233032, + "eval_global_dataset_runtime": 19.401, + "eval_global_dataset_samples_per_second": 21.442, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1140 + }, + { + "epoch": 0.8851823118696664, + "grad_norm": 3.9553866386413574, + "learning_rate": 3.290182230287153e-05, + "loss": 0.1317, + "step": 1141 + }, + { + "epoch": 0.8859581070597362, + "grad_norm": 4.835943222045898, + "learning_rate": 3.288751190871335e-05, + "loss": 0.195, + "step": 1142 + }, + { + "epoch": 0.8867339022498061, + "grad_norm": 7.370544910430908, + "learning_rate": 3.287315638613882e-05, + "loss": 0.4922, + "step": 1143 + }, + { + "epoch": 0.8875096974398758, + "grad_norm": 13.102246284484863, + "learning_rate": 3.285875578265239e-05, + "loss": 1.2099, + "step": 1144 + }, + { + "epoch": 0.8882854926299457, + "grad_norm": 7.960138320922852, + "learning_rate": 3.2844310145907626e-05, + "loss": 0.7187, + "step": 1145 + }, + { + "epoch": 0.8890612878200155, + "grad_norm": 7.201216220855713, + "learning_rate": 3.282981952370714e-05, + "loss": 0.3269, + "step": 1146 + }, + { + "epoch": 0.8898370830100854, + "grad_norm": 8.31344223022461, + "learning_rate": 3.281528396400242e-05, + "loss": 0.4128, + "step": 1147 + }, + { + "epoch": 0.8906128782001551, + "grad_norm": 8.452710151672363, + "learning_rate": 3.280070351489364e-05, + "loss": 0.5341, + "step": 1148 + }, + { + "epoch": 0.891388673390225, + "grad_norm": 10.184416770935059, + "learning_rate": 3.278607822462954e-05, + "loss": 0.6201, + "step": 1149 + }, + { + "epoch": 0.8921644685802949, + "grad_norm": 10.648394584655762, + "learning_rate": 3.277140814160722e-05, + "loss": 0.693, + "step": 1150 + }, + { + "epoch": 0.8929402637703646, + "grad_norm": 10.509428977966309, + "learning_rate": 3.275669331437201e-05, + "loss": 0.6365, + "step": 1151 + }, + { + "epoch": 0.8937160589604345, + "grad_norm": 11.796065330505371, + "learning_rate": 3.2741933791617324e-05, + "loss": 0.9307, + "step": 1152 + }, + { + "epoch": 0.8944918541505043, + "grad_norm": 1.182183861732483, + "learning_rate": 3.272712962218445e-05, + "loss": 0.0301, + "step": 1153 + }, + { + "epoch": 0.8952676493405741, + "grad_norm": 10.76645278930664, + "learning_rate": 3.271228085506247e-05, + "loss": 0.6476, + "step": 1154 + }, + { + "epoch": 0.8960434445306439, + "grad_norm": 11.166770935058594, + "learning_rate": 3.269738753938798e-05, + "loss": 0.9999, + "step": 1155 + }, + { + "epoch": 0.8968192397207138, + "grad_norm": 5.118910789489746, + "learning_rate": 3.268244972444504e-05, + "loss": 0.2122, + "step": 1156 + }, + { + "epoch": 0.8975950349107835, + "grad_norm": 8.023819923400879, + "learning_rate": 3.266746745966494e-05, + "loss": 0.893, + "step": 1157 + }, + { + "epoch": 0.8983708301008534, + "grad_norm": 7.175548076629639, + "learning_rate": 3.2652440794626086e-05, + "loss": 0.4017, + "step": 1158 + }, + { + "epoch": 0.8991466252909232, + "grad_norm": 10.204700469970703, + "learning_rate": 3.263736977905378e-05, + "loss": 0.6845, + "step": 1159 + }, + { + "epoch": 0.899922420480993, + "grad_norm": 1.8034473657608032, + "learning_rate": 3.262225446282009e-05, + "loss": 0.0683, + "step": 1160 + }, + { + "epoch": 0.899922420480993, + "eval_Qnli-dev_cosine_accuracy": 0.6640625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8921190500259399, + "eval_Qnli-dev_cosine_ap": 0.7040066291792803, + "eval_Qnli-dev_cosine_f1": 0.6790697674418604, + "eval_Qnli-dev_cosine_f1_threshold": 0.7328351140022278, + "eval_Qnli-dev_cosine_precision": 0.5354523227383863, + "eval_Qnli-dev_cosine_recall": 0.9279661016949152, + "eval_Qnli-dev_dot_accuracy": 0.640625, + "eval_Qnli-dev_dot_accuracy_threshold": 429.1894226074219, + "eval_Qnli-dev_dot_ap": 0.6093531457523953, + "eval_Qnli-dev_dot_f1": 0.6666666666666667, + "eval_Qnli-dev_dot_f1_threshold": 347.73388671875, + "eval_Qnli-dev_dot_precision": 0.5150115473441108, + "eval_Qnli-dev_dot_recall": 0.9449152542372882, + "eval_Qnli-dev_euclidean_accuracy": 0.662109375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 10.652650833129883, + "eval_Qnli-dev_euclidean_ap": 0.7039494977460571, + "eval_Qnli-dev_euclidean_f1": 0.6813186813186812, + "eval_Qnli-dev_euclidean_f1_threshold": 16.35511589050293, + "eval_Qnli-dev_euclidean_precision": 0.5411471321695761, + "eval_Qnli-dev_euclidean_recall": 0.9194915254237288, + "eval_Qnli-dev_manhattan_accuracy": 0.671875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 233.37313842773438, + "eval_Qnli-dev_manhattan_ap": 0.7010698941511505, + "eval_Qnli-dev_manhattan_f1": 0.6770670826833073, + "eval_Qnli-dev_manhattan_f1_threshold": 299.50537109375, + "eval_Qnli-dev_manhattan_precision": 0.5358024691358024, + "eval_Qnli-dev_manhattan_recall": 0.9194915254237288, + "eval_Qnli-dev_max_accuracy": 0.671875, + "eval_Qnli-dev_max_accuracy_threshold": 429.1894226074219, + "eval_Qnli-dev_max_ap": 0.7040066291792803, + "eval_Qnli-dev_max_f1": 0.6813186813186812, + "eval_Qnli-dev_max_f1_threshold": 347.73388671875, + "eval_Qnli-dev_max_precision": 0.5411471321695761, + "eval_Qnli-dev_max_recall": 0.9449152542372882, + "eval_allNLI-dev_cosine_accuracy": 0.720703125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8813140392303467, + "eval_allNLI-dev_cosine_ap": 0.5796734635494606, + "eval_allNLI-dev_cosine_f1": 0.6045454545454545, + "eval_allNLI-dev_cosine_f1_threshold": 0.8260512351989746, + "eval_allNLI-dev_cosine_precision": 0.49812734082397003, + "eval_allNLI-dev_cosine_recall": 0.7687861271676301, + "eval_allNLI-dev_dot_accuracy": 0.673828125, + "eval_allNLI-dev_dot_accuracy_threshold": 420.40533447265625, + "eval_allNLI-dev_dot_ap": 0.4683631605272184, + "eval_allNLI-dev_dot_f1": 0.5666666666666667, + "eval_allNLI-dev_dot_f1_threshold": 381.01446533203125, + "eval_allNLI-dev_dot_precision": 0.44299674267100975, + "eval_allNLI-dev_dot_recall": 0.7861271676300579, + "eval_allNLI-dev_euclidean_accuracy": 0.72265625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.13191032409668, + "eval_allNLI-dev_euclidean_ap": 0.5811069224601753, + "eval_allNLI-dev_euclidean_f1": 0.6216216216216217, + "eval_allNLI-dev_euclidean_f1_threshold": 13.219721794128418, + "eval_allNLI-dev_euclidean_precision": 0.5092250922509225, + "eval_allNLI-dev_euclidean_recall": 0.7976878612716763, + "eval_allNLI-dev_manhattan_accuracy": 0.720703125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 185.72914123535156, + "eval_allNLI-dev_manhattan_ap": 0.5840391331047329, + "eval_allNLI-dev_manhattan_f1": 0.6125290023201856, + "eval_allNLI-dev_manhattan_f1_threshold": 234.300048828125, + "eval_allNLI-dev_manhattan_precision": 0.5116279069767442, + "eval_allNLI-dev_manhattan_recall": 0.7630057803468208, + "eval_allNLI-dev_max_accuracy": 0.72265625, + "eval_allNLI-dev_max_accuracy_threshold": 420.40533447265625, + "eval_allNLI-dev_max_ap": 0.5840391331047329, + "eval_allNLI-dev_max_f1": 0.6216216216216217, + "eval_allNLI-dev_max_f1_threshold": 381.01446533203125, + "eval_allNLI-dev_max_precision": 0.5116279069767442, + "eval_allNLI-dev_max_recall": 0.7976878612716763, + "eval_sequential_score": 0.7040066291792803, + "eval_sts-test_pearson_cosine": 0.831515109639266, + "eval_sts-test_pearson_dot": 0.7648449414420361, + "eval_sts-test_pearson_euclidean": 0.8697880700065697, + "eval_sts-test_pearson_manhattan": 0.869084559282981, + "eval_sts-test_pearson_max": 0.8697880700065697, + "eval_sts-test_spearman_cosine": 0.8720129656877365, + "eval_sts-test_spearman_dot": 0.7547109593286, + "eval_sts-test_spearman_euclidean": 0.8681458749371214, + "eval_sts-test_spearman_manhattan": 0.8684129952163667, + "eval_sts-test_spearman_max": 0.8720129656877365, + "eval_vitaminc-pairs_loss": 3.9453749656677246, + "eval_vitaminc-pairs_runtime": 5.4739, + "eval_vitaminc-pairs_samples_per_second": 23.384, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 1160 + }, + { + "epoch": 0.899922420480993, + "eval_negation-triplets_loss": 1.7579838037490845, + "eval_negation-triplets_runtime": 1.071, + "eval_negation-triplets_samples_per_second": 119.516, + "eval_negation-triplets_steps_per_second": 1.867, + "step": 1160 + }, + { + "epoch": 0.899922420480993, + "eval_scitail-pairs-pos_loss": 0.1167030856013298, + "eval_scitail-pairs-pos_runtime": 1.2724, + "eval_scitail-pairs-pos_samples_per_second": 100.6, + "eval_scitail-pairs-pos_steps_per_second": 1.572, + "step": 1160 + }, + { + "epoch": 0.899922420480993, + "eval_scitail-pairs-qa_loss": 0.012686913833022118, + "eval_scitail-pairs-qa_runtime": 0.7725, + "eval_scitail-pairs-qa_samples_per_second": 165.702, + "eval_scitail-pairs-qa_steps_per_second": 2.589, + "step": 1160 + }, + { + "epoch": 0.899922420480993, + "eval_xsum-pairs_loss": 0.17685747146606445, + "eval_xsum-pairs_runtime": 3.9204, + "eval_xsum-pairs_samples_per_second": 32.65, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 1160 + }, + { + "epoch": 0.899922420480993, + "eval_sciq_pairs_loss": 0.0558762326836586, + "eval_sciq_pairs_runtime": 6.0519, + "eval_sciq_pairs_samples_per_second": 21.15, + "eval_sciq_pairs_steps_per_second": 0.33, + "step": 1160 + }, + { + "epoch": 0.899922420480993, + "eval_qasc_pairs_loss": 0.26454707980155945, + "eval_qasc_pairs_runtime": 0.9182, + "eval_qasc_pairs_samples_per_second": 139.409, + "eval_qasc_pairs_steps_per_second": 2.178, + "step": 1160 + }, + { + "epoch": 0.899922420480993, + "eval_openbookqa_pairs_loss": 1.2756041288375854, + "eval_openbookqa_pairs_runtime": 0.8671, + "eval_openbookqa_pairs_samples_per_second": 147.623, + "eval_openbookqa_pairs_steps_per_second": 2.307, + "step": 1160 + }, + { + "epoch": 0.899922420480993, + "eval_msmarco_pairs_loss": 0.449916809797287, + "eval_msmarco_pairs_runtime": 1.9762, + "eval_msmarco_pairs_samples_per_second": 64.772, + "eval_msmarco_pairs_steps_per_second": 1.012, + "step": 1160 + }, + { + "epoch": 0.899922420480993, + "eval_nq_pairs_loss": 0.7896221280097961, + "eval_nq_pairs_runtime": 3.3436, + "eval_nq_pairs_samples_per_second": 38.282, + "eval_nq_pairs_steps_per_second": 0.598, + "step": 1160 + }, + { + "epoch": 0.899922420480993, + "eval_trivia_pairs_loss": 0.8409191370010376, + "eval_trivia_pairs_runtime": 4.4782, + "eval_trivia_pairs_samples_per_second": 28.583, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 1160 + }, + { + "epoch": 0.899922420480993, + "eval_gooaq_pairs_loss": 0.7546132802963257, + "eval_gooaq_pairs_runtime": 1.465, + "eval_gooaq_pairs_samples_per_second": 87.373, + "eval_gooaq_pairs_steps_per_second": 1.365, + "step": 1160 + }, + { + "epoch": 0.899922420480993, + "eval_paws-pos_loss": 0.033789414912462234, + "eval_paws-pos_runtime": 1.0491, + "eval_paws-pos_samples_per_second": 122.008, + "eval_paws-pos_steps_per_second": 1.906, + "step": 1160 + }, + { + "epoch": 0.899922420480993, + "eval_global_dataset_loss": 0.43357256054878235, + "eval_global_dataset_runtime": 19.4216, + "eval_global_dataset_samples_per_second": 21.419, + "eval_global_dataset_steps_per_second": 0.36, + "step": 1160 + }, + { + "epoch": 0.9006982156710628, + "grad_norm": 6.303240776062012, + "learning_rate": 3.260709489594372e-05, + "loss": 0.3874, + "step": 1161 + }, + { + "epoch": 0.9014740108611327, + "grad_norm": 6.447384357452393, + "learning_rate": 3.259189112858976e-05, + "loss": 0.1167, + "step": 1162 + }, + { + "epoch": 0.9022498060512025, + "grad_norm": 6.725493907928467, + "learning_rate": 3.2576643211069586e-05, + "loss": 0.3241, + "step": 1163 + }, + { + "epoch": 0.9030256012412723, + "grad_norm": 7.5674614906311035, + "learning_rate": 3.2561351193840667e-05, + "loss": 0.2304, + "step": 1164 + }, + { + "epoch": 0.9038013964313422, + "grad_norm": 5.545563220977783, + "learning_rate": 3.2546015127506415e-05, + "loss": 0.1767, + "step": 1165 + }, + { + "epoch": 0.9045771916214119, + "grad_norm": 12.802922248840332, + "learning_rate": 3.253063506281598e-05, + "loss": 1.0422, + "step": 1166 + }, + { + "epoch": 0.9053529868114818, + "grad_norm": 9.864930152893066, + "learning_rate": 3.251521105066415e-05, + "loss": 0.5599, + "step": 1167 + }, + { + "epoch": 0.9061287820015516, + "grad_norm": 14.451776504516602, + "learning_rate": 3.24997431420911e-05, + "loss": 1.2632, + "step": 1168 + }, + { + "epoch": 0.9069045771916214, + "grad_norm": 13.052535057067871, + "learning_rate": 3.248423138828229e-05, + "loss": 1.2131, + "step": 1169 + }, + { + "epoch": 0.9076803723816912, + "grad_norm": 2.4571800231933594, + "learning_rate": 3.246867584056827e-05, + "loss": 0.0449, + "step": 1170 + }, + { + "epoch": 0.9084561675717611, + "grad_norm": 7.134913444519043, + "learning_rate": 3.245307655042449e-05, + "loss": 0.3326, + "step": 1171 + }, + { + "epoch": 0.9092319627618308, + "grad_norm": 5.698978900909424, + "learning_rate": 3.2437433569471166e-05, + "loss": 0.3237, + "step": 1172 + }, + { + "epoch": 0.9100077579519007, + "grad_norm": 7.162409782409668, + "learning_rate": 3.2421746949473086e-05, + "loss": 0.3914, + "step": 1173 + }, + { + "epoch": 0.9107835531419706, + "grad_norm": 9.979856491088867, + "learning_rate": 3.240601674233946e-05, + "loss": 0.6666, + "step": 1174 + }, + { + "epoch": 0.9115593483320403, + "grad_norm": 13.519302368164062, + "learning_rate": 3.239024300012371e-05, + "loss": 2.0492, + "step": 1175 + }, + { + "epoch": 0.9123351435221102, + "grad_norm": 6.480223655700684, + "learning_rate": 3.2374425775023344e-05, + "loss": 0.314, + "step": 1176 + }, + { + "epoch": 0.91311093871218, + "grad_norm": 4.045516014099121, + "learning_rate": 3.235856511937975e-05, + "loss": 0.2207, + "step": 1177 + }, + { + "epoch": 0.9138867339022498, + "grad_norm": 8.722333908081055, + "learning_rate": 3.234266108567803e-05, + "loss": 0.3565, + "step": 1178 + }, + { + "epoch": 0.9146625290923196, + "grad_norm": 1.492080569267273, + "learning_rate": 3.232671372654683e-05, + "loss": 0.0341, + "step": 1179 + }, + { + "epoch": 0.9154383242823895, + "grad_norm": 7.871532917022705, + "learning_rate": 3.231072309475819e-05, + "loss": 0.2821, + "step": 1180 + }, + { + "epoch": 0.9154383242823895, + "eval_Qnli-dev_cosine_accuracy": 0.654296875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8147125244140625, + "eval_Qnli-dev_cosine_ap": 0.6997827040276372, + "eval_Qnli-dev_cosine_f1": 0.6748681898066785, + "eval_Qnli-dev_cosine_f1_threshold": 0.7724531888961792, + "eval_Qnli-dev_cosine_precision": 0.5765765765765766, + "eval_Qnli-dev_cosine_recall": 0.8135593220338984, + "eval_Qnli-dev_dot_accuracy": 0.62890625, + "eval_Qnli-dev_dot_accuracy_threshold": 426.5829162597656, + "eval_Qnli-dev_dot_ap": 0.6346953696529666, + "eval_Qnli-dev_dot_f1": 0.6687797147385103, + "eval_Qnli-dev_dot_f1_threshold": 354.88006591796875, + "eval_Qnli-dev_dot_precision": 0.5341772151898734, + "eval_Qnli-dev_dot_recall": 0.8940677966101694, + "eval_Qnli-dev_euclidean_accuracy": 0.65625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.091718673706055, + "eval_Qnli-dev_euclidean_ap": 0.6991715062525676, + "eval_Qnli-dev_euclidean_f1": 0.6794258373205743, + "eval_Qnli-dev_euclidean_f1_threshold": 16.516164779663086, + "eval_Qnli-dev_euclidean_precision": 0.5447570332480819, + "eval_Qnli-dev_euclidean_recall": 0.902542372881356, + "eval_Qnli-dev_manhattan_accuracy": 0.65625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 227.53836059570312, + "eval_Qnli-dev_manhattan_ap": 0.70060128992247, + "eval_Qnli-dev_manhattan_f1": 0.67430441898527, + "eval_Qnli-dev_manhattan_f1_threshold": 292.7884521484375, + "eval_Qnli-dev_manhattan_precision": 0.5493333333333333, + "eval_Qnli-dev_manhattan_recall": 0.8728813559322034, + "eval_Qnli-dev_max_accuracy": 0.65625, + "eval_Qnli-dev_max_accuracy_threshold": 426.5829162597656, + "eval_Qnli-dev_max_ap": 0.70060128992247, + "eval_Qnli-dev_max_f1": 0.6794258373205743, + "eval_Qnli-dev_max_f1_threshold": 354.88006591796875, + "eval_Qnli-dev_max_precision": 0.5765765765765766, + "eval_Qnli-dev_max_recall": 0.902542372881356, + "eval_allNLI-dev_cosine_accuracy": 0.708984375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9036482572555542, + "eval_allNLI-dev_cosine_ap": 0.5793069595342613, + "eval_allNLI-dev_cosine_f1": 0.6135831381733021, + "eval_allNLI-dev_cosine_f1_threshold": 0.8388439416885376, + "eval_allNLI-dev_cosine_precision": 0.515748031496063, + "eval_allNLI-dev_cosine_recall": 0.7572254335260116, + "eval_allNLI-dev_dot_accuracy": 0.67578125, + "eval_allNLI-dev_dot_accuracy_threshold": 422.08563232421875, + "eval_allNLI-dev_dot_ap": 0.4807905477839307, + "eval_allNLI-dev_dot_f1": 0.5675675675675675, + "eval_allNLI-dev_dot_f1_threshold": 357.0093994140625, + "eval_allNLI-dev_dot_precision": 0.4260869565217391, + "eval_allNLI-dev_dot_recall": 0.8497109826589595, + "eval_allNLI-dev_euclidean_accuracy": 0.71484375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.423340797424316, + "eval_allNLI-dev_euclidean_ap": 0.5817200018308355, + "eval_allNLI-dev_euclidean_f1": 0.6132075471698113, + "eval_allNLI-dev_euclidean_f1_threshold": 12.199594497680664, + "eval_allNLI-dev_euclidean_precision": 0.5179282868525896, + "eval_allNLI-dev_euclidean_recall": 0.7514450867052023, + "eval_allNLI-dev_manhattan_accuracy": 0.71875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 177.607177734375, + "eval_allNLI-dev_manhattan_ap": 0.5850872206316992, + "eval_allNLI-dev_manhattan_f1": 0.6145833333333334, + "eval_allNLI-dev_manhattan_f1_threshold": 206.32032775878906, + "eval_allNLI-dev_manhattan_precision": 0.5592417061611374, + "eval_allNLI-dev_manhattan_recall": 0.6820809248554913, + "eval_allNLI-dev_max_accuracy": 0.71875, + "eval_allNLI-dev_max_accuracy_threshold": 422.08563232421875, + "eval_allNLI-dev_max_ap": 0.5850872206316992, + "eval_allNLI-dev_max_f1": 0.6145833333333334, + "eval_allNLI-dev_max_f1_threshold": 357.0093994140625, + "eval_allNLI-dev_max_precision": 0.5592417061611374, + "eval_allNLI-dev_max_recall": 0.8497109826589595, + "eval_sequential_score": 0.70060128992247, + "eval_sts-test_pearson_cosine": 0.8246418240865199, + "eval_sts-test_pearson_dot": 0.7650783443443958, + "eval_sts-test_pearson_euclidean": 0.8657713266995477, + "eval_sts-test_pearson_manhattan": 0.8650141735698451, + "eval_sts-test_pearson_max": 0.8657713266995477, + "eval_sts-test_spearman_cosine": 0.8709703507998398, + "eval_sts-test_spearman_dot": 0.7532125528434318, + "eval_sts-test_spearman_euclidean": 0.8664657335794427, + "eval_sts-test_spearman_manhattan": 0.8662740639269033, + "eval_sts-test_spearman_max": 0.8709703507998398, + "eval_vitaminc-pairs_loss": 4.117798805236816, + "eval_vitaminc-pairs_runtime": 5.4547, + "eval_vitaminc-pairs_samples_per_second": 23.466, + "eval_vitaminc-pairs_steps_per_second": 0.367, + "step": 1180 + }, + { + "epoch": 0.9154383242823895, + "eval_negation-triplets_loss": 1.7201412916183472, + "eval_negation-triplets_runtime": 1.077, + "eval_negation-triplets_samples_per_second": 118.852, + "eval_negation-triplets_steps_per_second": 1.857, + "step": 1180 + }, + { + "epoch": 0.9154383242823895, + "eval_scitail-pairs-pos_loss": 0.16175013780593872, + "eval_scitail-pairs-pos_runtime": 1.3278, + "eval_scitail-pairs-pos_samples_per_second": 96.404, + "eval_scitail-pairs-pos_steps_per_second": 1.506, + "step": 1180 + }, + { + "epoch": 0.9154383242823895, + "eval_scitail-pairs-qa_loss": 0.007280169520527124, + "eval_scitail-pairs-qa_runtime": 0.7948, + "eval_scitail-pairs-qa_samples_per_second": 161.045, + "eval_scitail-pairs-qa_steps_per_second": 2.516, + "step": 1180 + }, + { + "epoch": 0.9154383242823895, + "eval_xsum-pairs_loss": 0.1715097427368164, + "eval_xsum-pairs_runtime": 3.911, + "eval_xsum-pairs_samples_per_second": 32.728, + "eval_xsum-pairs_steps_per_second": 0.511, + "step": 1180 + }, + { + "epoch": 0.9154383242823895, + "eval_sciq_pairs_loss": 0.0543162077665329, + "eval_sciq_pairs_runtime": 5.9518, + "eval_sciq_pairs_samples_per_second": 21.506, + "eval_sciq_pairs_steps_per_second": 0.336, + "step": 1180 + }, + { + "epoch": 0.9154383242823895, + "eval_qasc_pairs_loss": 0.3288556933403015, + "eval_qasc_pairs_runtime": 0.9121, + "eval_qasc_pairs_samples_per_second": 140.337, + "eval_qasc_pairs_steps_per_second": 2.193, + "step": 1180 + }, + { + "epoch": 0.9154383242823895, + "eval_openbookqa_pairs_loss": 1.2357536554336548, + "eval_openbookqa_pairs_runtime": 0.8726, + "eval_openbookqa_pairs_samples_per_second": 146.694, + "eval_openbookqa_pairs_steps_per_second": 2.292, + "step": 1180 + }, + { + "epoch": 0.9154383242823895, + "eval_msmarco_pairs_loss": 0.5136918425559998, + "eval_msmarco_pairs_runtime": 1.9693, + "eval_msmarco_pairs_samples_per_second": 64.996, + "eval_msmarco_pairs_steps_per_second": 1.016, + "step": 1180 + }, + { + "epoch": 0.9154383242823895, + "eval_nq_pairs_loss": 0.7663967609405518, + "eval_nq_pairs_runtime": 3.3236, + "eval_nq_pairs_samples_per_second": 38.512, + "eval_nq_pairs_steps_per_second": 0.602, + "step": 1180 + }, + { + "epoch": 0.9154383242823895, + "eval_trivia_pairs_loss": 0.8957348465919495, + "eval_trivia_pairs_runtime": 4.4796, + "eval_trivia_pairs_samples_per_second": 28.574, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 1180 + }, + { + "epoch": 0.9154383242823895, + "eval_gooaq_pairs_loss": 0.6759052872657776, + "eval_gooaq_pairs_runtime": 1.4635, + "eval_gooaq_pairs_samples_per_second": 87.461, + "eval_gooaq_pairs_steps_per_second": 1.367, + "step": 1180 + }, + { + "epoch": 0.9154383242823895, + "eval_paws-pos_loss": 0.03511948883533478, + "eval_paws-pos_runtime": 1.0259, + "eval_paws-pos_samples_per_second": 124.772, + "eval_paws-pos_steps_per_second": 1.95, + "step": 1180 + }, + { + "epoch": 0.9154383242823895, + "eval_global_dataset_loss": 0.4371417164802551, + "eval_global_dataset_runtime": 19.3984, + "eval_global_dataset_samples_per_second": 21.445, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1180 + }, + { + "epoch": 0.9162141194724592, + "grad_norm": 1.2721761465072632, + "learning_rate": 3.229468924322731e-05, + "loss": 0.0443, + "step": 1181 + }, + { + "epoch": 0.9169899146625291, + "grad_norm": 10.766345977783203, + "learning_rate": 3.227861222501243e-05, + "loss": 0.7913, + "step": 1182 + }, + { + "epoch": 0.9177657098525989, + "grad_norm": 8.42568588256836, + "learning_rate": 3.226249209331463e-05, + "loss": 0.4149, + "step": 1183 + }, + { + "epoch": 0.9185415050426687, + "grad_norm": 0.06603731960058212, + "learning_rate": 3.2246328901477664e-05, + "loss": 0.0002, + "step": 1184 + }, + { + "epoch": 0.9193173002327386, + "grad_norm": 13.39123249053955, + "learning_rate": 3.2230122702987766e-05, + "loss": 0.8619, + "step": 1185 + }, + { + "epoch": 0.9200930954228084, + "grad_norm": 3.478794574737549, + "learning_rate": 3.22138735514735e-05, + "loss": 0.0709, + "step": 1186 + }, + { + "epoch": 0.9208688906128782, + "grad_norm": 8.111926078796387, + "learning_rate": 3.219758150070555e-05, + "loss": 0.2993, + "step": 1187 + }, + { + "epoch": 0.921644685802948, + "grad_norm": 41.20555114746094, + "learning_rate": 3.218124660459657e-05, + "loss": 1.0337, + "step": 1188 + }, + { + "epoch": 0.9224204809930179, + "grad_norm": 11.006499290466309, + "learning_rate": 3.2164868917201e-05, + "loss": 0.4867, + "step": 1189 + }, + { + "epoch": 0.9231962761830876, + "grad_norm": 9.926593780517578, + "learning_rate": 3.214844849271488e-05, + "loss": 0.4353, + "step": 1190 + }, + { + "epoch": 0.9239720713731575, + "grad_norm": 12.641822814941406, + "learning_rate": 3.213198538547567e-05, + "loss": 0.9588, + "step": 1191 + }, + { + "epoch": 0.9247478665632273, + "grad_norm": 13.263992309570312, + "learning_rate": 3.211547964996206e-05, + "loss": 1.1448, + "step": 1192 + }, + { + "epoch": 0.9255236617532971, + "grad_norm": 11.031960487365723, + "learning_rate": 3.209893134079382e-05, + "loss": 0.8281, + "step": 1193 + }, + { + "epoch": 0.9262994569433669, + "grad_norm": 0.0, + "learning_rate": 3.2082340512731614e-05, + "loss": 0.0, + "step": 1194 + }, + { + "epoch": 0.9270752521334368, + "grad_norm": 6.612181663513184, + "learning_rate": 3.2065707220676775e-05, + "loss": 0.3219, + "step": 1195 + }, + { + "epoch": 0.9278510473235065, + "grad_norm": 0.3819209635257721, + "learning_rate": 3.204903151967116e-05, + "loss": 0.0071, + "step": 1196 + }, + { + "epoch": 0.9286268425135764, + "grad_norm": 8.089438438415527, + "learning_rate": 3.2032313464897e-05, + "loss": 0.3425, + "step": 1197 + }, + { + "epoch": 0.9294026377036463, + "grad_norm": 12.403762817382812, + "learning_rate": 3.2015553111676636e-05, + "loss": 0.8969, + "step": 1198 + }, + { + "epoch": 0.930178432893716, + "grad_norm": 4.145405292510986, + "learning_rate": 3.199875051547241e-05, + "loss": 0.1154, + "step": 1199 + }, + { + "epoch": 0.9309542280837859, + "grad_norm": 0.0, + "learning_rate": 3.1981905731886434e-05, + "loss": 0.0, + "step": 1200 + }, + { + "epoch": 0.9309542280837859, + "eval_Qnli-dev_cosine_accuracy": 0.6796875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8335632085800171, + "eval_Qnli-dev_cosine_ap": 0.706524459225737, + "eval_Qnli-dev_cosine_f1": 0.6761006289308176, + "eval_Qnli-dev_cosine_f1_threshold": 0.725448489189148, + "eval_Qnli-dev_cosine_precision": 0.5375, + "eval_Qnli-dev_cosine_recall": 0.9110169491525424, + "eval_Qnli-dev_dot_accuracy": 0.658203125, + "eval_Qnli-dev_dot_accuracy_threshold": 448.72308349609375, + "eval_Qnli-dev_dot_ap": 0.6116373918170548, + "eval_Qnli-dev_dot_f1": 0.6647058823529411, + "eval_Qnli-dev_dot_f1_threshold": 349.4371337890625, + "eval_Qnli-dev_dot_precision": 0.509009009009009, + "eval_Qnli-dev_dot_recall": 0.9576271186440678, + "eval_Qnli-dev_euclidean_accuracy": 0.6640625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.576730728149414, + "eval_Qnli-dev_euclidean_ap": 0.6998439663321232, + "eval_Qnli-dev_euclidean_f1": 0.6806451612903225, + "eval_Qnli-dev_euclidean_f1_threshold": 16.829500198364258, + "eval_Qnli-dev_euclidean_precision": 0.5494791666666666, + "eval_Qnli-dev_euclidean_recall": 0.8940677966101694, + "eval_Qnli-dev_manhattan_accuracy": 0.658203125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 226.66720581054688, + "eval_Qnli-dev_manhattan_ap": 0.7022455729310744, + "eval_Qnli-dev_manhattan_f1": 0.6774716369529984, + "eval_Qnli-dev_manhattan_f1_threshold": 303.7059631347656, + "eval_Qnli-dev_manhattan_precision": 0.5485564304461942, + "eval_Qnli-dev_manhattan_recall": 0.885593220338983, + "eval_Qnli-dev_max_accuracy": 0.6796875, + "eval_Qnli-dev_max_accuracy_threshold": 448.72308349609375, + "eval_Qnli-dev_max_ap": 0.706524459225737, + "eval_Qnli-dev_max_f1": 0.6806451612903225, + "eval_Qnli-dev_max_f1_threshold": 349.4371337890625, + "eval_Qnli-dev_max_precision": 0.5494791666666666, + "eval_Qnli-dev_max_recall": 0.9576271186440678, + "eval_allNLI-dev_cosine_accuracy": 0.71875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9080227017402649, + "eval_allNLI-dev_cosine_ap": 0.5921924463615581, + "eval_allNLI-dev_cosine_f1": 0.6183574879227053, + "eval_allNLI-dev_cosine_f1_threshold": 0.8490327596664429, + "eval_allNLI-dev_cosine_precision": 0.5311203319502075, + "eval_allNLI-dev_cosine_recall": 0.7398843930635838, + "eval_allNLI-dev_dot_accuracy": 0.6875, + "eval_allNLI-dev_dot_accuracy_threshold": 477.9401550292969, + "eval_allNLI-dev_dot_ap": 0.46842358252304855, + "eval_allNLI-dev_dot_f1": 0.5680933852140079, + "eval_allNLI-dev_dot_f1_threshold": 399.18157958984375, + "eval_allNLI-dev_dot_precision": 0.4281524926686217, + "eval_allNLI-dev_dot_recall": 0.8439306358381503, + "eval_allNLI-dev_euclidean_accuracy": 0.724609375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.70788288116455, + "eval_allNLI-dev_euclidean_ap": 0.5972942265102964, + "eval_allNLI-dev_euclidean_f1": 0.618510158013544, + "eval_allNLI-dev_euclidean_f1_threshold": 13.176319122314453, + "eval_allNLI-dev_euclidean_precision": 0.5074074074074074, + "eval_allNLI-dev_euclidean_recall": 0.791907514450867, + "eval_allNLI-dev_manhattan_accuracy": 0.734375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 208.86460876464844, + "eval_allNLI-dev_manhattan_ap": 0.59842837330715, + "eval_allNLI-dev_manhattan_f1": 0.6272727272727273, + "eval_allNLI-dev_manhattan_f1_threshold": 240.3743896484375, + "eval_allNLI-dev_manhattan_precision": 0.5168539325842697, + "eval_allNLI-dev_manhattan_recall": 0.7976878612716763, + "eval_allNLI-dev_max_accuracy": 0.734375, + "eval_allNLI-dev_max_accuracy_threshold": 477.9401550292969, + "eval_allNLI-dev_max_ap": 0.59842837330715, + "eval_allNLI-dev_max_f1": 0.6272727272727273, + "eval_allNLI-dev_max_f1_threshold": 399.18157958984375, + "eval_allNLI-dev_max_precision": 0.5311203319502075, + "eval_allNLI-dev_max_recall": 0.8439306358381503, + "eval_sequential_score": 0.706524459225737, + "eval_sts-test_pearson_cosine": 0.8170081731198993, + "eval_sts-test_pearson_dot": 0.7459592467694864, + "eval_sts-test_pearson_euclidean": 0.8541544062149957, + "eval_sts-test_pearson_manhattan": 0.8518428815379085, + "eval_sts-test_pearson_max": 0.8541544062149957, + "eval_sts-test_spearman_cosine": 0.8576533151862483, + "eval_sts-test_spearman_dot": 0.7300865329967211, + "eval_sts-test_spearman_euclidean": 0.8520380164093783, + "eval_sts-test_spearman_manhattan": 0.8519810033023657, + "eval_sts-test_spearman_max": 0.8576533151862483, + "eval_vitaminc-pairs_loss": 3.6992790699005127, + "eval_vitaminc-pairs_runtime": 5.4818, + "eval_vitaminc-pairs_samples_per_second": 23.35, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 1200 + }, + { + "epoch": 0.9309542280837859, + "eval_negation-triplets_loss": 1.7432525157928467, + "eval_negation-triplets_runtime": 1.0684, + "eval_negation-triplets_samples_per_second": 119.804, + "eval_negation-triplets_steps_per_second": 1.872, + "step": 1200 + }, + { + "epoch": 0.9309542280837859, + "eval_scitail-pairs-pos_loss": 0.16948328912258148, + "eval_scitail-pairs-pos_runtime": 1.2812, + "eval_scitail-pairs-pos_samples_per_second": 99.91, + "eval_scitail-pairs-pos_steps_per_second": 1.561, + "step": 1200 + }, + { + "epoch": 0.9309542280837859, + "eval_scitail-pairs-qa_loss": 0.0073561277240514755, + "eval_scitail-pairs-qa_runtime": 0.7771, + "eval_scitail-pairs-qa_samples_per_second": 164.722, + "eval_scitail-pairs-qa_steps_per_second": 2.574, + "step": 1200 + }, + { + "epoch": 0.9309542280837859, + "eval_xsum-pairs_loss": 0.18582597374916077, + "eval_xsum-pairs_runtime": 3.9319, + "eval_xsum-pairs_samples_per_second": 32.554, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 1200 + }, + { + "epoch": 0.9309542280837859, + "eval_sciq_pairs_loss": 0.05760980397462845, + "eval_sciq_pairs_runtime": 5.9461, + "eval_sciq_pairs_samples_per_second": 21.527, + "eval_sciq_pairs_steps_per_second": 0.336, + "step": 1200 + }, + { + "epoch": 0.9309542280837859, + "eval_qasc_pairs_loss": 0.3552781045436859, + "eval_qasc_pairs_runtime": 0.9094, + "eval_qasc_pairs_samples_per_second": 140.75, + "eval_qasc_pairs_steps_per_second": 2.199, + "step": 1200 + }, + { + "epoch": 0.9309542280837859, + "eval_openbookqa_pairs_loss": 1.2319754362106323, + "eval_openbookqa_pairs_runtime": 0.8622, + "eval_openbookqa_pairs_samples_per_second": 148.463, + "eval_openbookqa_pairs_steps_per_second": 2.32, + "step": 1200 + }, + { + "epoch": 0.9309542280837859, + "eval_msmarco_pairs_loss": 0.4920467436313629, + "eval_msmarco_pairs_runtime": 1.9674, + "eval_msmarco_pairs_samples_per_second": 65.059, + "eval_msmarco_pairs_steps_per_second": 1.017, + "step": 1200 + }, + { + "epoch": 0.9309542280837859, + "eval_nq_pairs_loss": 0.7864431738853455, + "eval_nq_pairs_runtime": 3.3299, + "eval_nq_pairs_samples_per_second": 38.439, + "eval_nq_pairs_steps_per_second": 0.601, + "step": 1200 + }, + { + "epoch": 0.9309542280837859, + "eval_trivia_pairs_loss": 0.9242772459983826, + "eval_trivia_pairs_runtime": 4.4741, + "eval_trivia_pairs_samples_per_second": 28.609, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 1200 + }, + { + "epoch": 0.9309542280837859, + "eval_gooaq_pairs_loss": 0.6514115929603577, + "eval_gooaq_pairs_runtime": 1.4637, + "eval_gooaq_pairs_samples_per_second": 87.452, + "eval_gooaq_pairs_steps_per_second": 1.366, + "step": 1200 + }, + { + "epoch": 0.9309542280837859, + "eval_paws-pos_loss": 0.03220964968204498, + "eval_paws-pos_runtime": 1.0242, + "eval_paws-pos_samples_per_second": 124.978, + "eval_paws-pos_steps_per_second": 1.953, + "step": 1200 + }, + { + "epoch": 0.9309542280837859, + "eval_global_dataset_loss": 0.4274618327617645, + "eval_global_dataset_runtime": 19.3992, + "eval_global_dataset_samples_per_second": 21.444, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1200 + }, + { + "epoch": 0.9317300232738557, + "grad_norm": 5.502035617828369, + "learning_rate": 3.196501881666043e-05, + "loss": 0.3298, + "step": 1201 + }, + { + "epoch": 0.9325058184639256, + "grad_norm": 1.5313353538513184, + "learning_rate": 3.194808982567555e-05, + "loss": 0.0379, + "step": 1202 + }, + { + "epoch": 0.9332816136539953, + "grad_norm": 14.137077331542969, + "learning_rate": 3.1931118814952156e-05, + "loss": 1.2161, + "step": 1203 + }, + { + "epoch": 0.9340574088440652, + "grad_norm": 5.789236545562744, + "learning_rate": 3.191410584064968e-05, + "loss": 0.1717, + "step": 1204 + }, + { + "epoch": 0.934833204034135, + "grad_norm": 10.070456504821777, + "learning_rate": 3.189705095906642e-05, + "loss": 0.6055, + "step": 1205 + }, + { + "epoch": 0.9356089992242048, + "grad_norm": 12.627296447753906, + "learning_rate": 3.1879954226639326e-05, + "loss": 0.8773, + "step": 1206 + }, + { + "epoch": 0.9363847944142746, + "grad_norm": 12.135488510131836, + "learning_rate": 3.1862815699943864e-05, + "loss": 1.0106, + "step": 1207 + }, + { + "epoch": 0.9371605896043445, + "grad_norm": 1.4644407033920288, + "learning_rate": 3.184563543569379e-05, + "loss": 0.0719, + "step": 1208 + }, + { + "epoch": 0.9379363847944143, + "grad_norm": 8.121184349060059, + "learning_rate": 3.182841349074099e-05, + "loss": 0.4155, + "step": 1209 + }, + { + "epoch": 0.9387121799844841, + "grad_norm": 8.184906959533691, + "learning_rate": 3.181114992207525e-05, + "loss": 0.5019, + "step": 1210 + }, + { + "epoch": 0.939487975174554, + "grad_norm": 11.244579315185547, + "learning_rate": 3.179384478682411e-05, + "loss": 1.003, + "step": 1211 + }, + { + "epoch": 0.9402637703646237, + "grad_norm": 9.746201515197754, + "learning_rate": 3.177649814225267e-05, + "loss": 0.6809, + "step": 1212 + }, + { + "epoch": 0.9410395655546936, + "grad_norm": 4.968193531036377, + "learning_rate": 3.175911004576338e-05, + "loss": 0.3111, + "step": 1213 + }, + { + "epoch": 0.9418153607447634, + "grad_norm": 8.098217010498047, + "learning_rate": 3.174168055489586e-05, + "loss": 0.6943, + "step": 1214 + }, + { + "epoch": 0.9425911559348332, + "grad_norm": 6.622547626495361, + "learning_rate": 3.17242097273267e-05, + "loss": 0.5279, + "step": 1215 + }, + { + "epoch": 0.943366951124903, + "grad_norm": 0.0, + "learning_rate": 3.17066976208693e-05, + "loss": 0.0, + "step": 1216 + }, + { + "epoch": 0.9441427463149729, + "grad_norm": 5.503591060638428, + "learning_rate": 3.168914429347365e-05, + "loss": 0.1415, + "step": 1217 + }, + { + "epoch": 0.9449185415050426, + "grad_norm": 11.31903076171875, + "learning_rate": 3.167154980322613e-05, + "loss": 0.8001, + "step": 1218 + }, + { + "epoch": 0.9456943366951125, + "grad_norm": 7.2771172523498535, + "learning_rate": 3.1653914208349346e-05, + "loss": 0.269, + "step": 1219 + }, + { + "epoch": 0.9464701318851824, + "grad_norm": 12.079676628112793, + "learning_rate": 3.163623756720194e-05, + "loss": 1.0562, + "step": 1220 + }, + { + "epoch": 0.9464701318851824, + "eval_Qnli-dev_cosine_accuracy": 0.6640625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8834493160247803, + "eval_Qnli-dev_cosine_ap": 0.7056467103762114, + "eval_Qnli-dev_cosine_f1": 0.6772655007949125, + "eval_Qnli-dev_cosine_f1_threshold": 0.7154059410095215, + "eval_Qnli-dev_cosine_precision": 0.5419847328244275, + "eval_Qnli-dev_cosine_recall": 0.902542372881356, + "eval_Qnli-dev_dot_accuracy": 0.62890625, + "eval_Qnli-dev_dot_accuracy_threshold": 370.88232421875, + "eval_Qnli-dev_dot_ap": 0.6266639711897898, + "eval_Qnli-dev_dot_f1": 0.6677471636952999, + "eval_Qnli-dev_dot_f1_threshold": 342.46575927734375, + "eval_Qnli-dev_dot_precision": 0.5406824146981627, + "eval_Qnli-dev_dot_recall": 0.8728813559322034, + "eval_Qnli-dev_euclidean_accuracy": 0.673828125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.051568031311035, + "eval_Qnli-dev_euclidean_ap": 0.7062359154691228, + "eval_Qnli-dev_euclidean_f1": 0.6763285024154589, + "eval_Qnli-dev_euclidean_f1_threshold": 16.254940032958984, + "eval_Qnli-dev_euclidean_precision": 0.5454545454545454, + "eval_Qnli-dev_euclidean_recall": 0.8898305084745762, + "eval_Qnli-dev_manhattan_accuracy": 0.658203125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 198.36468505859375, + "eval_Qnli-dev_manhattan_ap": 0.7078759667156502, + "eval_Qnli-dev_manhattan_f1": 0.6740858505564389, + "eval_Qnli-dev_manhattan_f1_threshold": 298.33349609375, + "eval_Qnli-dev_manhattan_precision": 0.539440203562341, + "eval_Qnli-dev_manhattan_recall": 0.8983050847457628, + "eval_Qnli-dev_max_accuracy": 0.673828125, + "eval_Qnli-dev_max_accuracy_threshold": 370.88232421875, + "eval_Qnli-dev_max_ap": 0.7078759667156502, + "eval_Qnli-dev_max_f1": 0.6772655007949125, + "eval_Qnli-dev_max_f1_threshold": 342.46575927734375, + "eval_Qnli-dev_max_precision": 0.5454545454545454, + "eval_Qnli-dev_max_recall": 0.902542372881356, + "eval_allNLI-dev_cosine_accuracy": 0.720703125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.855778694152832, + "eval_allNLI-dev_cosine_ap": 0.5852758190059832, + "eval_allNLI-dev_cosine_f1": 0.5979381443298969, + "eval_allNLI-dev_cosine_f1_threshold": 0.7472450733184814, + "eval_allNLI-dev_cosine_precision": 0.46474358974358976, + "eval_allNLI-dev_cosine_recall": 0.838150289017341, + "eval_allNLI-dev_dot_accuracy": 0.68359375, + "eval_allNLI-dev_dot_accuracy_threshold": 437.57275390625, + "eval_allNLI-dev_dot_ap": 0.5103040316844185, + "eval_allNLI-dev_dot_f1": 0.5785123966942148, + "eval_allNLI-dev_dot_f1_threshold": 335.12640380859375, + "eval_allNLI-dev_dot_precision": 0.45016077170418006, + "eval_allNLI-dev_dot_recall": 0.8092485549132948, + "eval_allNLI-dev_euclidean_accuracy": 0.728515625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.916200637817383, + "eval_allNLI-dev_euclidean_ap": 0.5878260405613935, + "eval_allNLI-dev_euclidean_f1": 0.6176470588235294, + "eval_allNLI-dev_euclidean_f1_threshold": 14.955665588378906, + "eval_allNLI-dev_euclidean_precision": 0.48514851485148514, + "eval_allNLI-dev_euclidean_recall": 0.8497109826589595, + "eval_allNLI-dev_manhattan_accuracy": 0.728515625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 211.16436767578125, + "eval_allNLI-dev_manhattan_ap": 0.5923754559397374, + "eval_allNLI-dev_manhattan_f1": 0.6096491228070176, + "eval_allNLI-dev_manhattan_f1_threshold": 263.7450866699219, + "eval_allNLI-dev_manhattan_precision": 0.4911660777385159, + "eval_allNLI-dev_manhattan_recall": 0.8034682080924855, + "eval_allNLI-dev_max_accuracy": 0.728515625, + "eval_allNLI-dev_max_accuracy_threshold": 437.57275390625, + "eval_allNLI-dev_max_ap": 0.5923754559397374, + "eval_allNLI-dev_max_f1": 0.6176470588235294, + "eval_allNLI-dev_max_f1_threshold": 335.12640380859375, + "eval_allNLI-dev_max_precision": 0.4911660777385159, + "eval_allNLI-dev_max_recall": 0.8497109826589595, + "eval_sequential_score": 0.7078759667156502, + "eval_sts-test_pearson_cosine": 0.8203734441669495, + "eval_sts-test_pearson_dot": 0.7687420687703166, + "eval_sts-test_pearson_euclidean": 0.8607362195418966, + "eval_sts-test_pearson_manhattan": 0.8600804421642239, + "eval_sts-test_pearson_max": 0.8607362195418966, + "eval_sts-test_spearman_cosine": 0.8622918281807284, + "eval_sts-test_spearman_dot": 0.7485517537207789, + "eval_sts-test_spearman_euclidean": 0.8593975403502396, + "eval_sts-test_spearman_manhattan": 0.8593323377623923, + "eval_sts-test_spearman_max": 0.8622918281807284, + "eval_vitaminc-pairs_loss": 3.530754804611206, + "eval_vitaminc-pairs_runtime": 5.4391, + "eval_vitaminc-pairs_samples_per_second": 23.533, + "eval_vitaminc-pairs_steps_per_second": 0.368, + "step": 1220 + }, + { + "epoch": 0.9464701318851824, + "eval_negation-triplets_loss": 1.730074167251587, + "eval_negation-triplets_runtime": 1.0689, + "eval_negation-triplets_samples_per_second": 119.75, + "eval_negation-triplets_steps_per_second": 1.871, + "step": 1220 + }, + { + "epoch": 0.9464701318851824, + "eval_scitail-pairs-pos_loss": 0.17221961915493011, + "eval_scitail-pairs-pos_runtime": 1.2631, + "eval_scitail-pairs-pos_samples_per_second": 101.338, + "eval_scitail-pairs-pos_steps_per_second": 1.583, + "step": 1220 + }, + { + "epoch": 0.9464701318851824, + "eval_scitail-pairs-qa_loss": 0.008606269024312496, + "eval_scitail-pairs-qa_runtime": 0.765, + "eval_scitail-pairs-qa_samples_per_second": 167.318, + "eval_scitail-pairs-qa_steps_per_second": 2.614, + "step": 1220 + }, + { + "epoch": 0.9464701318851824, + "eval_xsum-pairs_loss": 0.13978269696235657, + "eval_xsum-pairs_runtime": 3.9128, + "eval_xsum-pairs_samples_per_second": 32.713, + "eval_xsum-pairs_steps_per_second": 0.511, + "step": 1220 + }, + { + "epoch": 0.9464701318851824, + "eval_sciq_pairs_loss": 0.05261810123920441, + "eval_sciq_pairs_runtime": 5.9531, + "eval_sciq_pairs_samples_per_second": 21.501, + "eval_sciq_pairs_steps_per_second": 0.336, + "step": 1220 + }, + { + "epoch": 0.9464701318851824, + "eval_qasc_pairs_loss": 0.3038148283958435, + "eval_qasc_pairs_runtime": 0.9109, + "eval_qasc_pairs_samples_per_second": 140.523, + "eval_qasc_pairs_steps_per_second": 2.196, + "step": 1220 + }, + { + "epoch": 0.9464701318851824, + "eval_openbookqa_pairs_loss": 1.3340705633163452, + "eval_openbookqa_pairs_runtime": 0.8666, + "eval_openbookqa_pairs_samples_per_second": 147.709, + "eval_openbookqa_pairs_steps_per_second": 2.308, + "step": 1220 + }, + { + "epoch": 0.9464701318851824, + "eval_msmarco_pairs_loss": 0.4787042737007141, + "eval_msmarco_pairs_runtime": 1.9618, + "eval_msmarco_pairs_samples_per_second": 65.246, + "eval_msmarco_pairs_steps_per_second": 1.019, + "step": 1220 + }, + { + "epoch": 0.9464701318851824, + "eval_nq_pairs_loss": 0.7848641872406006, + "eval_nq_pairs_runtime": 3.331, + "eval_nq_pairs_samples_per_second": 38.427, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 1220 + }, + { + "epoch": 0.9464701318851824, + "eval_trivia_pairs_loss": 0.9958114624023438, + "eval_trivia_pairs_runtime": 4.4735, + "eval_trivia_pairs_samples_per_second": 28.613, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 1220 + }, + { + "epoch": 0.9464701318851824, + "eval_gooaq_pairs_loss": 0.6044450998306274, + "eval_gooaq_pairs_runtime": 1.4567, + "eval_gooaq_pairs_samples_per_second": 87.873, + "eval_gooaq_pairs_steps_per_second": 1.373, + "step": 1220 + }, + { + "epoch": 0.9464701318851824, + "eval_paws-pos_loss": 0.031799837946891785, + "eval_paws-pos_runtime": 1.0232, + "eval_paws-pos_samples_per_second": 125.103, + "eval_paws-pos_steps_per_second": 1.955, + "step": 1220 + }, + { + "epoch": 0.9464701318851824, + "eval_global_dataset_loss": 0.3775564432144165, + "eval_global_dataset_runtime": 19.3727, + "eval_global_dataset_samples_per_second": 21.474, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1220 + }, + { + "epoch": 0.9472459270752521, + "grad_norm": 9.041296005249023, + "learning_rate": 3.1618519938278346e-05, + "loss": 0.7592, + "step": 1221 + }, + { + "epoch": 0.948021722265322, + "grad_norm": 6.344674587249756, + "learning_rate": 3.1600761380208666e-05, + "loss": 0.2419, + "step": 1222 + }, + { + "epoch": 0.9487975174553918, + "grad_norm": 8.407943725585938, + "learning_rate": 3.1582961951758426e-05, + "loss": 0.4579, + "step": 1223 + }, + { + "epoch": 0.9495733126454616, + "grad_norm": 9.430070877075195, + "learning_rate": 3.156512171182841e-05, + "loss": 0.5014, + "step": 1224 + }, + { + "epoch": 0.9503491078355314, + "grad_norm": 7.2023701667785645, + "learning_rate": 3.154724071945442e-05, + "loss": 0.3161, + "step": 1225 + }, + { + "epoch": 0.9511249030256013, + "grad_norm": 8.055824279785156, + "learning_rate": 3.152931903380716e-05, + "loss": 0.5806, + "step": 1226 + }, + { + "epoch": 0.951900698215671, + "grad_norm": 10.482738494873047, + "learning_rate": 3.151135671419197e-05, + "loss": 0.7771, + "step": 1227 + }, + { + "epoch": 0.9526764934057409, + "grad_norm": 7.506356239318848, + "learning_rate": 3.1493353820048636e-05, + "loss": 0.4159, + "step": 1228 + }, + { + "epoch": 0.9534522885958107, + "grad_norm": 11.379199981689453, + "learning_rate": 3.147531041095124e-05, + "loss": 1.0459, + "step": 1229 + }, + { + "epoch": 0.9542280837858805, + "grad_norm": 5.536096096038818, + "learning_rate": 3.1457226546607924e-05, + "loss": 0.2795, + "step": 1230 + }, + { + "epoch": 0.9550038789759504, + "grad_norm": 5.343005180358887, + "learning_rate": 3.14391022868607e-05, + "loss": 0.1772, + "step": 1231 + }, + { + "epoch": 0.9557796741660202, + "grad_norm": 11.733698844909668, + "learning_rate": 3.142093769168525e-05, + "loss": 1.0023, + "step": 1232 + }, + { + "epoch": 0.95655546935609, + "grad_norm": 5.304755687713623, + "learning_rate": 3.140273282119075e-05, + "loss": 0.1929, + "step": 1233 + }, + { + "epoch": 0.9573312645461598, + "grad_norm": 9.428386688232422, + "learning_rate": 3.1384487735619616e-05, + "loss": 0.7223, + "step": 1234 + }, + { + "epoch": 0.9581070597362297, + "grad_norm": 6.395488739013672, + "learning_rate": 3.136620249534738e-05, + "loss": 0.3477, + "step": 1235 + }, + { + "epoch": 0.9588828549262994, + "grad_norm": 5.173681735992432, + "learning_rate": 3.134787716088244e-05, + "loss": 0.1797, + "step": 1236 + }, + { + "epoch": 0.9596586501163693, + "grad_norm": 7.247193336486816, + "learning_rate": 3.132951179286584e-05, + "loss": 0.6005, + "step": 1237 + }, + { + "epoch": 0.9604344453064391, + "grad_norm": 6.825124740600586, + "learning_rate": 3.131110645207116e-05, + "loss": 0.63, + "step": 1238 + }, + { + "epoch": 0.9612102404965089, + "grad_norm": 9.605498313903809, + "learning_rate": 3.1292661199404206e-05, + "loss": 0.7902, + "step": 1239 + }, + { + "epoch": 0.9619860356865787, + "grad_norm": 8.841053009033203, + "learning_rate": 3.127417609590287e-05, + "loss": 0.5181, + "step": 1240 + }, + { + "epoch": 0.9619860356865787, + "eval_Qnli-dev_cosine_accuracy": 0.6796875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8547252416610718, + "eval_Qnli-dev_cosine_ap": 0.7126178010235956, + "eval_Qnli-dev_cosine_f1": 0.6757575757575758, + "eval_Qnli-dev_cosine_f1_threshold": 0.7053539752960205, + "eval_Qnli-dev_cosine_precision": 0.5259433962264151, + "eval_Qnli-dev_cosine_recall": 0.9449152542372882, + "eval_Qnli-dev_dot_accuracy": 0.6484375, + "eval_Qnli-dev_dot_accuracy_threshold": 403.96002197265625, + "eval_Qnli-dev_dot_ap": 0.6244771158579566, + "eval_Qnli-dev_dot_f1": 0.6739811912225705, + "eval_Qnli-dev_dot_f1_threshold": 352.4027099609375, + "eval_Qnli-dev_dot_precision": 0.5348258706467661, + "eval_Qnli-dev_dot_recall": 0.9110169491525424, + "eval_Qnli-dev_euclidean_accuracy": 0.671875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.902843475341797, + "eval_Qnli-dev_euclidean_ap": 0.7120934267795361, + "eval_Qnli-dev_euclidean_f1": 0.6737804878048781, + "eval_Qnli-dev_euclidean_f1_threshold": 17.013092041015625, + "eval_Qnli-dev_euclidean_precision": 0.5261904761904762, + "eval_Qnli-dev_euclidean_recall": 0.9364406779661016, + "eval_Qnli-dev_manhattan_accuracy": 0.671875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 222.78826904296875, + "eval_Qnli-dev_manhattan_ap": 0.7107520979714589, + "eval_Qnli-dev_manhattan_f1": 0.679245283018868, + "eval_Qnli-dev_manhattan_f1_threshold": 273.74493408203125, + "eval_Qnli-dev_manhattan_precision": 0.5706051873198847, + "eval_Qnli-dev_manhattan_recall": 0.8389830508474576, + "eval_Qnli-dev_max_accuracy": 0.6796875, + "eval_Qnli-dev_max_accuracy_threshold": 403.96002197265625, + "eval_Qnli-dev_max_ap": 0.7126178010235956, + "eval_Qnli-dev_max_f1": 0.679245283018868, + "eval_Qnli-dev_max_f1_threshold": 352.4027099609375, + "eval_Qnli-dev_max_precision": 0.5706051873198847, + "eval_Qnli-dev_max_recall": 0.9449152542372882, + "eval_allNLI-dev_cosine_accuracy": 0.712890625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8636312484741211, + "eval_allNLI-dev_cosine_ap": 0.5878964507645789, + "eval_allNLI-dev_cosine_f1": 0.6011560693641619, + "eval_allNLI-dev_cosine_f1_threshold": 0.7359181046485901, + "eval_allNLI-dev_cosine_precision": 0.4508670520231214, + "eval_allNLI-dev_cosine_recall": 0.9017341040462428, + "eval_allNLI-dev_dot_accuracy": 0.6796875, + "eval_allNLI-dev_dot_accuracy_threshold": 451.7778625488281, + "eval_allNLI-dev_dot_ap": 0.5074656594607511, + "eval_allNLI-dev_dot_f1": 0.5769230769230769, + "eval_allNLI-dev_dot_f1_threshold": 342.3611755371094, + "eval_allNLI-dev_dot_precision": 0.4322766570605187, + "eval_allNLI-dev_dot_recall": 0.8670520231213873, + "eval_allNLI-dev_euclidean_accuracy": 0.71484375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.188664436340332, + "eval_allNLI-dev_euclidean_ap": 0.5923100886475902, + "eval_allNLI-dev_euclidean_f1": 0.6062992125984251, + "eval_allNLI-dev_euclidean_f1_threshold": 15.60263729095459, + "eval_allNLI-dev_euclidean_precision": 0.4597014925373134, + "eval_allNLI-dev_euclidean_recall": 0.8901734104046243, + "eval_allNLI-dev_manhattan_accuracy": 0.716796875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 215.37570190429688, + "eval_allNLI-dev_manhattan_ap": 0.5958311705114666, + "eval_allNLI-dev_manhattan_f1": 0.6086956521739131, + "eval_allNLI-dev_manhattan_f1_threshold": 261.849853515625, + "eval_allNLI-dev_manhattan_precision": 0.4878048780487805, + "eval_allNLI-dev_manhattan_recall": 0.8092485549132948, + "eval_allNLI-dev_max_accuracy": 0.716796875, + "eval_allNLI-dev_max_accuracy_threshold": 451.7778625488281, + "eval_allNLI-dev_max_ap": 0.5958311705114666, + "eval_allNLI-dev_max_f1": 0.6086956521739131, + "eval_allNLI-dev_max_f1_threshold": 342.3611755371094, + "eval_allNLI-dev_max_precision": 0.4878048780487805, + "eval_allNLI-dev_max_recall": 0.9017341040462428, + "eval_sequential_score": 0.7126178010235956, + "eval_sts-test_pearson_cosine": 0.8135591042692009, + "eval_sts-test_pearson_dot": 0.7718295344954452, + "eval_sts-test_pearson_euclidean": 0.855490219102103, + "eval_sts-test_pearson_manhattan": 0.8529488880312616, + "eval_sts-test_pearson_max": 0.855490219102103, + "eval_sts-test_spearman_cosine": 0.8581036471296298, + "eval_sts-test_spearman_dot": 0.751045562513076, + "eval_sts-test_spearman_euclidean": 0.8563246502355377, + "eval_sts-test_spearman_manhattan": 0.8549231302312817, + "eval_sts-test_spearman_max": 0.8581036471296298, + "eval_vitaminc-pairs_loss": 3.4552667140960693, + "eval_vitaminc-pairs_runtime": 5.4515, + "eval_vitaminc-pairs_samples_per_second": 23.48, + "eval_vitaminc-pairs_steps_per_second": 0.367, + "step": 1240 + }, + { + "epoch": 0.9619860356865787, + "eval_negation-triplets_loss": 1.7634042501449585, + "eval_negation-triplets_runtime": 1.0719, + "eval_negation-triplets_samples_per_second": 119.416, + "eval_negation-triplets_steps_per_second": 1.866, + "step": 1240 + }, + { + "epoch": 0.9619860356865787, + "eval_scitail-pairs-pos_loss": 0.16564783453941345, + "eval_scitail-pairs-pos_runtime": 1.2779, + "eval_scitail-pairs-pos_samples_per_second": 100.168, + "eval_scitail-pairs-pos_steps_per_second": 1.565, + "step": 1240 + }, + { + "epoch": 0.9619860356865787, + "eval_scitail-pairs-qa_loss": 0.010840480215847492, + "eval_scitail-pairs-qa_runtime": 0.7739, + "eval_scitail-pairs-qa_samples_per_second": 165.398, + "eval_scitail-pairs-qa_steps_per_second": 2.584, + "step": 1240 + }, + { + "epoch": 0.9619860356865787, + "eval_xsum-pairs_loss": 0.15841726958751678, + "eval_xsum-pairs_runtime": 3.9176, + "eval_xsum-pairs_samples_per_second": 32.673, + "eval_xsum-pairs_steps_per_second": 0.511, + "step": 1240 + }, + { + "epoch": 0.9619860356865787, + "eval_sciq_pairs_loss": 0.050957903265953064, + "eval_sciq_pairs_runtime": 5.9222, + "eval_sciq_pairs_samples_per_second": 21.614, + "eval_sciq_pairs_steps_per_second": 0.338, + "step": 1240 + }, + { + "epoch": 0.9619860356865787, + "eval_qasc_pairs_loss": 0.3406738340854645, + "eval_qasc_pairs_runtime": 0.9344, + "eval_qasc_pairs_samples_per_second": 136.988, + "eval_qasc_pairs_steps_per_second": 2.14, + "step": 1240 + }, + { + "epoch": 0.9619860356865787, + "eval_openbookqa_pairs_loss": 1.2715380191802979, + "eval_openbookqa_pairs_runtime": 0.8708, + "eval_openbookqa_pairs_samples_per_second": 146.993, + "eval_openbookqa_pairs_steps_per_second": 2.297, + "step": 1240 + }, + { + "epoch": 0.9619860356865787, + "eval_msmarco_pairs_loss": 0.4859205484390259, + "eval_msmarco_pairs_runtime": 1.9602, + "eval_msmarco_pairs_samples_per_second": 65.301, + "eval_msmarco_pairs_steps_per_second": 1.02, + "step": 1240 + }, + { + "epoch": 0.9619860356865787, + "eval_nq_pairs_loss": 0.7601814270019531, + "eval_nq_pairs_runtime": 3.3323, + "eval_nq_pairs_samples_per_second": 38.412, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 1240 + }, + { + "epoch": 0.9619860356865787, + "eval_trivia_pairs_loss": 1.0151591300964355, + "eval_trivia_pairs_runtime": 4.4756, + "eval_trivia_pairs_samples_per_second": 28.599, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 1240 + }, + { + "epoch": 0.9619860356865787, + "eval_gooaq_pairs_loss": 0.7332832217216492, + "eval_gooaq_pairs_runtime": 1.4567, + "eval_gooaq_pairs_samples_per_second": 87.87, + "eval_gooaq_pairs_steps_per_second": 1.373, + "step": 1240 + }, + { + "epoch": 0.9619860356865787, + "eval_paws-pos_loss": 0.034515175968408585, + "eval_paws-pos_runtime": 1.023, + "eval_paws-pos_samples_per_second": 125.126, + "eval_paws-pos_steps_per_second": 1.955, + "step": 1240 + }, + { + "epoch": 0.9619860356865787, + "eval_global_dataset_loss": 0.38951435685157776, + "eval_global_dataset_runtime": 19.3838, + "eval_global_dataset_samples_per_second": 21.461, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1240 + }, + { + "epoch": 0.9627618308766486, + "grad_norm": 5.5492730140686035, + "learning_rate": 3.1255651202736935e-05, + "loss": 0.1658, + "step": 1241 + }, + { + "epoch": 0.9635376260667183, + "grad_norm": 0.0, + "learning_rate": 3.1237086581207826e-05, + "loss": 0.0, + "step": 1242 + }, + { + "epoch": 0.9643134212567882, + "grad_norm": 7.026987552642822, + "learning_rate": 3.1218482292748454e-05, + "loss": 0.251, + "step": 1243 + }, + { + "epoch": 0.9650892164468581, + "grad_norm": 0.0, + "learning_rate": 3.1199838398923e-05, + "loss": 0.0, + "step": 1244 + }, + { + "epoch": 0.9658650116369278, + "grad_norm": 7.9339375495910645, + "learning_rate": 3.1181154961426664e-05, + "loss": 0.3589, + "step": 1245 + }, + { + "epoch": 0.9666408068269977, + "grad_norm": 14.850165367126465, + "learning_rate": 3.1162432042085536e-05, + "loss": 1.3156, + "step": 1246 + }, + { + "epoch": 0.9674166020170675, + "grad_norm": 2.0797557830810547, + "learning_rate": 3.1143669702856366e-05, + "loss": 0.0649, + "step": 1247 + }, + { + "epoch": 0.9681923972071373, + "grad_norm": 1.3811256885528564, + "learning_rate": 3.112486800582633e-05, + "loss": 0.0389, + "step": 1248 + }, + { + "epoch": 0.9689681923972071, + "grad_norm": 6.178694725036621, + "learning_rate": 3.110602701321284e-05, + "loss": 0.2085, + "step": 1249 + }, + { + "epoch": 0.969743987587277, + "grad_norm": 17.20705223083496, + "learning_rate": 3.108714678736335e-05, + "loss": 1.4844, + "step": 1250 + }, + { + "epoch": 0.9705197827773467, + "grad_norm": 7.540727615356445, + "learning_rate": 3.1068227390755146e-05, + "loss": 0.4047, + "step": 1251 + }, + { + "epoch": 0.9712955779674166, + "grad_norm": 8.611518859863281, + "learning_rate": 3.104926888599514e-05, + "loss": 0.5689, + "step": 1252 + }, + { + "epoch": 0.9720713731574864, + "grad_norm": 8.494779586791992, + "learning_rate": 3.103027133581964e-05, + "loss": 0.2998, + "step": 1253 + }, + { + "epoch": 0.9728471683475562, + "grad_norm": 13.081733703613281, + "learning_rate": 3.101123480309418e-05, + "loss": 1.1925, + "step": 1254 + }, + { + "epoch": 0.9736229635376261, + "grad_norm": 11.005424499511719, + "learning_rate": 3.0992159350813283e-05, + "loss": 0.7421, + "step": 1255 + }, + { + "epoch": 0.9743987587276959, + "grad_norm": 6.095644474029541, + "learning_rate": 3.0973045042100254e-05, + "loss": 0.2006, + "step": 1256 + }, + { + "epoch": 0.9751745539177658, + "grad_norm": 9.286148071289062, + "learning_rate": 3.0953891940207e-05, + "loss": 0.4527, + "step": 1257 + }, + { + "epoch": 0.9759503491078355, + "grad_norm": 6.884605884552002, + "learning_rate": 3.0934700108513784e-05, + "loss": 0.2712, + "step": 1258 + }, + { + "epoch": 0.9767261442979054, + "grad_norm": 1.2576189041137695, + "learning_rate": 3.091546961052904e-05, + "loss": 0.0259, + "step": 1259 + }, + { + "epoch": 0.9775019394879751, + "grad_norm": 12.61975383758545, + "learning_rate": 3.089620050988914e-05, + "loss": 0.8246, + "step": 1260 + }, + { + "epoch": 0.9775019394879751, + "eval_Qnli-dev_cosine_accuracy": 0.68359375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8506690263748169, + "eval_Qnli-dev_cosine_ap": 0.7089749612691468, + "eval_Qnli-dev_cosine_f1": 0.6817391304347826, + "eval_Qnli-dev_cosine_f1_threshold": 0.7615749835968018, + "eval_Qnli-dev_cosine_precision": 0.5781710914454278, + "eval_Qnli-dev_cosine_recall": 0.8305084745762712, + "eval_Qnli-dev_dot_accuracy": 0.662109375, + "eval_Qnli-dev_dot_accuracy_threshold": 402.4147644042969, + "eval_Qnli-dev_dot_ap": 0.6343324821001513, + "eval_Qnli-dev_dot_f1": 0.6656394453004623, + "eval_Qnli-dev_dot_f1_threshold": 335.5284118652344, + "eval_Qnli-dev_dot_precision": 0.5230024213075061, + "eval_Qnli-dev_dot_recall": 0.9152542372881356, + "eval_Qnli-dev_euclidean_accuracy": 0.6796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.597921371459961, + "eval_Qnli-dev_euclidean_ap": 0.710237817427888, + "eval_Qnli-dev_euclidean_f1": 0.6768707482993197, + "eval_Qnli-dev_euclidean_f1_threshold": 15.75560474395752, + "eval_Qnli-dev_euclidean_precision": 0.5653409090909091, + "eval_Qnli-dev_euclidean_recall": 0.8432203389830508, + "eval_Qnli-dev_manhattan_accuracy": 0.669921875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 246.60052490234375, + "eval_Qnli-dev_manhattan_ap": 0.7081534892697814, + "eval_Qnli-dev_manhattan_f1": 0.6762820512820512, + "eval_Qnli-dev_manhattan_f1_threshold": 302.3358459472656, + "eval_Qnli-dev_manhattan_precision": 0.5438144329896907, + "eval_Qnli-dev_manhattan_recall": 0.8940677966101694, + "eval_Qnli-dev_max_accuracy": 0.68359375, + "eval_Qnli-dev_max_accuracy_threshold": 402.4147644042969, + "eval_Qnli-dev_max_ap": 0.710237817427888, + "eval_Qnli-dev_max_f1": 0.6817391304347826, + "eval_Qnli-dev_max_f1_threshold": 335.5284118652344, + "eval_Qnli-dev_max_precision": 0.5781710914454278, + "eval_Qnli-dev_max_recall": 0.9152542372881356, + "eval_allNLI-dev_cosine_accuracy": 0.712890625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8452476263046265, + "eval_allNLI-dev_cosine_ap": 0.591388755434723, + "eval_allNLI-dev_cosine_f1": 0.6086956521739131, + "eval_allNLI-dev_cosine_f1_threshold": 0.7601401805877686, + "eval_allNLI-dev_cosine_precision": 0.4878048780487805, + "eval_allNLI-dev_cosine_recall": 0.8092485549132948, + "eval_allNLI-dev_dot_accuracy": 0.677734375, + "eval_allNLI-dev_dot_accuracy_threshold": 465.6206359863281, + "eval_allNLI-dev_dot_ap": 0.49586628198932237, + "eval_allNLI-dev_dot_f1": 0.5807962529274004, + "eval_allNLI-dev_dot_f1_threshold": 358.9529724121094, + "eval_allNLI-dev_dot_precision": 0.4881889763779528, + "eval_allNLI-dev_dot_recall": 0.7167630057803468, + "eval_allNLI-dev_euclidean_accuracy": 0.716796875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.403022766113281, + "eval_allNLI-dev_euclidean_ap": 0.594665113892039, + "eval_allNLI-dev_euclidean_f1": 0.6200873362445415, + "eval_allNLI-dev_euclidean_f1_threshold": 14.826539993286133, + "eval_allNLI-dev_euclidean_precision": 0.4982456140350877, + "eval_allNLI-dev_euclidean_recall": 0.8208092485549133, + "eval_allNLI-dev_manhattan_accuracy": 0.716796875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 192.29055786132812, + "eval_allNLI-dev_manhattan_ap": 0.5946754780353875, + "eval_allNLI-dev_manhattan_f1": 0.617283950617284, + "eval_allNLI-dev_manhattan_f1_threshold": 248.37490844726562, + "eval_allNLI-dev_manhattan_precision": 0.5387931034482759, + "eval_allNLI-dev_manhattan_recall": 0.7225433526011561, + "eval_allNLI-dev_max_accuracy": 0.716796875, + "eval_allNLI-dev_max_accuracy_threshold": 465.6206359863281, + "eval_allNLI-dev_max_ap": 0.5946754780353875, + "eval_allNLI-dev_max_f1": 0.6200873362445415, + "eval_allNLI-dev_max_f1_threshold": 358.9529724121094, + "eval_allNLI-dev_max_precision": 0.5387931034482759, + "eval_allNLI-dev_max_recall": 0.8208092485549133, + "eval_sequential_score": 0.710237817427888, + "eval_sts-test_pearson_cosine": 0.8202394215764615, + "eval_sts-test_pearson_dot": 0.7614838953893317, + "eval_sts-test_pearson_euclidean": 0.8603399427412477, + "eval_sts-test_pearson_manhattan": 0.8593513492118714, + "eval_sts-test_pearson_max": 0.8603399427412477, + "eval_sts-test_spearman_cosine": 0.8629879788029329, + "eval_sts-test_spearman_dot": 0.7405705452566683, + "eval_sts-test_spearman_euclidean": 0.8606810512782063, + "eval_sts-test_spearman_manhattan": 0.8599923293653927, + "eval_sts-test_spearman_max": 0.8629879788029329, + "eval_vitaminc-pairs_loss": 3.4737749099731445, + "eval_vitaminc-pairs_runtime": 5.4429, + "eval_vitaminc-pairs_samples_per_second": 23.517, + "eval_vitaminc-pairs_steps_per_second": 0.367, + "step": 1260 + }, + { + "epoch": 0.9775019394879751, + "eval_negation-triplets_loss": 1.866385817527771, + "eval_negation-triplets_runtime": 1.0692, + "eval_negation-triplets_samples_per_second": 119.71, + "eval_negation-triplets_steps_per_second": 1.87, + "step": 1260 + }, + { + "epoch": 0.9775019394879751, + "eval_scitail-pairs-pos_loss": 0.1383531242609024, + "eval_scitail-pairs-pos_runtime": 1.2729, + "eval_scitail-pairs-pos_samples_per_second": 100.555, + "eval_scitail-pairs-pos_steps_per_second": 1.571, + "step": 1260 + }, + { + "epoch": 0.9775019394879751, + "eval_scitail-pairs-qa_loss": 0.0053575593046844006, + "eval_scitail-pairs-qa_runtime": 0.7683, + "eval_scitail-pairs-qa_samples_per_second": 166.603, + "eval_scitail-pairs-qa_steps_per_second": 2.603, + "step": 1260 + }, + { + "epoch": 0.9775019394879751, + "eval_xsum-pairs_loss": 0.17526620626449585, + "eval_xsum-pairs_runtime": 3.9177, + "eval_xsum-pairs_samples_per_second": 32.672, + "eval_xsum-pairs_steps_per_second": 0.511, + "step": 1260 + }, + { + "epoch": 0.9775019394879751, + "eval_sciq_pairs_loss": 0.060723017901182175, + "eval_sciq_pairs_runtime": 5.9795, + "eval_sciq_pairs_samples_per_second": 21.406, + "eval_sciq_pairs_steps_per_second": 0.334, + "step": 1260 + }, + { + "epoch": 0.9775019394879751, + "eval_qasc_pairs_loss": 0.333357036113739, + "eval_qasc_pairs_runtime": 0.9373, + "eval_qasc_pairs_samples_per_second": 136.559, + "eval_qasc_pairs_steps_per_second": 2.134, + "step": 1260 + }, + { + "epoch": 0.9775019394879751, + "eval_openbookqa_pairs_loss": 1.277976632118225, + "eval_openbookqa_pairs_runtime": 0.9057, + "eval_openbookqa_pairs_samples_per_second": 141.324, + "eval_openbookqa_pairs_steps_per_second": 2.208, + "step": 1260 + }, + { + "epoch": 0.9775019394879751, + "eval_msmarco_pairs_loss": 0.44262370467185974, + "eval_msmarco_pairs_runtime": 1.9705, + "eval_msmarco_pairs_samples_per_second": 64.959, + "eval_msmarco_pairs_steps_per_second": 1.015, + "step": 1260 + }, + { + "epoch": 0.9775019394879751, + "eval_nq_pairs_loss": 0.7622057199478149, + "eval_nq_pairs_runtime": 3.3389, + "eval_nq_pairs_samples_per_second": 38.336, + "eval_nq_pairs_steps_per_second": 0.599, + "step": 1260 + }, + { + "epoch": 0.9775019394879751, + "eval_trivia_pairs_loss": 1.0362054109573364, + "eval_trivia_pairs_runtime": 4.4779, + "eval_trivia_pairs_samples_per_second": 28.585, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 1260 + }, + { + "epoch": 0.9775019394879751, + "eval_gooaq_pairs_loss": 0.6584834456443787, + "eval_gooaq_pairs_runtime": 1.4663, + "eval_gooaq_pairs_samples_per_second": 87.293, + "eval_gooaq_pairs_steps_per_second": 1.364, + "step": 1260 + }, + { + "epoch": 0.9775019394879751, + "eval_paws-pos_loss": 0.03573489561676979, + "eval_paws-pos_runtime": 1.0303, + "eval_paws-pos_samples_per_second": 124.238, + "eval_paws-pos_steps_per_second": 1.941, + "step": 1260 + }, + { + "epoch": 0.9775019394879751, + "eval_global_dataset_loss": 0.4087735116481781, + "eval_global_dataset_runtime": 19.4242, + "eval_global_dataset_samples_per_second": 21.417, + "eval_global_dataset_steps_per_second": 0.36, + "step": 1260 + }, + { + "epoch": 0.978277734678045, + "grad_norm": 11.124897003173828, + "learning_rate": 3.087689287035822e-05, + "loss": 1.0637, + "step": 1261 + }, + { + "epoch": 0.9790535298681148, + "grad_norm": 9.754111289978027, + "learning_rate": 3.085754675582792e-05, + "loss": 0.6064, + "step": 1262 + }, + { + "epoch": 0.9798293250581847, + "grad_norm": 7.557218551635742, + "learning_rate": 3.083816223031724e-05, + "loss": 0.6687, + "step": 1263 + }, + { + "epoch": 0.9806051202482544, + "grad_norm": 8.061969757080078, + "learning_rate": 3.081873935797223e-05, + "loss": 0.3731, + "step": 1264 + }, + { + "epoch": 0.9813809154383243, + "grad_norm": 8.666831016540527, + "learning_rate": 3.079927820306588e-05, + "loss": 0.5555, + "step": 1265 + }, + { + "epoch": 0.9821567106283942, + "grad_norm": 11.47012996673584, + "learning_rate": 3.077977882999785e-05, + "loss": 0.6958, + "step": 1266 + }, + { + "epoch": 0.9829325058184639, + "grad_norm": 1.2331620454788208, + "learning_rate": 3.076024130329425e-05, + "loss": 0.0339, + "step": 1267 + }, + { + "epoch": 0.9837083010085338, + "grad_norm": 9.674966812133789, + "learning_rate": 3.074066568760749e-05, + "loss": 0.6395, + "step": 1268 + }, + { + "epoch": 0.9844840961986036, + "grad_norm": 11.353792190551758, + "learning_rate": 3.072105204771597e-05, + "loss": 0.8679, + "step": 1269 + }, + { + "epoch": 0.9852598913886734, + "grad_norm": 8.691879272460938, + "learning_rate": 3.070140044852396e-05, + "loss": 0.4221, + "step": 1270 + }, + { + "epoch": 0.9860356865787432, + "grad_norm": 9.81008529663086, + "learning_rate": 3.0681710955061316e-05, + "loss": 0.5803, + "step": 1271 + }, + { + "epoch": 0.9868114817688131, + "grad_norm": 0.0, + "learning_rate": 3.0661983632483295e-05, + "loss": 0.0, + "step": 1272 + }, + { + "epoch": 0.9875872769588828, + "grad_norm": 20.598369598388672, + "learning_rate": 3.064221854607035e-05, + "loss": 1.9141, + "step": 1273 + }, + { + "epoch": 0.9883630721489527, + "grad_norm": 7.952115058898926, + "learning_rate": 3.062241576122788e-05, + "loss": 0.3395, + "step": 1274 + }, + { + "epoch": 0.9891388673390225, + "grad_norm": 6.923105239868164, + "learning_rate": 3.060257534348606e-05, + "loss": 0.3166, + "step": 1275 + }, + { + "epoch": 0.9899146625290923, + "grad_norm": 1.0444159507751465, + "learning_rate": 3.058269735849956e-05, + "loss": 0.0565, + "step": 1276 + }, + { + "epoch": 0.9906904577191621, + "grad_norm": 11.81807804107666, + "learning_rate": 3.05627818720474e-05, + "loss": 0.5926, + "step": 1277 + }, + { + "epoch": 0.991466252909232, + "grad_norm": 6.953481197357178, + "learning_rate": 3.054282895003269e-05, + "loss": 0.2625, + "step": 1278 + }, + { + "epoch": 0.9922420480993018, + "grad_norm": 0.0002872368204407394, + "learning_rate": 3.052283865848239e-05, + "loss": 0.0, + "step": 1279 + }, + { + "epoch": 0.9930178432893716, + "grad_norm": 5.5912394523620605, + "learning_rate": 3.0502811063547165e-05, + "loss": 0.2279, + "step": 1280 + }, + { + "epoch": 0.9930178432893716, + "eval_Qnli-dev_cosine_accuracy": 0.69140625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8351902961730957, + "eval_Qnli-dev_cosine_ap": 0.7092983222475606, + "eval_Qnli-dev_cosine_f1": 0.6813559322033899, + "eval_Qnli-dev_cosine_f1_threshold": 0.7565773725509644, + "eval_Qnli-dev_cosine_precision": 0.5677966101694916, + "eval_Qnli-dev_cosine_recall": 0.8516949152542372, + "eval_Qnli-dev_dot_accuracy": 0.65625, + "eval_Qnli-dev_dot_accuracy_threshold": 414.1812438964844, + "eval_Qnli-dev_dot_ap": 0.6359935253753102, + "eval_Qnli-dev_dot_f1": 0.665680473372781, + "eval_Qnli-dev_dot_f1_threshold": 320.97711181640625, + "eval_Qnli-dev_dot_precision": 0.5113636363636364, + "eval_Qnli-dev_dot_recall": 0.9533898305084746, + "eval_Qnli-dev_euclidean_accuracy": 0.6796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.622146606445312, + "eval_Qnli-dev_euclidean_ap": 0.7043692045938181, + "eval_Qnli-dev_euclidean_f1": 0.6759581881533101, + "eval_Qnli-dev_euclidean_f1_threshold": 15.105302810668945, + "eval_Qnli-dev_euclidean_precision": 0.5739644970414202, + "eval_Qnli-dev_euclidean_recall": 0.8220338983050848, + "eval_Qnli-dev_manhattan_accuracy": 0.677734375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 219.85577392578125, + "eval_Qnli-dev_manhattan_ap": 0.7085446261219764, + "eval_Qnli-dev_manhattan_f1": 0.6830985915492958, + "eval_Qnli-dev_manhattan_f1_threshold": 267.67681884765625, + "eval_Qnli-dev_manhattan_precision": 0.5843373493975904, + "eval_Qnli-dev_manhattan_recall": 0.8220338983050848, + "eval_Qnli-dev_max_accuracy": 0.69140625, + "eval_Qnli-dev_max_accuracy_threshold": 414.1812438964844, + "eval_Qnli-dev_max_ap": 0.7092983222475606, + "eval_Qnli-dev_max_f1": 0.6830985915492958, + "eval_Qnli-dev_max_f1_threshold": 320.97711181640625, + "eval_Qnli-dev_max_precision": 0.5843373493975904, + "eval_Qnli-dev_max_recall": 0.9533898305084746, + "eval_allNLI-dev_cosine_accuracy": 0.708984375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8770594596862793, + "eval_allNLI-dev_cosine_ap": 0.5917903838302063, + "eval_allNLI-dev_cosine_f1": 0.6081081081081081, + "eval_allNLI-dev_cosine_f1_threshold": 0.7945371866226196, + "eval_allNLI-dev_cosine_precision": 0.4981549815498155, + "eval_allNLI-dev_cosine_recall": 0.7803468208092486, + "eval_allNLI-dev_dot_accuracy": 0.6796875, + "eval_allNLI-dev_dot_accuracy_threshold": 468.1973571777344, + "eval_allNLI-dev_dot_ap": 0.4946300040854543, + "eval_allNLI-dev_dot_f1": 0.5679442508710801, + "eval_allNLI-dev_dot_f1_threshold": 314.8370361328125, + "eval_allNLI-dev_dot_precision": 0.40648379052369077, + "eval_allNLI-dev_dot_recall": 0.9421965317919075, + "eval_allNLI-dev_euclidean_accuracy": 0.716796875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.88510513305664, + "eval_allNLI-dev_euclidean_ap": 0.5983021647334662, + "eval_allNLI-dev_euclidean_f1": 0.6275395033860045, + "eval_allNLI-dev_euclidean_f1_threshold": 13.795380592346191, + "eval_allNLI-dev_euclidean_precision": 0.5148148148148148, + "eval_allNLI-dev_euclidean_recall": 0.8034682080924855, + "eval_allNLI-dev_manhattan_accuracy": 0.720703125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 176.0701446533203, + "eval_allNLI-dev_manhattan_ap": 0.6035950928452588, + "eval_allNLI-dev_manhattan_f1": 0.6247191011235955, + "eval_allNLI-dev_manhattan_f1_threshold": 251.88307189941406, + "eval_allNLI-dev_manhattan_precision": 0.5110294117647058, + "eval_allNLI-dev_manhattan_recall": 0.8034682080924855, + "eval_allNLI-dev_max_accuracy": 0.720703125, + "eval_allNLI-dev_max_accuracy_threshold": 468.1973571777344, + "eval_allNLI-dev_max_ap": 0.6035950928452588, + "eval_allNLI-dev_max_f1": 0.6275395033860045, + "eval_allNLI-dev_max_f1_threshold": 314.8370361328125, + "eval_allNLI-dev_max_precision": 0.5148148148148148, + "eval_allNLI-dev_max_recall": 0.9421965317919075, + "eval_sequential_score": 0.7092983222475606, + "eval_sts-test_pearson_cosine": 0.8233748715569453, + "eval_sts-test_pearson_dot": 0.7479125117682563, + "eval_sts-test_pearson_euclidean": 0.8642860364494312, + "eval_sts-test_pearson_manhattan": 0.8619461302560482, + "eval_sts-test_pearson_max": 0.8642860364494312, + "eval_sts-test_spearman_cosine": 0.8646337064406017, + "eval_sts-test_spearman_dot": 0.7338321688239199, + "eval_sts-test_spearman_euclidean": 0.8622879348209871, + "eval_sts-test_spearman_manhattan": 0.8621249954782612, + "eval_sts-test_spearman_max": 0.8646337064406017, + "eval_vitaminc-pairs_loss": 3.264768362045288, + "eval_vitaminc-pairs_runtime": 5.4771, + "eval_vitaminc-pairs_samples_per_second": 23.37, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 1280 + }, + { + "epoch": 0.9930178432893716, + "eval_negation-triplets_loss": 1.742498517036438, + "eval_negation-triplets_runtime": 1.0742, + "eval_negation-triplets_samples_per_second": 119.156, + "eval_negation-triplets_steps_per_second": 1.862, + "step": 1280 + }, + { + "epoch": 0.9930178432893716, + "eval_scitail-pairs-pos_loss": 0.10851787030696869, + "eval_scitail-pairs-pos_runtime": 1.2567, + "eval_scitail-pairs-pos_samples_per_second": 101.854, + "eval_scitail-pairs-pos_steps_per_second": 1.591, + "step": 1280 + }, + { + "epoch": 0.9930178432893716, + "eval_scitail-pairs-qa_loss": 0.008289136923849583, + "eval_scitail-pairs-qa_runtime": 0.7777, + "eval_scitail-pairs-qa_samples_per_second": 164.582, + "eval_scitail-pairs-qa_steps_per_second": 2.572, + "step": 1280 + }, + { + "epoch": 0.9930178432893716, + "eval_xsum-pairs_loss": 0.12274169921875, + "eval_xsum-pairs_runtime": 3.9132, + "eval_xsum-pairs_samples_per_second": 32.71, + "eval_xsum-pairs_steps_per_second": 0.511, + "step": 1280 + }, + { + "epoch": 0.9930178432893716, + "eval_sciq_pairs_loss": 0.05998640134930611, + "eval_sciq_pairs_runtime": 5.9667, + "eval_sciq_pairs_samples_per_second": 21.452, + "eval_sciq_pairs_steps_per_second": 0.335, + "step": 1280 + }, + { + "epoch": 0.9930178432893716, + "eval_qasc_pairs_loss": 0.3739040493965149, + "eval_qasc_pairs_runtime": 0.9121, + "eval_qasc_pairs_samples_per_second": 140.34, + "eval_qasc_pairs_steps_per_second": 2.193, + "step": 1280 + }, + { + "epoch": 0.9930178432893716, + "eval_openbookqa_pairs_loss": 1.2809137105941772, + "eval_openbookqa_pairs_runtime": 0.8682, + "eval_openbookqa_pairs_samples_per_second": 147.431, + "eval_openbookqa_pairs_steps_per_second": 2.304, + "step": 1280 + }, + { + "epoch": 0.9930178432893716, + "eval_msmarco_pairs_loss": 0.44774818420410156, + "eval_msmarco_pairs_runtime": 1.9732, + "eval_msmarco_pairs_samples_per_second": 64.868, + "eval_msmarco_pairs_steps_per_second": 1.014, + "step": 1280 + }, + { + "epoch": 0.9930178432893716, + "eval_nq_pairs_loss": 0.7762097120285034, + "eval_nq_pairs_runtime": 3.3275, + "eval_nq_pairs_samples_per_second": 38.467, + "eval_nq_pairs_steps_per_second": 0.601, + "step": 1280 + }, + { + "epoch": 0.9930178432893716, + "eval_trivia_pairs_loss": 1.040014624595642, + "eval_trivia_pairs_runtime": 4.8519, + "eval_trivia_pairs_samples_per_second": 26.382, + "eval_trivia_pairs_steps_per_second": 0.412, + "step": 1280 + }, + { + "epoch": 0.9930178432893716, + "eval_gooaq_pairs_loss": 0.66016685962677, + "eval_gooaq_pairs_runtime": 1.4713, + "eval_gooaq_pairs_samples_per_second": 87.0, + "eval_gooaq_pairs_steps_per_second": 1.359, + "step": 1280 + }, + { + "epoch": 0.9930178432893716, + "eval_paws-pos_loss": 0.03437270596623421, + "eval_paws-pos_runtime": 1.0394, + "eval_paws-pos_samples_per_second": 123.149, + "eval_paws-pos_steps_per_second": 1.924, + "step": 1280 + }, + { + "epoch": 0.9930178432893716, + "eval_global_dataset_loss": 0.42161038517951965, + "eval_global_dataset_runtime": 19.3841, + "eval_global_dataset_samples_per_second": 21.461, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1280 + }, + { + "epoch": 0.9937936384794415, + "grad_norm": 10.868171691894531, + "learning_rate": 3.0482746231501095e-05, + "loss": 0.7749, + "step": 1281 + }, + { + "epoch": 0.9945694336695112, + "grad_norm": 8.361343383789062, + "learning_rate": 3.0462644228741494e-05, + "loss": 0.4142, + "step": 1282 + }, + { + "epoch": 0.9953452288595811, + "grad_norm": 0.3114175498485565, + "learning_rate": 3.044250512178867e-05, + "loss": 0.0036, + "step": 1283 + }, + { + "epoch": 0.9961210240496509, + "grad_norm": 11.332352638244629, + "learning_rate": 3.042232897728571e-05, + "loss": 0.6992, + "step": 1284 + }, + { + "epoch": 0.9968968192397207, + "grad_norm": 15.02707290649414, + "learning_rate": 3.0402115861998284e-05, + "loss": 1.0585, + "step": 1285 + }, + { + "epoch": 0.9976726144297905, + "grad_norm": 6.417912483215332, + "learning_rate": 3.0381865842814383e-05, + "loss": 0.2649, + "step": 1286 + }, + { + "epoch": 0.9984484096198604, + "grad_norm": 9.296407699584961, + "learning_rate": 3.0361578986744128e-05, + "loss": 0.5067, + "step": 1287 + }, + { + "epoch": 0.9992242048099301, + "grad_norm": 1.1284188032150269, + "learning_rate": 3.034125536091954e-05, + "loss": 0.0064, + "step": 1288 + }, + { + "epoch": 1.0, + "grad_norm": 13.12823486328125, + "learning_rate": 3.0320895032594315e-05, + "loss": 0.8847, + "step": 1289 + }, + { + "epoch": 1.0007757951900698, + "grad_norm": 9.512046813964844, + "learning_rate": 3.0300498069143584e-05, + "loss": 0.7364, + "step": 1290 + }, + { + "epoch": 1.0015515903801397, + "grad_norm": 5.954893112182617, + "learning_rate": 3.0280064538063744e-05, + "loss": 0.342, + "step": 1291 + }, + { + "epoch": 1.0023273855702095, + "grad_norm": 9.340106964111328, + "learning_rate": 3.0259594506972175e-05, + "loss": 0.6892, + "step": 1292 + }, + { + "epoch": 1.0031031807602793, + "grad_norm": 9.20679759979248, + "learning_rate": 3.023908804360704e-05, + "loss": 0.7998, + "step": 1293 + }, + { + "epoch": 1.003878975950349, + "grad_norm": 7.529881477355957, + "learning_rate": 3.0218545215827067e-05, + "loss": 0.6517, + "step": 1294 + }, + { + "epoch": 1.004654771140419, + "grad_norm": 8.472724914550781, + "learning_rate": 3.019796609161133e-05, + "loss": 0.364, + "step": 1295 + }, + { + "epoch": 1.0054305663304888, + "grad_norm": 13.168395042419434, + "learning_rate": 3.017735073905899e-05, + "loss": 1.1281, + "step": 1296 + }, + { + "epoch": 1.0062063615205585, + "grad_norm": 10.656558990478516, + "learning_rate": 3.0156699226389114e-05, + "loss": 0.8739, + "step": 1297 + }, + { + "epoch": 1.0069821567106283, + "grad_norm": 8.201665878295898, + "learning_rate": 3.013601162194041e-05, + "loss": 0.7529, + "step": 1298 + }, + { + "epoch": 1.0077579519006983, + "grad_norm": 11.546370506286621, + "learning_rate": 3.0115287994171038e-05, + "loss": 0.9978, + "step": 1299 + }, + { + "epoch": 1.008533747090768, + "grad_norm": 8.171534538269043, + "learning_rate": 3.0094528411658346e-05, + "loss": 0.6605, + "step": 1300 + }, + { + "epoch": 1.008533747090768, + "eval_Qnli-dev_cosine_accuracy": 0.67578125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8662157654762268, + "eval_Qnli-dev_cosine_ap": 0.7126391371337122, + "eval_Qnli-dev_cosine_f1": 0.6791171477079796, + "eval_Qnli-dev_cosine_f1_threshold": 0.7792766094207764, + "eval_Qnli-dev_cosine_precision": 0.56657223796034, + "eval_Qnli-dev_cosine_recall": 0.847457627118644, + "eval_Qnli-dev_dot_accuracy": 0.64453125, + "eval_Qnli-dev_dot_accuracy_threshold": 451.7261962890625, + "eval_Qnli-dev_dot_ap": 0.6343530035260454, + "eval_Qnli-dev_dot_f1": 0.6677018633540374, + "eval_Qnli-dev_dot_f1_threshold": 373.7069396972656, + "eval_Qnli-dev_dot_precision": 0.5269607843137255, + "eval_Qnli-dev_dot_recall": 0.9110169491525424, + "eval_Qnli-dev_euclidean_accuracy": 0.681640625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.7845458984375, + "eval_Qnli-dev_euclidean_ap": 0.7125774259119511, + "eval_Qnli-dev_euclidean_f1": 0.6801346801346801, + "eval_Qnli-dev_euclidean_f1_threshold": 15.394538879394531, + "eval_Qnli-dev_euclidean_precision": 0.5642458100558659, + "eval_Qnli-dev_euclidean_recall": 0.8559322033898306, + "eval_Qnli-dev_manhattan_accuracy": 0.677734375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 217.98236083984375, + "eval_Qnli-dev_manhattan_ap": 0.7085019682563813, + "eval_Qnli-dev_manhattan_f1": 0.6767676767676768, + "eval_Qnli-dev_manhattan_f1_threshold": 276.0415954589844, + "eval_Qnli-dev_manhattan_precision": 0.5614525139664804, + "eval_Qnli-dev_manhattan_recall": 0.8516949152542372, + "eval_Qnli-dev_max_accuracy": 0.681640625, + "eval_Qnli-dev_max_accuracy_threshold": 451.7261962890625, + "eval_Qnli-dev_max_ap": 0.7126391371337122, + "eval_Qnli-dev_max_f1": 0.6801346801346801, + "eval_Qnli-dev_max_f1_threshold": 373.7069396972656, + "eval_Qnli-dev_max_precision": 0.56657223796034, + "eval_Qnli-dev_max_recall": 0.9110169491525424, + "eval_allNLI-dev_cosine_accuracy": 0.724609375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8767399191856384, + "eval_allNLI-dev_cosine_ap": 0.6004391363079808, + "eval_allNLI-dev_cosine_f1": 0.610752688172043, + "eval_allNLI-dev_cosine_f1_threshold": 0.8027051687240601, + "eval_allNLI-dev_cosine_precision": 0.4863013698630137, + "eval_allNLI-dev_cosine_recall": 0.8208092485549133, + "eval_allNLI-dev_dot_accuracy": 0.681640625, + "eval_allNLI-dev_dot_accuracy_threshold": 498.48583984375, + "eval_allNLI-dev_dot_ap": 0.5009487258992915, + "eval_allNLI-dev_dot_f1": 0.5666041275797373, + "eval_allNLI-dev_dot_f1_threshold": 358.7185363769531, + "eval_allNLI-dev_dot_precision": 0.41944444444444445, + "eval_allNLI-dev_dot_recall": 0.8728323699421965, + "eval_allNLI-dev_euclidean_accuracy": 0.72265625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.624082565307617, + "eval_allNLI-dev_euclidean_ap": 0.607039147376904, + "eval_allNLI-dev_euclidean_f1": 0.6258823529411764, + "eval_allNLI-dev_euclidean_f1_threshold": 13.012201309204102, + "eval_allNLI-dev_euclidean_precision": 0.5277777777777778, + "eval_allNLI-dev_euclidean_recall": 0.7687861271676301, + "eval_allNLI-dev_manhattan_accuracy": 0.724609375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 205.20468139648438, + "eval_allNLI-dev_manhattan_ap": 0.6086341198042668, + "eval_allNLI-dev_manhattan_f1": 0.6213151927437641, + "eval_allNLI-dev_manhattan_f1_threshold": 240.67556762695312, + "eval_allNLI-dev_manhattan_precision": 0.5111940298507462, + "eval_allNLI-dev_manhattan_recall": 0.791907514450867, + "eval_allNLI-dev_max_accuracy": 0.724609375, + "eval_allNLI-dev_max_accuracy_threshold": 498.48583984375, + "eval_allNLI-dev_max_ap": 0.6086341198042668, + "eval_allNLI-dev_max_f1": 0.6258823529411764, + "eval_allNLI-dev_max_f1_threshold": 358.7185363769531, + "eval_allNLI-dev_max_precision": 0.5277777777777778, + "eval_allNLI-dev_max_recall": 0.8728323699421965, + "eval_sequential_score": 0.7126391371337122, + "eval_sts-test_pearson_cosine": 0.8226136573877003, + "eval_sts-test_pearson_dot": 0.7311230590694432, + "eval_sts-test_pearson_euclidean": 0.8621074958093748, + "eval_sts-test_pearson_manhattan": 0.8589052568915001, + "eval_sts-test_pearson_max": 0.8621074958093748, + "eval_sts-test_spearman_cosine": 0.8617564240894416, + "eval_sts-test_spearman_dot": 0.7088352329922187, + "eval_sts-test_spearman_euclidean": 0.8597470029504449, + "eval_sts-test_spearman_manhattan": 0.8590048480315299, + "eval_sts-test_spearman_max": 0.8617564240894416, + "eval_vitaminc-pairs_loss": 3.3502156734466553, + "eval_vitaminc-pairs_runtime": 5.5935, + "eval_vitaminc-pairs_samples_per_second": 22.884, + "eval_vitaminc-pairs_steps_per_second": 0.358, + "step": 1300 + }, + { + "epoch": 1.008533747090768, + "eval_negation-triplets_loss": 1.6735045909881592, + "eval_negation-triplets_runtime": 1.1719, + "eval_negation-triplets_samples_per_second": 109.226, + "eval_negation-triplets_steps_per_second": 1.707, + "step": 1300 + }, + { + "epoch": 1.008533747090768, + "eval_scitail-pairs-pos_loss": 0.13635899126529694, + "eval_scitail-pairs-pos_runtime": 1.4389, + "eval_scitail-pairs-pos_samples_per_second": 88.954, + "eval_scitail-pairs-pos_steps_per_second": 1.39, + "step": 1300 + }, + { + "epoch": 1.008533747090768, + "eval_scitail-pairs-qa_loss": 0.015252473764121532, + "eval_scitail-pairs-qa_runtime": 0.8056, + "eval_scitail-pairs-qa_samples_per_second": 158.88, + "eval_scitail-pairs-qa_steps_per_second": 2.483, + "step": 1300 + }, + { + "epoch": 1.008533747090768, + "eval_xsum-pairs_loss": 0.13077150285243988, + "eval_xsum-pairs_runtime": 4.0345, + "eval_xsum-pairs_samples_per_second": 31.726, + "eval_xsum-pairs_steps_per_second": 0.496, + "step": 1300 + }, + { + "epoch": 1.008533747090768, + "eval_sciq_pairs_loss": 0.051347579807043076, + "eval_sciq_pairs_runtime": 6.1086, + "eval_sciq_pairs_samples_per_second": 20.954, + "eval_sciq_pairs_steps_per_second": 0.327, + "step": 1300 + }, + { + "epoch": 1.008533747090768, + "eval_qasc_pairs_loss": 0.4130658805370331, + "eval_qasc_pairs_runtime": 0.9342, + "eval_qasc_pairs_samples_per_second": 137.017, + "eval_qasc_pairs_steps_per_second": 2.141, + "step": 1300 + }, + { + "epoch": 1.008533747090768, + "eval_openbookqa_pairs_loss": 1.2348370552062988, + "eval_openbookqa_pairs_runtime": 0.8861, + "eval_openbookqa_pairs_samples_per_second": 144.454, + "eval_openbookqa_pairs_steps_per_second": 2.257, + "step": 1300 + }, + { + "epoch": 1.008533747090768, + "eval_msmarco_pairs_loss": 0.47880667448043823, + "eval_msmarco_pairs_runtime": 1.9788, + "eval_msmarco_pairs_samples_per_second": 64.686, + "eval_msmarco_pairs_steps_per_second": 1.011, + "step": 1300 + }, + { + "epoch": 1.008533747090768, + "eval_nq_pairs_loss": 0.7434379458427429, + "eval_nq_pairs_runtime": 3.3447, + "eval_nq_pairs_samples_per_second": 38.269, + "eval_nq_pairs_steps_per_second": 0.598, + "step": 1300 + }, + { + "epoch": 1.008533747090768, + "eval_trivia_pairs_loss": 0.9103936553001404, + "eval_trivia_pairs_runtime": 4.4783, + "eval_trivia_pairs_samples_per_second": 28.582, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 1300 + }, + { + "epoch": 1.008533747090768, + "eval_gooaq_pairs_loss": 0.6433140635490417, + "eval_gooaq_pairs_runtime": 1.4723, + "eval_gooaq_pairs_samples_per_second": 86.94, + "eval_gooaq_pairs_steps_per_second": 1.358, + "step": 1300 + }, + { + "epoch": 1.008533747090768, + "eval_paws-pos_loss": 0.03517117351293564, + "eval_paws-pos_runtime": 1.0305, + "eval_paws-pos_samples_per_second": 124.216, + "eval_paws-pos_steps_per_second": 1.941, + "step": 1300 + }, + { + "epoch": 1.008533747090768, + "eval_global_dataset_loss": 0.4110895097255707, + "eval_global_dataset_runtime": 19.3858, + "eval_global_dataset_samples_per_second": 21.459, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1300 + }, + { + "epoch": 1.0093095422808378, + "grad_norm": 1.6833008527755737, + "learning_rate": 3.007373294309867e-05, + "loss": 0.0452, + "step": 1301 + }, + { + "epoch": 1.0100853374709078, + "grad_norm": 1.8815747499465942, + "learning_rate": 3.0052901657307095e-05, + "loss": 0.0565, + "step": 1302 + }, + { + "epoch": 1.0108611326609775, + "grad_norm": 6.545415878295898, + "learning_rate": 3.0032034623217233e-05, + "loss": 0.2766, + "step": 1303 + }, + { + "epoch": 1.0116369278510473, + "grad_norm": 11.524443626403809, + "learning_rate": 3.0011131909880996e-05, + "loss": 0.8971, + "step": 1304 + }, + { + "epoch": 1.012412723041117, + "grad_norm": 5.498656749725342, + "learning_rate": 2.999019358646835e-05, + "loss": 0.1824, + "step": 1305 + }, + { + "epoch": 1.013188518231187, + "grad_norm": 10.798274040222168, + "learning_rate": 2.9969219722267115e-05, + "loss": 0.6842, + "step": 1306 + }, + { + "epoch": 1.0139643134212568, + "grad_norm": 13.164130210876465, + "learning_rate": 2.9948210386682704e-05, + "loss": 1.1295, + "step": 1307 + }, + { + "epoch": 1.0147401086113266, + "grad_norm": 2.0560712814331055, + "learning_rate": 2.9927165649237935e-05, + "loss": 0.0504, + "step": 1308 + }, + { + "epoch": 1.0155159038013963, + "grad_norm": 12.251069068908691, + "learning_rate": 2.9906085579572747e-05, + "loss": 0.9367, + "step": 1309 + }, + { + "epoch": 1.0162916989914663, + "grad_norm": 8.721585273742676, + "learning_rate": 2.988497024744401e-05, + "loss": 0.4695, + "step": 1310 + }, + { + "epoch": 1.017067494181536, + "grad_norm": 7.184176921844482, + "learning_rate": 2.9863819722725294e-05, + "loss": 0.3577, + "step": 1311 + }, + { + "epoch": 1.0178432893716058, + "grad_norm": 10.247795104980469, + "learning_rate": 2.9842634075406604e-05, + "loss": 0.7925, + "step": 1312 + }, + { + "epoch": 1.0186190845616758, + "grad_norm": 16.04217529296875, + "learning_rate": 2.9821413375594193e-05, + "loss": 2.2451, + "step": 1313 + }, + { + "epoch": 1.0193948797517456, + "grad_norm": 0.9659941792488098, + "learning_rate": 2.980015769351029e-05, + "loss": 0.0148, + "step": 1314 + }, + { + "epoch": 1.0201706749418153, + "grad_norm": 11.656105041503906, + "learning_rate": 2.9778867099492897e-05, + "loss": 0.8865, + "step": 1315 + }, + { + "epoch": 1.020946470131885, + "grad_norm": 7.7234296798706055, + "learning_rate": 2.9757541663995527e-05, + "loss": 0.368, + "step": 1316 + }, + { + "epoch": 1.021722265321955, + "grad_norm": 16.25791358947754, + "learning_rate": 2.9736181457587022e-05, + "loss": 2.5455, + "step": 1317 + }, + { + "epoch": 1.0224980605120249, + "grad_norm": 7.6581549644470215, + "learning_rate": 2.9714786550951243e-05, + "loss": 0.6449, + "step": 1318 + }, + { + "epoch": 1.0232738557020946, + "grad_norm": 5.596193313598633, + "learning_rate": 2.969335701488691e-05, + "loss": 0.2032, + "step": 1319 + }, + { + "epoch": 1.0240496508921644, + "grad_norm": 8.82541561126709, + "learning_rate": 2.9671892920307326e-05, + "loss": 0.7724, + "step": 1320 + }, + { + "epoch": 1.0240496508921644, + "eval_Qnli-dev_cosine_accuracy": 0.6875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8841333389282227, + "eval_Qnli-dev_cosine_ap": 0.7244910029802504, + "eval_Qnli-dev_cosine_f1": 0.6876090750436299, + "eval_Qnli-dev_cosine_f1_threshold": 0.7878941893577576, + "eval_Qnli-dev_cosine_precision": 0.5845697329376854, + "eval_Qnli-dev_cosine_recall": 0.8347457627118644, + "eval_Qnli-dev_dot_accuracy": 0.65234375, + "eval_Qnli-dev_dot_accuracy_threshold": 402.2781982421875, + "eval_Qnli-dev_dot_ap": 0.6520083000928956, + "eval_Qnli-dev_dot_f1": 0.6655683690280066, + "eval_Qnli-dev_dot_f1_threshold": 368.6608581542969, + "eval_Qnli-dev_dot_precision": 0.5444743935309974, + "eval_Qnli-dev_dot_recall": 0.8559322033898306, + "eval_Qnli-dev_euclidean_accuracy": 0.677734375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 10.566871643066406, + "eval_Qnli-dev_euclidean_ap": 0.7224269807681223, + "eval_Qnli-dev_euclidean_f1": 0.6774774774774774, + "eval_Qnli-dev_euclidean_f1_threshold": 14.181439399719238, + "eval_Qnli-dev_euclidean_precision": 0.5893416927899686, + "eval_Qnli-dev_euclidean_recall": 0.7966101694915254, + "eval_Qnli-dev_manhattan_accuracy": 0.673828125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 199.04324340820312, + "eval_Qnli-dev_manhattan_ap": 0.7241420462316805, + "eval_Qnli-dev_manhattan_f1": 0.6806883365200765, + "eval_Qnli-dev_manhattan_f1_threshold": 248.01312255859375, + "eval_Qnli-dev_manhattan_precision": 0.6202090592334495, + "eval_Qnli-dev_manhattan_recall": 0.7542372881355932, + "eval_Qnli-dev_max_accuracy": 0.6875, + "eval_Qnli-dev_max_accuracy_threshold": 402.2781982421875, + "eval_Qnli-dev_max_ap": 0.7244910029802504, + "eval_Qnli-dev_max_f1": 0.6876090750436299, + "eval_Qnli-dev_max_f1_threshold": 368.6608581542969, + "eval_Qnli-dev_max_precision": 0.6202090592334495, + "eval_Qnli-dev_max_recall": 0.8559322033898306, + "eval_allNLI-dev_cosine_accuracy": 0.7265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8809912204742432, + "eval_allNLI-dev_cosine_ap": 0.6071569884862367, + "eval_allNLI-dev_cosine_f1": 0.6153846153846154, + "eval_allNLI-dev_cosine_f1_threshold": 0.801473081111908, + "eval_allNLI-dev_cosine_precision": 0.49645390070921985, + "eval_allNLI-dev_cosine_recall": 0.8092485549132948, + "eval_allNLI-dev_dot_accuracy": 0.685546875, + "eval_allNLI-dev_dot_accuracy_threshold": 424.3419189453125, + "eval_allNLI-dev_dot_ap": 0.5084326628405879, + "eval_allNLI-dev_dot_f1": 0.5829596412556054, + "eval_allNLI-dev_dot_f1_threshold": 371.3386535644531, + "eval_allNLI-dev_dot_precision": 0.47619047619047616, + "eval_allNLI-dev_dot_recall": 0.7514450867052023, + "eval_allNLI-dev_euclidean_accuracy": 0.724609375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.782122611999512, + "eval_allNLI-dev_euclidean_ap": 0.6115654987198076, + "eval_allNLI-dev_euclidean_f1": 0.6216216216216217, + "eval_allNLI-dev_euclidean_f1_threshold": 13.280250549316406, + "eval_allNLI-dev_euclidean_precision": 0.5092250922509225, + "eval_allNLI-dev_euclidean_recall": 0.7976878612716763, + "eval_allNLI-dev_manhattan_accuracy": 0.732421875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 186.94989013671875, + "eval_allNLI-dev_manhattan_ap": 0.6170559300923076, + "eval_allNLI-dev_manhattan_f1": 0.620525059665871, + "eval_allNLI-dev_manhattan_f1_threshold": 228.82212829589844, + "eval_allNLI-dev_manhattan_precision": 0.5284552845528455, + "eval_allNLI-dev_manhattan_recall": 0.7514450867052023, + "eval_allNLI-dev_max_accuracy": 0.732421875, + "eval_allNLI-dev_max_accuracy_threshold": 424.3419189453125, + "eval_allNLI-dev_max_ap": 0.6170559300923076, + "eval_allNLI-dev_max_f1": 0.6216216216216217, + "eval_allNLI-dev_max_f1_threshold": 371.3386535644531, + "eval_allNLI-dev_max_precision": 0.5284552845528455, + "eval_allNLI-dev_max_recall": 0.8092485549132948, + "eval_sequential_score": 0.7244910029802504, + "eval_sts-test_pearson_cosine": 0.8337262323115868, + "eval_sts-test_pearson_dot": 0.7813875247031166, + "eval_sts-test_pearson_euclidean": 0.8707652210344399, + "eval_sts-test_pearson_manhattan": 0.8674722000754059, + "eval_sts-test_pearson_max": 0.8707652210344399, + "eval_sts-test_spearman_cosine": 0.8720057159833909, + "eval_sts-test_spearman_dot": 0.7693733968646623, + "eval_sts-test_spearman_euclidean": 0.8687448258195839, + "eval_sts-test_spearman_manhattan": 0.8665528195341102, + "eval_sts-test_spearman_max": 0.8720057159833909, + "eval_vitaminc-pairs_loss": 3.3064844608306885, + "eval_vitaminc-pairs_runtime": 5.4381, + "eval_vitaminc-pairs_samples_per_second": 23.537, + "eval_vitaminc-pairs_steps_per_second": 0.368, + "step": 1320 + }, + { + "epoch": 1.0240496508921644, + "eval_negation-triplets_loss": 1.8170727491378784, + "eval_negation-triplets_runtime": 1.1011, + "eval_negation-triplets_samples_per_second": 116.25, + "eval_negation-triplets_steps_per_second": 1.816, + "step": 1320 + }, + { + "epoch": 1.0240496508921644, + "eval_scitail-pairs-pos_loss": 0.10172401368618011, + "eval_scitail-pairs-pos_runtime": 1.2867, + "eval_scitail-pairs-pos_samples_per_second": 99.483, + "eval_scitail-pairs-pos_steps_per_second": 1.554, + "step": 1320 + }, + { + "epoch": 1.0240496508921644, + "eval_scitail-pairs-qa_loss": 0.005887420382350683, + "eval_scitail-pairs-qa_runtime": 0.7796, + "eval_scitail-pairs-qa_samples_per_second": 164.186, + "eval_scitail-pairs-qa_steps_per_second": 2.565, + "step": 1320 + }, + { + "epoch": 1.0240496508921644, + "eval_xsum-pairs_loss": 0.17405182123184204, + "eval_xsum-pairs_runtime": 3.9205, + "eval_xsum-pairs_samples_per_second": 32.649, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 1320 + }, + { + "epoch": 1.0240496508921644, + "eval_sciq_pairs_loss": 0.05215142294764519, + "eval_sciq_pairs_runtime": 5.9285, + "eval_sciq_pairs_samples_per_second": 21.591, + "eval_sciq_pairs_steps_per_second": 0.337, + "step": 1320 + }, + { + "epoch": 1.0240496508921644, + "eval_qasc_pairs_loss": 0.33946433663368225, + "eval_qasc_pairs_runtime": 0.9103, + "eval_qasc_pairs_samples_per_second": 140.62, + "eval_qasc_pairs_steps_per_second": 2.197, + "step": 1320 + }, + { + "epoch": 1.0240496508921644, + "eval_openbookqa_pairs_loss": 1.2653858661651611, + "eval_openbookqa_pairs_runtime": 0.8614, + "eval_openbookqa_pairs_samples_per_second": 148.592, + "eval_openbookqa_pairs_steps_per_second": 2.322, + "step": 1320 + }, + { + "epoch": 1.0240496508921644, + "eval_msmarco_pairs_loss": 0.44474682211875916, + "eval_msmarco_pairs_runtime": 1.9649, + "eval_msmarco_pairs_samples_per_second": 65.144, + "eval_msmarco_pairs_steps_per_second": 1.018, + "step": 1320 + }, + { + "epoch": 1.0240496508921644, + "eval_nq_pairs_loss": 0.7020614743232727, + "eval_nq_pairs_runtime": 3.3353, + "eval_nq_pairs_samples_per_second": 38.377, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 1320 + }, + { + "epoch": 1.0240496508921644, + "eval_trivia_pairs_loss": 0.853732168674469, + "eval_trivia_pairs_runtime": 4.473, + "eval_trivia_pairs_samples_per_second": 28.616, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 1320 + }, + { + "epoch": 1.0240496508921644, + "eval_gooaq_pairs_loss": 0.6402598023414612, + "eval_gooaq_pairs_runtime": 1.4595, + "eval_gooaq_pairs_samples_per_second": 87.701, + "eval_gooaq_pairs_steps_per_second": 1.37, + "step": 1320 + }, + { + "epoch": 1.0240496508921644, + "eval_paws-pos_loss": 0.03393023461103439, + "eval_paws-pos_runtime": 1.0278, + "eval_paws-pos_samples_per_second": 124.536, + "eval_paws-pos_steps_per_second": 1.946, + "step": 1320 + }, + { + "epoch": 1.0240496508921644, + "eval_global_dataset_loss": 0.4150521457195282, + "eval_global_dataset_runtime": 19.3739, + "eval_global_dataset_samples_per_second": 21.472, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1320 + }, + { + "epoch": 1.0248254460822344, + "grad_norm": 0.6641380786895752, + "learning_rate": 2.9650394338240158e-05, + "loss": 0.0159, + "step": 1321 + }, + { + "epoch": 1.0256012412723041, + "grad_norm": 4.955172061920166, + "learning_rate": 2.9628861339827186e-05, + "loss": 0.1812, + "step": 1322 + }, + { + "epoch": 1.0263770364623739, + "grad_norm": 5.1858391761779785, + "learning_rate": 2.9607293996324088e-05, + "loss": 0.1828, + "step": 1323 + }, + { + "epoch": 1.0271528316524439, + "grad_norm": 6.474076747894287, + "learning_rate": 2.958569237910019e-05, + "loss": 0.2537, + "step": 1324 + }, + { + "epoch": 1.0279286268425136, + "grad_norm": 6.703513145446777, + "learning_rate": 2.956405655963824e-05, + "loss": 0.359, + "step": 1325 + }, + { + "epoch": 1.0287044220325834, + "grad_norm": 17.583660125732422, + "learning_rate": 2.9542386609534154e-05, + "loss": 2.602, + "step": 1326 + }, + { + "epoch": 1.0294802172226531, + "grad_norm": 7.523731231689453, + "learning_rate": 2.952068260049681e-05, + "loss": 0.4756, + "step": 1327 + }, + { + "epoch": 1.0302560124127231, + "grad_norm": 5.792099475860596, + "learning_rate": 2.9498944604347772e-05, + "loss": 0.2169, + "step": 1328 + }, + { + "epoch": 1.031031807602793, + "grad_norm": 8.036328315734863, + "learning_rate": 2.9477172693021086e-05, + "loss": 0.6649, + "step": 1329 + }, + { + "epoch": 1.0318076027928627, + "grad_norm": 6.497225761413574, + "learning_rate": 2.9455366938563022e-05, + "loss": 0.3527, + "step": 1330 + }, + { + "epoch": 1.0325833979829324, + "grad_norm": 7.400163650512695, + "learning_rate": 2.9433527413131842e-05, + "loss": 0.4558, + "step": 1331 + }, + { + "epoch": 1.0333591931730024, + "grad_norm": 15.190314292907715, + "learning_rate": 2.9411654188997563e-05, + "loss": 2.2663, + "step": 1332 + }, + { + "epoch": 1.0341349883630722, + "grad_norm": 6.587426662445068, + "learning_rate": 2.9389747338541708e-05, + "loss": 0.405, + "step": 1333 + }, + { + "epoch": 1.034910783553142, + "grad_norm": 5.712998390197754, + "learning_rate": 2.936780693425709e-05, + "loss": 0.1744, + "step": 1334 + }, + { + "epoch": 1.0356865787432117, + "grad_norm": 8.554718971252441, + "learning_rate": 2.9345833048747542e-05, + "loss": 0.4001, + "step": 1335 + }, + { + "epoch": 1.0364623739332817, + "grad_norm": 6.324635028839111, + "learning_rate": 2.932382575472769e-05, + "loss": 0.2784, + "step": 1336 + }, + { + "epoch": 1.0372381691233514, + "grad_norm": 6.66073751449585, + "learning_rate": 2.9301785125022724e-05, + "loss": 0.5451, + "step": 1337 + }, + { + "epoch": 1.0380139643134212, + "grad_norm": 1.0588122606277466, + "learning_rate": 2.9279711232568146e-05, + "loss": 0.0327, + "step": 1338 + }, + { + "epoch": 1.0387897595034912, + "grad_norm": 7.928436756134033, + "learning_rate": 2.925760415040952e-05, + "loss": 0.4629, + "step": 1339 + }, + { + "epoch": 1.039565554693561, + "grad_norm": 6.430562496185303, + "learning_rate": 2.9235463951702242e-05, + "loss": 0.3098, + "step": 1340 + }, + { + "epoch": 1.039565554693561, + "eval_Qnli-dev_cosine_accuracy": 0.697265625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8529124855995178, + "eval_Qnli-dev_cosine_ap": 0.7330469706537892, + "eval_Qnli-dev_cosine_f1": 0.6835016835016835, + "eval_Qnli-dev_cosine_f1_threshold": 0.7754085063934326, + "eval_Qnli-dev_cosine_precision": 0.5670391061452514, + "eval_Qnli-dev_cosine_recall": 0.8601694915254238, + "eval_Qnli-dev_dot_accuracy": 0.671875, + "eval_Qnli-dev_dot_accuracy_threshold": 431.09881591796875, + "eval_Qnli-dev_dot_ap": 0.6374556353572148, + "eval_Qnli-dev_dot_f1": 0.6787878787878787, + "eval_Qnli-dev_dot_f1_threshold": 350.5918884277344, + "eval_Qnli-dev_dot_precision": 0.5283018867924528, + "eval_Qnli-dev_dot_recall": 0.9491525423728814, + "eval_Qnli-dev_euclidean_accuracy": 0.693359375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.708356857299805, + "eval_Qnli-dev_euclidean_ap": 0.7321980522389582, + "eval_Qnli-dev_euclidean_f1": 0.6830065359477124, + "eval_Qnli-dev_euclidean_f1_threshold": 15.453522682189941, + "eval_Qnli-dev_euclidean_precision": 0.5558510638297872, + "eval_Qnli-dev_euclidean_recall": 0.885593220338983, + "eval_Qnli-dev_manhattan_accuracy": 0.693359375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 215.2725067138672, + "eval_Qnli-dev_manhattan_ap": 0.7310642425935305, + "eval_Qnli-dev_manhattan_f1": 0.6840390879478827, + "eval_Qnli-dev_manhattan_f1_threshold": 281.79852294921875, + "eval_Qnli-dev_manhattan_precision": 0.5555555555555556, + "eval_Qnli-dev_manhattan_recall": 0.8898305084745762, + "eval_Qnli-dev_max_accuracy": 0.697265625, + "eval_Qnli-dev_max_accuracy_threshold": 431.09881591796875, + "eval_Qnli-dev_max_ap": 0.7330469706537892, + "eval_Qnli-dev_max_f1": 0.6840390879478827, + "eval_Qnli-dev_max_f1_threshold": 350.5918884277344, + "eval_Qnli-dev_max_precision": 0.5670391061452514, + "eval_Qnli-dev_max_recall": 0.9491525423728814, + "eval_allNLI-dev_cosine_accuracy": 0.7265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8704944849014282, + "eval_allNLI-dev_cosine_ap": 0.6002965974513637, + "eval_allNLI-dev_cosine_f1": 0.6064516129032258, + "eval_allNLI-dev_cosine_f1_threshold": 0.7752574682235718, + "eval_allNLI-dev_cosine_precision": 0.4828767123287671, + "eval_allNLI-dev_cosine_recall": 0.815028901734104, + "eval_allNLI-dev_dot_accuracy": 0.685546875, + "eval_allNLI-dev_dot_accuracy_threshold": 405.99005126953125, + "eval_allNLI-dev_dot_ap": 0.5124535821533236, + "eval_allNLI-dev_dot_f1": 0.57421875, + "eval_allNLI-dev_dot_f1_threshold": 333.994873046875, + "eval_allNLI-dev_dot_precision": 0.4336283185840708, + "eval_allNLI-dev_dot_recall": 0.8497109826589595, + "eval_allNLI-dev_euclidean_accuracy": 0.724609375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.281926155090332, + "eval_allNLI-dev_euclidean_ap": 0.6048200994502669, + "eval_allNLI-dev_euclidean_f1": 0.6258503401360545, + "eval_allNLI-dev_euclidean_f1_threshold": 13.6641206741333, + "eval_allNLI-dev_euclidean_precision": 0.5149253731343284, + "eval_allNLI-dev_euclidean_recall": 0.7976878612716763, + "eval_allNLI-dev_manhattan_accuracy": 0.73046875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 203.3751983642578, + "eval_allNLI-dev_manhattan_ap": 0.6087643052515129, + "eval_allNLI-dev_manhattan_f1": 0.6118721461187215, + "eval_allNLI-dev_manhattan_f1_threshold": 247.16043090820312, + "eval_allNLI-dev_manhattan_precision": 0.5056603773584906, + "eval_allNLI-dev_manhattan_recall": 0.7745664739884393, + "eval_allNLI-dev_max_accuracy": 0.73046875, + "eval_allNLI-dev_max_accuracy_threshold": 405.99005126953125, + "eval_allNLI-dev_max_ap": 0.6087643052515129, + "eval_allNLI-dev_max_f1": 0.6258503401360545, + "eval_allNLI-dev_max_f1_threshold": 333.994873046875, + "eval_allNLI-dev_max_precision": 0.5149253731343284, + "eval_allNLI-dev_max_recall": 0.8497109826589595, + "eval_sequential_score": 0.7330469706537892, + "eval_sts-test_pearson_cosine": 0.8271237611352065, + "eval_sts-test_pearson_dot": 0.7606416240038454, + "eval_sts-test_pearson_euclidean": 0.8675965363122529, + "eval_sts-test_pearson_manhattan": 0.8652647844385679, + "eval_sts-test_pearson_max": 0.8675965363122529, + "eval_sts-test_spearman_cosine": 0.8691351015701784, + "eval_sts-test_spearman_dot": 0.7381261866082088, + "eval_sts-test_spearman_euclidean": 0.8673084445827043, + "eval_sts-test_spearman_manhattan": 0.8657553073048544, + "eval_sts-test_spearman_max": 0.8691351015701784, + "eval_vitaminc-pairs_loss": 3.2540602684020996, + "eval_vitaminc-pairs_runtime": 5.435, + "eval_vitaminc-pairs_samples_per_second": 23.551, + "eval_vitaminc-pairs_steps_per_second": 0.368, + "step": 1340 + }, + { + "epoch": 1.039565554693561, + "eval_negation-triplets_loss": 1.8031283617019653, + "eval_negation-triplets_runtime": 1.0948, + "eval_negation-triplets_samples_per_second": 116.916, + "eval_negation-triplets_steps_per_second": 1.827, + "step": 1340 + }, + { + "epoch": 1.039565554693561, + "eval_scitail-pairs-pos_loss": 0.11065179854631424, + "eval_scitail-pairs-pos_runtime": 1.2923, + "eval_scitail-pairs-pos_samples_per_second": 99.045, + "eval_scitail-pairs-pos_steps_per_second": 1.548, + "step": 1340 + }, + { + "epoch": 1.039565554693561, + "eval_scitail-pairs-qa_loss": 0.011520191095769405, + "eval_scitail-pairs-qa_runtime": 0.781, + "eval_scitail-pairs-qa_samples_per_second": 163.897, + "eval_scitail-pairs-qa_steps_per_second": 2.561, + "step": 1340 + }, + { + "epoch": 1.039565554693561, + "eval_xsum-pairs_loss": 0.182429701089859, + "eval_xsum-pairs_runtime": 3.9084, + "eval_xsum-pairs_samples_per_second": 32.75, + "eval_xsum-pairs_steps_per_second": 0.512, + "step": 1340 + }, + { + "epoch": 1.039565554693561, + "eval_sciq_pairs_loss": 0.053086020052433014, + "eval_sciq_pairs_runtime": 5.9877, + "eval_sciq_pairs_samples_per_second": 21.377, + "eval_sciq_pairs_steps_per_second": 0.334, + "step": 1340 + }, + { + "epoch": 1.039565554693561, + "eval_qasc_pairs_loss": 0.358184814453125, + "eval_qasc_pairs_runtime": 0.9114, + "eval_qasc_pairs_samples_per_second": 140.438, + "eval_qasc_pairs_steps_per_second": 2.194, + "step": 1340 + }, + { + "epoch": 1.039565554693561, + "eval_openbookqa_pairs_loss": 1.126702070236206, + "eval_openbookqa_pairs_runtime": 0.8739, + "eval_openbookqa_pairs_samples_per_second": 146.474, + "eval_openbookqa_pairs_steps_per_second": 2.289, + "step": 1340 + }, + { + "epoch": 1.039565554693561, + "eval_msmarco_pairs_loss": 0.4774639904499054, + "eval_msmarco_pairs_runtime": 1.9683, + "eval_msmarco_pairs_samples_per_second": 65.03, + "eval_msmarco_pairs_steps_per_second": 1.016, + "step": 1340 + }, + { + "epoch": 1.039565554693561, + "eval_nq_pairs_loss": 0.7456953525543213, + "eval_nq_pairs_runtime": 3.3355, + "eval_nq_pairs_samples_per_second": 38.375, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 1340 + }, + { + "epoch": 1.039565554693561, + "eval_trivia_pairs_loss": 0.9217349886894226, + "eval_trivia_pairs_runtime": 4.4872, + "eval_trivia_pairs_samples_per_second": 28.525, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 1340 + }, + { + "epoch": 1.039565554693561, + "eval_gooaq_pairs_loss": 0.675987184047699, + "eval_gooaq_pairs_runtime": 1.4594, + "eval_gooaq_pairs_samples_per_second": 87.705, + "eval_gooaq_pairs_steps_per_second": 1.37, + "step": 1340 + }, + { + "epoch": 1.039565554693561, + "eval_paws-pos_loss": 0.033557094633579254, + "eval_paws-pos_runtime": 1.0258, + "eval_paws-pos_samples_per_second": 124.778, + "eval_paws-pos_steps_per_second": 1.95, + "step": 1340 + }, + { + "epoch": 1.039565554693561, + "eval_global_dataset_loss": 0.3858978748321533, + "eval_global_dataset_runtime": 19.4028, + "eval_global_dataset_samples_per_second": 21.44, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1340 + }, + { + "epoch": 1.0403413498836307, + "grad_norm": 9.66927433013916, + "learning_rate": 2.92132907097113e-05, + "loss": 0.8871, + "step": 1341 + }, + { + "epoch": 1.0411171450737005, + "grad_norm": 9.482085227966309, + "learning_rate": 2.919108449781102e-05, + "loss": 0.7023, + "step": 1342 + }, + { + "epoch": 1.0418929402637704, + "grad_norm": 13.155608177185059, + "learning_rate": 2.916884538948484e-05, + "loss": 2.2463, + "step": 1343 + }, + { + "epoch": 1.0426687354538402, + "grad_norm": 14.551504135131836, + "learning_rate": 2.914657345832504e-05, + "loss": 2.525, + "step": 1344 + }, + { + "epoch": 1.04344453064391, + "grad_norm": 8.199325561523438, + "learning_rate": 2.9124268778032536e-05, + "loss": 0.5623, + "step": 1345 + }, + { + "epoch": 1.0442203258339797, + "grad_norm": 6.5460028648376465, + "learning_rate": 2.91019314224166e-05, + "loss": 0.386, + "step": 1346 + }, + { + "epoch": 1.0449961210240497, + "grad_norm": 9.358715057373047, + "learning_rate": 2.9079561465394638e-05, + "loss": 0.7995, + "step": 1347 + }, + { + "epoch": 1.0457719162141195, + "grad_norm": 7.2858781814575195, + "learning_rate": 2.9057158980991945e-05, + "loss": 0.3221, + "step": 1348 + }, + { + "epoch": 1.0465477114041892, + "grad_norm": 5.400081634521484, + "learning_rate": 2.903472404334143e-05, + "loss": 0.3153, + "step": 1349 + }, + { + "epoch": 1.0473235065942592, + "grad_norm": 0.38519519567489624, + "learning_rate": 2.9012256726683416e-05, + "loss": 0.0072, + "step": 1350 + }, + { + "epoch": 1.048099301784329, + "grad_norm": 9.517110824584961, + "learning_rate": 2.8989757105365377e-05, + "loss": 0.9014, + "step": 1351 + }, + { + "epoch": 1.0488750969743987, + "grad_norm": 10.678223609924316, + "learning_rate": 2.8967225253841664e-05, + "loss": 0.7907, + "step": 1352 + }, + { + "epoch": 1.0496508921644685, + "grad_norm": 5.61782693862915, + "learning_rate": 2.8944661246673298e-05, + "loss": 0.2115, + "step": 1353 + }, + { + "epoch": 1.0504266873545385, + "grad_norm": 0.8489585518836975, + "learning_rate": 2.892206515852771e-05, + "loss": 0.0196, + "step": 1354 + }, + { + "epoch": 1.0512024825446082, + "grad_norm": 9.517561912536621, + "learning_rate": 2.8899437064178488e-05, + "loss": 0.4474, + "step": 1355 + }, + { + "epoch": 1.051978277734678, + "grad_norm": 8.063013076782227, + "learning_rate": 2.8876777038505116e-05, + "loss": 0.5143, + "step": 1356 + }, + { + "epoch": 1.0527540729247478, + "grad_norm": 12.481844902038574, + "learning_rate": 2.885408515649278e-05, + "loss": 1.0596, + "step": 1357 + }, + { + "epoch": 1.0535298681148177, + "grad_norm": 1.238405466079712, + "learning_rate": 2.8831361493232045e-05, + "loss": 0.0449, + "step": 1358 + }, + { + "epoch": 1.0543056633048875, + "grad_norm": 12.407630920410156, + "learning_rate": 2.8808606123918667e-05, + "loss": 1.0073, + "step": 1359 + }, + { + "epoch": 1.0550814584949573, + "grad_norm": 8.399218559265137, + "learning_rate": 2.8785819123853317e-05, + "loss": 0.5045, + "step": 1360 + }, + { + "epoch": 1.0550814584949573, + "eval_Qnli-dev_cosine_accuracy": 0.6875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8613344430923462, + "eval_Qnli-dev_cosine_ap": 0.7211172504821423, + "eval_Qnli-dev_cosine_f1": 0.6815286624203821, + "eval_Qnli-dev_cosine_f1_threshold": 0.726311206817627, + "eval_Qnli-dev_cosine_precision": 0.5459183673469388, + "eval_Qnli-dev_cosine_recall": 0.9067796610169492, + "eval_Qnli-dev_dot_accuracy": 0.64453125, + "eval_Qnli-dev_dot_accuracy_threshold": 436.36004638671875, + "eval_Qnli-dev_dot_ap": 0.6255653131556189, + "eval_Qnli-dev_dot_f1": 0.6784565916398714, + "eval_Qnli-dev_dot_f1_threshold": 354.810302734375, + "eval_Qnli-dev_dot_precision": 0.5466321243523317, + "eval_Qnli-dev_dot_recall": 0.8940677966101694, + "eval_Qnli-dev_euclidean_accuracy": 0.6875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.479454040527344, + "eval_Qnli-dev_euclidean_ap": 0.7259571635186417, + "eval_Qnli-dev_euclidean_f1": 0.6857142857142857, + "eval_Qnli-dev_euclidean_f1_threshold": 15.042064666748047, + "eval_Qnli-dev_euclidean_precision": 0.5925925925925926, + "eval_Qnli-dev_euclidean_recall": 0.8135593220338984, + "eval_Qnli-dev_manhattan_accuracy": 0.6796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 218.62094116210938, + "eval_Qnli-dev_manhattan_ap": 0.7227752139201886, + "eval_Qnli-dev_manhattan_f1": 0.6814814814814815, + "eval_Qnli-dev_manhattan_f1_threshold": 267.144775390625, + "eval_Qnli-dev_manhattan_precision": 0.6052631578947368, + "eval_Qnli-dev_manhattan_recall": 0.7796610169491526, + "eval_Qnli-dev_max_accuracy": 0.6875, + "eval_Qnli-dev_max_accuracy_threshold": 436.36004638671875, + "eval_Qnli-dev_max_ap": 0.7259571635186417, + "eval_Qnli-dev_max_f1": 0.6857142857142857, + "eval_Qnli-dev_max_f1_threshold": 354.810302734375, + "eval_Qnli-dev_max_precision": 0.6052631578947368, + "eval_Qnli-dev_max_recall": 0.9067796610169492, + "eval_allNLI-dev_cosine_accuracy": 0.71484375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8501720428466797, + "eval_allNLI-dev_cosine_ap": 0.5913925489338334, + "eval_allNLI-dev_cosine_f1": 0.6022471910112359, + "eval_allNLI-dev_cosine_f1_threshold": 0.7886293530464172, + "eval_allNLI-dev_cosine_precision": 0.49264705882352944, + "eval_allNLI-dev_cosine_recall": 0.7745664739884393, + "eval_allNLI-dev_dot_accuracy": 0.6796875, + "eval_allNLI-dev_dot_accuracy_threshold": 445.7347106933594, + "eval_allNLI-dev_dot_ap": 0.5166437912510755, + "eval_allNLI-dev_dot_f1": 0.5907172995780591, + "eval_allNLI-dev_dot_f1_threshold": 350.8880310058594, + "eval_allNLI-dev_dot_precision": 0.46511627906976744, + "eval_allNLI-dev_dot_recall": 0.8092485549132948, + "eval_allNLI-dev_euclidean_accuracy": 0.7109375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.20344352722168, + "eval_allNLI-dev_euclidean_ap": 0.5937562600736808, + "eval_allNLI-dev_euclidean_f1": 0.6146572104018913, + "eval_allNLI-dev_euclidean_f1_threshold": 13.609488487243652, + "eval_allNLI-dev_euclidean_precision": 0.52, + "eval_allNLI-dev_euclidean_recall": 0.7514450867052023, + "eval_allNLI-dev_manhattan_accuracy": 0.71875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 215.38082885742188, + "eval_allNLI-dev_manhattan_ap": 0.5993870759095782, + "eval_allNLI-dev_manhattan_f1": 0.6107226107226107, + "eval_allNLI-dev_manhattan_f1_threshold": 248.83123779296875, + "eval_allNLI-dev_manhattan_precision": 0.51171875, + "eval_allNLI-dev_manhattan_recall": 0.7572254335260116, + "eval_allNLI-dev_max_accuracy": 0.71875, + "eval_allNLI-dev_max_accuracy_threshold": 445.7347106933594, + "eval_allNLI-dev_max_ap": 0.5993870759095782, + "eval_allNLI-dev_max_f1": 0.6146572104018913, + "eval_allNLI-dev_max_f1_threshold": 350.8880310058594, + "eval_allNLI-dev_max_precision": 0.52, + "eval_allNLI-dev_max_recall": 0.8092485549132948, + "eval_sequential_score": 0.7259571635186417, + "eval_sts-test_pearson_cosine": 0.8272125429116522, + "eval_sts-test_pearson_dot": 0.7559063106445402, + "eval_sts-test_pearson_euclidean": 0.8695702768159681, + "eval_sts-test_pearson_manhattan": 0.868951645589432, + "eval_sts-test_pearson_max": 0.8695702768159681, + "eval_sts-test_spearman_cosine": 0.8742459641286444, + "eval_sts-test_spearman_dot": 0.7356675970586266, + "eval_sts-test_spearman_euclidean": 0.8713461271417432, + "eval_sts-test_spearman_manhattan": 0.8712254329897706, + "eval_sts-test_spearman_max": 0.8742459641286444, + "eval_vitaminc-pairs_loss": 3.570645332336426, + "eval_vitaminc-pairs_runtime": 5.4986, + "eval_vitaminc-pairs_samples_per_second": 23.278, + "eval_vitaminc-pairs_steps_per_second": 0.364, + "step": 1360 + }, + { + "epoch": 1.0550814584949573, + "eval_negation-triplets_loss": 1.795102596282959, + "eval_negation-triplets_runtime": 1.075, + "eval_negation-triplets_samples_per_second": 119.07, + "eval_negation-triplets_steps_per_second": 1.86, + "step": 1360 + }, + { + "epoch": 1.0550814584949573, + "eval_scitail-pairs-pos_loss": 0.10269102454185486, + "eval_scitail-pairs-pos_runtime": 1.2879, + "eval_scitail-pairs-pos_samples_per_second": 99.39, + "eval_scitail-pairs-pos_steps_per_second": 1.553, + "step": 1360 + }, + { + "epoch": 1.0550814584949573, + "eval_scitail-pairs-qa_loss": 0.00830863043665886, + "eval_scitail-pairs-qa_runtime": 0.7763, + "eval_scitail-pairs-qa_samples_per_second": 164.876, + "eval_scitail-pairs-qa_steps_per_second": 2.576, + "step": 1360 + }, + { + "epoch": 1.0550814584949573, + "eval_xsum-pairs_loss": 0.10404207557439804, + "eval_xsum-pairs_runtime": 3.9214, + "eval_xsum-pairs_samples_per_second": 32.642, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 1360 + }, + { + "epoch": 1.0550814584949573, + "eval_sciq_pairs_loss": 0.05425732582807541, + "eval_sciq_pairs_runtime": 6.0506, + "eval_sciq_pairs_samples_per_second": 21.155, + "eval_sciq_pairs_steps_per_second": 0.331, + "step": 1360 + }, + { + "epoch": 1.0550814584949573, + "eval_qasc_pairs_loss": 0.37198981642723083, + "eval_qasc_pairs_runtime": 0.9132, + "eval_qasc_pairs_samples_per_second": 140.168, + "eval_qasc_pairs_steps_per_second": 2.19, + "step": 1360 + }, + { + "epoch": 1.0550814584949573, + "eval_openbookqa_pairs_loss": 1.1328680515289307, + "eval_openbookqa_pairs_runtime": 0.8682, + "eval_openbookqa_pairs_samples_per_second": 147.43, + "eval_openbookqa_pairs_steps_per_second": 2.304, + "step": 1360 + }, + { + "epoch": 1.0550814584949573, + "eval_msmarco_pairs_loss": 0.4498012661933899, + "eval_msmarco_pairs_runtime": 1.9742, + "eval_msmarco_pairs_samples_per_second": 64.836, + "eval_msmarco_pairs_steps_per_second": 1.013, + "step": 1360 + }, + { + "epoch": 1.0550814584949573, + "eval_nq_pairs_loss": 0.7245843410491943, + "eval_nq_pairs_runtime": 3.3395, + "eval_nq_pairs_samples_per_second": 38.329, + "eval_nq_pairs_steps_per_second": 0.599, + "step": 1360 + }, + { + "epoch": 1.0550814584949573, + "eval_trivia_pairs_loss": 0.9737788438796997, + "eval_trivia_pairs_runtime": 4.486, + "eval_trivia_pairs_samples_per_second": 28.533, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 1360 + }, + { + "epoch": 1.0550814584949573, + "eval_gooaq_pairs_loss": 0.7342396378517151, + "eval_gooaq_pairs_runtime": 1.4732, + "eval_gooaq_pairs_samples_per_second": 86.886, + "eval_gooaq_pairs_steps_per_second": 1.358, + "step": 1360 + }, + { + "epoch": 1.0550814584949573, + "eval_paws-pos_loss": 0.03282538056373596, + "eval_paws-pos_runtime": 1.0358, + "eval_paws-pos_samples_per_second": 123.573, + "eval_paws-pos_steps_per_second": 1.931, + "step": 1360 + }, + { + "epoch": 1.0550814584949573, + "eval_global_dataset_loss": 0.45224833488464355, + "eval_global_dataset_runtime": 19.4405, + "eval_global_dataset_samples_per_second": 21.399, + "eval_global_dataset_steps_per_second": 0.36, + "step": 1360 + }, + { + "epoch": 1.0558572536850273, + "grad_norm": 7.948814392089844, + "learning_rate": 2.8763000568441326e-05, + "loss": 0.5415, + "step": 1361 + }, + { + "epoch": 1.056633048875097, + "grad_norm": 8.42543888092041, + "learning_rate": 2.874015053319247e-05, + "loss": 0.2782, + "step": 1362 + }, + { + "epoch": 1.0574088440651668, + "grad_norm": 11.478260040283203, + "learning_rate": 2.8717269093720674e-05, + "loss": 0.9338, + "step": 1363 + }, + { + "epoch": 1.0581846392552365, + "grad_norm": 9.428354263305664, + "learning_rate": 2.869435632574379e-05, + "loss": 0.67, + "step": 1364 + }, + { + "epoch": 1.0589604344453065, + "grad_norm": 10.49340534210205, + "learning_rate": 2.8671412305083345e-05, + "loss": 0.6476, + "step": 1365 + }, + { + "epoch": 1.0597362296353763, + "grad_norm": 7.920373916625977, + "learning_rate": 2.864843710766428e-05, + "loss": 0.5213, + "step": 1366 + }, + { + "epoch": 1.060512024825446, + "grad_norm": 11.716124534606934, + "learning_rate": 2.86254308095147e-05, + "loss": 0.8649, + "step": 1367 + }, + { + "epoch": 1.0612878200155158, + "grad_norm": 11.57068157196045, + "learning_rate": 2.860239348676565e-05, + "loss": 0.8868, + "step": 1368 + }, + { + "epoch": 1.0620636152055858, + "grad_norm": 6.165207386016846, + "learning_rate": 2.8579325215650815e-05, + "loss": 0.2237, + "step": 1369 + }, + { + "epoch": 1.0628394103956555, + "grad_norm": 8.181838035583496, + "learning_rate": 2.8556226072506285e-05, + "loss": 0.4523, + "step": 1370 + }, + { + "epoch": 1.0636152055857253, + "grad_norm": 11.498833656311035, + "learning_rate": 2.853309613377034e-05, + "loss": 0.6975, + "step": 1371 + }, + { + "epoch": 1.0643910007757953, + "grad_norm": 7.351899147033691, + "learning_rate": 2.8509935475983153e-05, + "loss": 0.3273, + "step": 1372 + }, + { + "epoch": 1.065166795965865, + "grad_norm": 15.589635848999023, + "learning_rate": 2.8486744175786547e-05, + "loss": 2.4117, + "step": 1373 + }, + { + "epoch": 1.0659425911559348, + "grad_norm": 0.9971689581871033, + "learning_rate": 2.8463522309923744e-05, + "loss": 0.0137, + "step": 1374 + }, + { + "epoch": 1.0667183863460046, + "grad_norm": 6.033649921417236, + "learning_rate": 2.8440269955239124e-05, + "loss": 0.2437, + "step": 1375 + }, + { + "epoch": 1.0674941815360746, + "grad_norm": 15.267115592956543, + "learning_rate": 2.8416987188677938e-05, + "loss": 2.2607, + "step": 1376 + }, + { + "epoch": 1.0682699767261443, + "grad_norm": 1.2329741716384888, + "learning_rate": 2.8393674087286096e-05, + "loss": 0.055, + "step": 1377 + }, + { + "epoch": 1.069045771916214, + "grad_norm": 13.139399528503418, + "learning_rate": 2.8370330728209868e-05, + "loss": 1.4697, + "step": 1378 + }, + { + "epoch": 1.0698215671062838, + "grad_norm": 8.643523216247559, + "learning_rate": 2.834695718869568e-05, + "loss": 0.5136, + "step": 1379 + }, + { + "epoch": 1.0705973622963538, + "grad_norm": 7.726968765258789, + "learning_rate": 2.83235535460898e-05, + "loss": 0.4496, + "step": 1380 + }, + { + "epoch": 1.0705973622963538, + "eval_Qnli-dev_cosine_accuracy": 0.69140625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8578822016716003, + "eval_Qnli-dev_cosine_ap": 0.7162556053194651, + "eval_Qnli-dev_cosine_f1": 0.687211093990755, + "eval_Qnli-dev_cosine_f1_threshold": 0.7253327369689941, + "eval_Qnli-dev_cosine_precision": 0.5399515738498789, + "eval_Qnli-dev_cosine_recall": 0.9449152542372882, + "eval_Qnli-dev_dot_accuracy": 0.6484375, + "eval_Qnli-dev_dot_accuracy_threshold": 384.8323669433594, + "eval_Qnli-dev_dot_ap": 0.614369255733588, + "eval_Qnli-dev_dot_f1": 0.675496688741722, + "eval_Qnli-dev_dot_f1_threshold": 354.275390625, + "eval_Qnli-dev_dot_precision": 0.5543478260869565, + "eval_Qnli-dev_dot_recall": 0.864406779661017, + "eval_Qnli-dev_euclidean_accuracy": 0.693359375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.722587585449219, + "eval_Qnli-dev_euclidean_ap": 0.7181419229737148, + "eval_Qnli-dev_euclidean_f1": 0.6854460093896713, + "eval_Qnli-dev_euclidean_f1_threshold": 15.871225357055664, + "eval_Qnli-dev_euclidean_precision": 0.543424317617866, + "eval_Qnli-dev_euclidean_recall": 0.9279661016949152, + "eval_Qnli-dev_manhattan_accuracy": 0.6796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 204.0523681640625, + "eval_Qnli-dev_manhattan_ap": 0.7179864080873932, + "eval_Qnli-dev_manhattan_f1": 0.6810477657935284, + "eval_Qnli-dev_manhattan_f1_threshold": 289.6634521484375, + "eval_Qnli-dev_manhattan_precision": 0.5351089588377724, + "eval_Qnli-dev_manhattan_recall": 0.9364406779661016, + "eval_Qnli-dev_max_accuracy": 0.693359375, + "eval_Qnli-dev_max_accuracy_threshold": 384.8323669433594, + "eval_Qnli-dev_max_ap": 0.7181419229737148, + "eval_Qnli-dev_max_f1": 0.687211093990755, + "eval_Qnli-dev_max_f1_threshold": 354.275390625, + "eval_Qnli-dev_max_precision": 0.5543478260869565, + "eval_Qnli-dev_max_recall": 0.9449152542372882, + "eval_allNLI-dev_cosine_accuracy": 0.708984375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.880507230758667, + "eval_allNLI-dev_cosine_ap": 0.5929644773666212, + "eval_allNLI-dev_cosine_f1": 0.5973025048169557, + "eval_allNLI-dev_cosine_f1_threshold": 0.7436983585357666, + "eval_allNLI-dev_cosine_precision": 0.4479768786127168, + "eval_allNLI-dev_cosine_recall": 0.8959537572254336, + "eval_allNLI-dev_dot_accuracy": 0.6796875, + "eval_allNLI-dev_dot_accuracy_threshold": 474.6477355957031, + "eval_allNLI-dev_dot_ap": 0.5034187802774018, + "eval_allNLI-dev_dot_f1": 0.582441113490364, + "eval_allNLI-dev_dot_f1_threshold": 360.86676025390625, + "eval_allNLI-dev_dot_precision": 0.46258503401360546, + "eval_allNLI-dev_dot_recall": 0.7861271676300579, + "eval_allNLI-dev_euclidean_accuracy": 0.7109375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.652876853942871, + "eval_allNLI-dev_euclidean_ap": 0.5943382076208767, + "eval_allNLI-dev_euclidean_f1": 0.6061946902654868, + "eval_allNLI-dev_euclidean_f1_threshold": 14.11815357208252, + "eval_allNLI-dev_euclidean_precision": 0.4910394265232975, + "eval_allNLI-dev_euclidean_recall": 0.791907514450867, + "eval_allNLI-dev_manhattan_accuracy": 0.720703125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 197.07733154296875, + "eval_allNLI-dev_manhattan_ap": 0.6008447691960824, + "eval_allNLI-dev_manhattan_f1": 0.6052631578947368, + "eval_allNLI-dev_manhattan_f1_threshold": 254.95416259765625, + "eval_allNLI-dev_manhattan_precision": 0.4876325088339223, + "eval_allNLI-dev_manhattan_recall": 0.7976878612716763, + "eval_allNLI-dev_max_accuracy": 0.720703125, + "eval_allNLI-dev_max_accuracy_threshold": 474.6477355957031, + "eval_allNLI-dev_max_ap": 0.6008447691960824, + "eval_allNLI-dev_max_f1": 0.6061946902654868, + "eval_allNLI-dev_max_f1_threshold": 360.86676025390625, + "eval_allNLI-dev_max_precision": 0.4910394265232975, + "eval_allNLI-dev_max_recall": 0.8959537572254336, + "eval_sequential_score": 0.7181419229737148, + "eval_sts-test_pearson_cosine": 0.8348867972719413, + "eval_sts-test_pearson_dot": 0.7562928859633559, + "eval_sts-test_pearson_euclidean": 0.8744535065290406, + "eval_sts-test_pearson_manhattan": 0.8748895746842977, + "eval_sts-test_pearson_max": 0.8748895746842977, + "eval_sts-test_spearman_cosine": 0.8759230176492872, + "eval_sts-test_spearman_dot": 0.7370030194498074, + "eval_sts-test_spearman_euclidean": 0.8738742154002662, + "eval_sts-test_spearman_manhattan": 0.8749486036837508, + "eval_sts-test_spearman_max": 0.8759230176492872, + "eval_vitaminc-pairs_loss": 3.586669921875, + "eval_vitaminc-pairs_runtime": 5.565, + "eval_vitaminc-pairs_samples_per_second": 23.001, + "eval_vitaminc-pairs_steps_per_second": 0.359, + "step": 1380 + }, + { + "epoch": 1.0705973622963538, + "eval_negation-triplets_loss": 1.7180016040802002, + "eval_negation-triplets_runtime": 1.1331, + "eval_negation-triplets_samples_per_second": 112.968, + "eval_negation-triplets_steps_per_second": 1.765, + "step": 1380 + }, + { + "epoch": 1.0705973622963538, + "eval_scitail-pairs-pos_loss": 0.1205751895904541, + "eval_scitail-pairs-pos_runtime": 1.4008, + "eval_scitail-pairs-pos_samples_per_second": 91.376, + "eval_scitail-pairs-pos_steps_per_second": 1.428, + "step": 1380 + }, + { + "epoch": 1.0705973622963538, + "eval_scitail-pairs-qa_loss": 0.014098682440817356, + "eval_scitail-pairs-qa_runtime": 0.7876, + "eval_scitail-pairs-qa_samples_per_second": 162.522, + "eval_scitail-pairs-qa_steps_per_second": 2.539, + "step": 1380 + }, + { + "epoch": 1.0705973622963538, + "eval_xsum-pairs_loss": 0.1235150694847107, + "eval_xsum-pairs_runtime": 3.9444, + "eval_xsum-pairs_samples_per_second": 32.451, + "eval_xsum-pairs_steps_per_second": 0.507, + "step": 1380 + }, + { + "epoch": 1.0705973622963538, + "eval_sciq_pairs_loss": 0.0668882504105568, + "eval_sciq_pairs_runtime": 6.1876, + "eval_sciq_pairs_samples_per_second": 20.687, + "eval_sciq_pairs_steps_per_second": 0.323, + "step": 1380 + }, + { + "epoch": 1.0705973622963538, + "eval_qasc_pairs_loss": 0.3115514814853668, + "eval_qasc_pairs_runtime": 0.9465, + "eval_qasc_pairs_samples_per_second": 135.231, + "eval_qasc_pairs_steps_per_second": 2.113, + "step": 1380 + }, + { + "epoch": 1.0705973622963538, + "eval_openbookqa_pairs_loss": 1.1189560890197754, + "eval_openbookqa_pairs_runtime": 0.8888, + "eval_openbookqa_pairs_samples_per_second": 144.01, + "eval_openbookqa_pairs_steps_per_second": 2.25, + "step": 1380 + }, + { + "epoch": 1.0705973622963538, + "eval_msmarco_pairs_loss": 0.429826021194458, + "eval_msmarco_pairs_runtime": 1.9917, + "eval_msmarco_pairs_samples_per_second": 64.266, + "eval_msmarco_pairs_steps_per_second": 1.004, + "step": 1380 + }, + { + "epoch": 1.0705973622963538, + "eval_nq_pairs_loss": 0.6376323699951172, + "eval_nq_pairs_runtime": 3.3502, + "eval_nq_pairs_samples_per_second": 38.207, + "eval_nq_pairs_steps_per_second": 0.597, + "step": 1380 + }, + { + "epoch": 1.0705973622963538, + "eval_trivia_pairs_loss": 0.9205784797668457, + "eval_trivia_pairs_runtime": 4.5054, + "eval_trivia_pairs_samples_per_second": 28.41, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 1380 + }, + { + "epoch": 1.0705973622963538, + "eval_gooaq_pairs_loss": 0.5904638767242432, + "eval_gooaq_pairs_runtime": 1.4922, + "eval_gooaq_pairs_samples_per_second": 85.78, + "eval_gooaq_pairs_steps_per_second": 1.34, + "step": 1380 + }, + { + "epoch": 1.0705973622963538, + "eval_paws-pos_loss": 0.03213280811905861, + "eval_paws-pos_runtime": 1.0583, + "eval_paws-pos_samples_per_second": 120.952, + "eval_paws-pos_steps_per_second": 1.89, + "step": 1380 + }, + { + "epoch": 1.0705973622963538, + "eval_global_dataset_loss": 0.3968307077884674, + "eval_global_dataset_runtime": 19.4886, + "eval_global_dataset_samples_per_second": 21.346, + "eval_global_dataset_steps_per_second": 0.359, + "step": 1380 + }, + { + "epoch": 1.0713731574864236, + "grad_norm": 8.91393756866455, + "learning_rate": 2.8300119877838135e-05, + "loss": 0.8315, + "step": 1381 + }, + { + "epoch": 1.0721489526764933, + "grad_norm": 10.955202102661133, + "learning_rate": 2.8276656261485924e-05, + "loss": 0.8345, + "step": 1382 + }, + { + "epoch": 1.0729247478665633, + "grad_norm": 10.415684700012207, + "learning_rate": 2.8253162774677545e-05, + "loss": 1.0055, + "step": 1383 + }, + { + "epoch": 1.073700543056633, + "grad_norm": 9.91122817993164, + "learning_rate": 2.822963949515618e-05, + "loss": 0.9484, + "step": 1384 + }, + { + "epoch": 1.0744763382467029, + "grad_norm": 5.9125871658325195, + "learning_rate": 2.8206086500763638e-05, + "loss": 0.269, + "step": 1385 + }, + { + "epoch": 1.0752521334367726, + "grad_norm": 5.701411247253418, + "learning_rate": 2.8182503869440023e-05, + "loss": 0.1885, + "step": 1386 + }, + { + "epoch": 1.0760279286268426, + "grad_norm": 11.101652145385742, + "learning_rate": 2.8158891679223548e-05, + "loss": 0.9251, + "step": 1387 + }, + { + "epoch": 1.0768037238169124, + "grad_norm": 8.271735191345215, + "learning_rate": 2.8135250008250204e-05, + "loss": 0.5942, + "step": 1388 + }, + { + "epoch": 1.0775795190069821, + "grad_norm": 11.120552062988281, + "learning_rate": 2.8111578934753567e-05, + "loss": 0.9002, + "step": 1389 + }, + { + "epoch": 1.0783553141970519, + "grad_norm": 5.966233253479004, + "learning_rate": 2.8087878537064495e-05, + "loss": 0.2367, + "step": 1390 + }, + { + "epoch": 1.0791311093871219, + "grad_norm": 6.305662155151367, + "learning_rate": 2.806414889361089e-05, + "loss": 0.2988, + "step": 1391 + }, + { + "epoch": 1.0799069045771916, + "grad_norm": 5.020266056060791, + "learning_rate": 2.8040390082917427e-05, + "loss": 0.1876, + "step": 1392 + }, + { + "epoch": 1.0806826997672614, + "grad_norm": 7.049361228942871, + "learning_rate": 2.8016602183605305e-05, + "loss": 0.3205, + "step": 1393 + }, + { + "epoch": 1.0814584949573312, + "grad_norm": 9.482956886291504, + "learning_rate": 2.7992785274391973e-05, + "loss": 0.6711, + "step": 1394 + }, + { + "epoch": 1.0822342901474011, + "grad_norm": 6.90962028503418, + "learning_rate": 2.796893943409089e-05, + "loss": 0.3592, + "step": 1395 + }, + { + "epoch": 1.083010085337471, + "grad_norm": 0.8566962480545044, + "learning_rate": 2.7945064741611243e-05, + "loss": 0.0587, + "step": 1396 + }, + { + "epoch": 1.0837858805275407, + "grad_norm": 9.329536437988281, + "learning_rate": 2.7921161275957693e-05, + "loss": 0.7236, + "step": 1397 + }, + { + "epoch": 1.0845616757176106, + "grad_norm": 8.454391479492188, + "learning_rate": 2.789722911623013e-05, + "loss": 0.6048, + "step": 1398 + }, + { + "epoch": 1.0853374709076804, + "grad_norm": 6.985804557800293, + "learning_rate": 2.7873268341623375e-05, + "loss": 0.2784, + "step": 1399 + }, + { + "epoch": 1.0861132660977502, + "grad_norm": 10.241326332092285, + "learning_rate": 2.7849279031426952e-05, + "loss": 0.7009, + "step": 1400 + }, + { + "epoch": 1.0861132660977502, + "eval_Qnli-dev_cosine_accuracy": 0.6875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8572052717208862, + "eval_Qnli-dev_cosine_ap": 0.7104746193631043, + "eval_Qnli-dev_cosine_f1": 0.6784, + "eval_Qnli-dev_cosine_f1_threshold": 0.7487257122993469, + "eval_Qnli-dev_cosine_precision": 0.5449871465295629, + "eval_Qnli-dev_cosine_recall": 0.8983050847457628, + "eval_Qnli-dev_dot_accuracy": 0.64453125, + "eval_Qnli-dev_dot_accuracy_threshold": 421.1455993652344, + "eval_Qnli-dev_dot_ap": 0.6170427043612466, + "eval_Qnli-dev_dot_f1": 0.6614420062695925, + "eval_Qnli-dev_dot_f1_threshold": 360.0792236328125, + "eval_Qnli-dev_dot_precision": 0.5248756218905473, + "eval_Qnli-dev_dot_recall": 0.8940677966101694, + "eval_Qnli-dev_euclidean_accuracy": 0.6796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.758161544799805, + "eval_Qnli-dev_euclidean_ap": 0.7146198055927375, + "eval_Qnli-dev_euclidean_f1": 0.678688524590164, + "eval_Qnli-dev_euclidean_f1_threshold": 15.794084548950195, + "eval_Qnli-dev_euclidean_precision": 0.553475935828877, + "eval_Qnli-dev_euclidean_recall": 0.8771186440677966, + "eval_Qnli-dev_manhattan_accuracy": 0.673828125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 212.63900756835938, + "eval_Qnli-dev_manhattan_ap": 0.7145087616460573, + "eval_Qnli-dev_manhattan_f1": 0.6748681898066785, + "eval_Qnli-dev_manhattan_f1_threshold": 263.5986328125, + "eval_Qnli-dev_manhattan_precision": 0.5765765765765766, + "eval_Qnli-dev_manhattan_recall": 0.8135593220338984, + "eval_Qnli-dev_max_accuracy": 0.6875, + "eval_Qnli-dev_max_accuracy_threshold": 421.1455993652344, + "eval_Qnli-dev_max_ap": 0.7146198055927375, + "eval_Qnli-dev_max_f1": 0.678688524590164, + "eval_Qnli-dev_max_f1_threshold": 360.0792236328125, + "eval_Qnli-dev_max_precision": 0.5765765765765766, + "eval_Qnli-dev_max_recall": 0.8983050847457628, + "eval_allNLI-dev_cosine_accuracy": 0.720703125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8823549151420593, + "eval_allNLI-dev_cosine_ap": 0.601039872886489, + "eval_allNLI-dev_cosine_f1": 0.6059225512528473, + "eval_allNLI-dev_cosine_f1_threshold": 0.8116500377655029, + "eval_allNLI-dev_cosine_precision": 0.5, + "eval_allNLI-dev_cosine_recall": 0.7687861271676301, + "eval_allNLI-dev_dot_accuracy": 0.68359375, + "eval_allNLI-dev_dot_accuracy_threshold": 431.97900390625, + "eval_allNLI-dev_dot_ap": 0.5106883268509779, + "eval_allNLI-dev_dot_f1": 0.5783664459161147, + "eval_allNLI-dev_dot_f1_threshold": 370.47662353515625, + "eval_allNLI-dev_dot_precision": 0.46785714285714286, + "eval_allNLI-dev_dot_recall": 0.7572254335260116, + "eval_allNLI-dev_euclidean_accuracy": 0.71875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.192578315734863, + "eval_allNLI-dev_euclidean_ap": 0.6058696779128714, + "eval_allNLI-dev_euclidean_f1": 0.6055776892430279, + "eval_allNLI-dev_euclidean_f1_threshold": 14.881963729858398, + "eval_allNLI-dev_euclidean_precision": 0.46200607902735563, + "eval_allNLI-dev_euclidean_recall": 0.8786127167630058, + "eval_allNLI-dev_manhattan_accuracy": 0.732421875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 199.05015563964844, + "eval_allNLI-dev_manhattan_ap": 0.608770798997006, + "eval_allNLI-dev_manhattan_f1": 0.610655737704918, + "eval_allNLI-dev_manhattan_f1_threshold": 267.8783264160156, + "eval_allNLI-dev_manhattan_precision": 0.473015873015873, + "eval_allNLI-dev_manhattan_recall": 0.861271676300578, + "eval_allNLI-dev_max_accuracy": 0.732421875, + "eval_allNLI-dev_max_accuracy_threshold": 431.97900390625, + "eval_allNLI-dev_max_ap": 0.608770798997006, + "eval_allNLI-dev_max_f1": 0.610655737704918, + "eval_allNLI-dev_max_f1_threshold": 370.47662353515625, + "eval_allNLI-dev_max_precision": 0.5, + "eval_allNLI-dev_max_recall": 0.8786127167630058, + "eval_sequential_score": 0.7146198055927375, + "eval_sts-test_pearson_cosine": 0.8307181199334542, + "eval_sts-test_pearson_dot": 0.7654185999654254, + "eval_sts-test_pearson_euclidean": 0.868453721936921, + "eval_sts-test_pearson_manhattan": 0.8671810162324795, + "eval_sts-test_pearson_max": 0.868453721936921, + "eval_sts-test_spearman_cosine": 0.8711181194649541, + "eval_sts-test_spearman_dot": 0.7521750395993336, + "eval_sts-test_spearman_euclidean": 0.8681601505895052, + "eval_sts-test_spearman_manhattan": 0.8681510213321815, + "eval_sts-test_spearman_max": 0.8711181194649541, + "eval_vitaminc-pairs_loss": 3.826028347015381, + "eval_vitaminc-pairs_runtime": 5.4466, + "eval_vitaminc-pairs_samples_per_second": 23.501, + "eval_vitaminc-pairs_steps_per_second": 0.367, + "step": 1400 + }, + { + "epoch": 1.0861132660977502, + "eval_negation-triplets_loss": 1.6951435804367065, + "eval_negation-triplets_runtime": 1.0711, + "eval_negation-triplets_samples_per_second": 119.506, + "eval_negation-triplets_steps_per_second": 1.867, + "step": 1400 + }, + { + "epoch": 1.0861132660977502, + "eval_scitail-pairs-pos_loss": 0.12009570002555847, + "eval_scitail-pairs-pos_runtime": 1.2777, + "eval_scitail-pairs-pos_samples_per_second": 100.179, + "eval_scitail-pairs-pos_steps_per_second": 1.565, + "step": 1400 + }, + { + "epoch": 1.0861132660977502, + "eval_scitail-pairs-qa_loss": 0.00959110539406538, + "eval_scitail-pairs-qa_runtime": 0.7713, + "eval_scitail-pairs-qa_samples_per_second": 165.963, + "eval_scitail-pairs-qa_steps_per_second": 2.593, + "step": 1400 + }, + { + "epoch": 1.0861132660977502, + "eval_xsum-pairs_loss": 0.10473430901765823, + "eval_xsum-pairs_runtime": 3.9216, + "eval_xsum-pairs_samples_per_second": 32.639, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 1400 + }, + { + "epoch": 1.0861132660977502, + "eval_sciq_pairs_loss": 0.05343456193804741, + "eval_sciq_pairs_runtime": 6.015, + "eval_sciq_pairs_samples_per_second": 21.28, + "eval_sciq_pairs_steps_per_second": 0.333, + "step": 1400 + }, + { + "epoch": 1.0861132660977502, + "eval_qasc_pairs_loss": 0.318198561668396, + "eval_qasc_pairs_runtime": 0.9263, + "eval_qasc_pairs_samples_per_second": 138.182, + "eval_qasc_pairs_steps_per_second": 2.159, + "step": 1400 + }, + { + "epoch": 1.0861132660977502, + "eval_openbookqa_pairs_loss": 1.0410014390945435, + "eval_openbookqa_pairs_runtime": 0.8845, + "eval_openbookqa_pairs_samples_per_second": 144.719, + "eval_openbookqa_pairs_steps_per_second": 2.261, + "step": 1400 + }, + { + "epoch": 1.0861132660977502, + "eval_msmarco_pairs_loss": 0.382579505443573, + "eval_msmarco_pairs_runtime": 1.9777, + "eval_msmarco_pairs_samples_per_second": 64.722, + "eval_msmarco_pairs_steps_per_second": 1.011, + "step": 1400 + }, + { + "epoch": 1.0861132660977502, + "eval_nq_pairs_loss": 0.6045695543289185, + "eval_nq_pairs_runtime": 3.3426, + "eval_nq_pairs_samples_per_second": 38.294, + "eval_nq_pairs_steps_per_second": 0.598, + "step": 1400 + }, + { + "epoch": 1.0861132660977502, + "eval_trivia_pairs_loss": 0.9258496165275574, + "eval_trivia_pairs_runtime": 4.4765, + "eval_trivia_pairs_samples_per_second": 28.594, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 1400 + }, + { + "epoch": 1.0861132660977502, + "eval_gooaq_pairs_loss": 0.6181907653808594, + "eval_gooaq_pairs_runtime": 1.4752, + "eval_gooaq_pairs_samples_per_second": 86.769, + "eval_gooaq_pairs_steps_per_second": 1.356, + "step": 1400 + }, + { + "epoch": 1.0861132660977502, + "eval_paws-pos_loss": 0.030857058241963387, + "eval_paws-pos_runtime": 1.0339, + "eval_paws-pos_samples_per_second": 123.809, + "eval_paws-pos_steps_per_second": 1.935, + "step": 1400 + }, + { + "epoch": 1.0861132660977502, + "eval_global_dataset_loss": 0.4166296422481537, + "eval_global_dataset_runtime": 19.4072, + "eval_global_dataset_samples_per_second": 21.435, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1400 + }, + { + "epoch": 1.08688906128782, + "grad_norm": 5.65408182144165, + "learning_rate": 2.7825261265024825e-05, + "loss": 0.2188, + "step": 1401 + }, + { + "epoch": 1.08766485647789, + "grad_norm": 8.779122352600098, + "learning_rate": 2.780121512189509e-05, + "loss": 0.6728, + "step": 1402 + }, + { + "epoch": 1.0884406516679597, + "grad_norm": 6.550997734069824, + "learning_rate": 2.7777140681609782e-05, + "loss": 0.2361, + "step": 1403 + }, + { + "epoch": 1.0892164468580294, + "grad_norm": 11.684943199157715, + "learning_rate": 2.7753038023834563e-05, + "loss": 0.9778, + "step": 1404 + }, + { + "epoch": 1.0899922420480994, + "grad_norm": 6.218662261962891, + "learning_rate": 2.7728907228328448e-05, + "loss": 0.2757, + "step": 1405 + }, + { + "epoch": 1.0907680372381692, + "grad_norm": 7.610084533691406, + "learning_rate": 2.7704748374943597e-05, + "loss": 0.2378, + "step": 1406 + }, + { + "epoch": 1.091543832428239, + "grad_norm": 0.9815082550048828, + "learning_rate": 2.7680561543625e-05, + "loss": 0.0248, + "step": 1407 + }, + { + "epoch": 1.0923196276183087, + "grad_norm": 9.517618179321289, + "learning_rate": 2.7656346814410226e-05, + "loss": 0.7233, + "step": 1408 + }, + { + "epoch": 1.0930954228083787, + "grad_norm": 5.917538642883301, + "learning_rate": 2.763210426742918e-05, + "loss": 0.2239, + "step": 1409 + }, + { + "epoch": 1.0938712179984484, + "grad_norm": 5.5288238525390625, + "learning_rate": 2.7607833982903802e-05, + "loss": 0.2196, + "step": 1410 + }, + { + "epoch": 1.0946470131885182, + "grad_norm": 0.941170334815979, + "learning_rate": 2.7583536041147827e-05, + "loss": 0.0197, + "step": 1411 + }, + { + "epoch": 1.095422808378588, + "grad_norm": 7.4091339111328125, + "learning_rate": 2.7559210522566514e-05, + "loss": 0.4134, + "step": 1412 + }, + { + "epoch": 1.096198603568658, + "grad_norm": 8.706399917602539, + "learning_rate": 2.7534857507656365e-05, + "loss": 0.5449, + "step": 1413 + }, + { + "epoch": 1.0969743987587277, + "grad_norm": 11.42735767364502, + "learning_rate": 2.751047707700489e-05, + "loss": 0.787, + "step": 1414 + }, + { + "epoch": 1.0977501939487975, + "grad_norm": 7.394610404968262, + "learning_rate": 2.7486069311290317e-05, + "loss": 0.2942, + "step": 1415 + }, + { + "epoch": 1.0985259891388672, + "grad_norm": 0.7645777463912964, + "learning_rate": 2.74616342912813e-05, + "loss": 0.0119, + "step": 1416 + }, + { + "epoch": 1.0993017843289372, + "grad_norm": 11.57833480834961, + "learning_rate": 2.743717209783673e-05, + "loss": 0.9274, + "step": 1417 + }, + { + "epoch": 1.100077579519007, + "grad_norm": 11.273788452148438, + "learning_rate": 2.7412682811905386e-05, + "loss": 0.8511, + "step": 1418 + }, + { + "epoch": 1.1008533747090767, + "grad_norm": 10.449344635009766, + "learning_rate": 2.738816651452571e-05, + "loss": 0.5329, + "step": 1419 + }, + { + "epoch": 1.1016291698991467, + "grad_norm": 6.331997394561768, + "learning_rate": 2.7363623286825515e-05, + "loss": 0.4823, + "step": 1420 + }, + { + "epoch": 1.1016291698991467, + "eval_Qnli-dev_cosine_accuracy": 0.673828125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8215996026992798, + "eval_Qnli-dev_cosine_ap": 0.7134200489767937, + "eval_Qnli-dev_cosine_f1": 0.6815068493150686, + "eval_Qnli-dev_cosine_f1_threshold": 0.738730788230896, + "eval_Qnli-dev_cosine_precision": 0.5718390804597702, + "eval_Qnli-dev_cosine_recall": 0.8432203389830508, + "eval_Qnli-dev_dot_accuracy": 0.642578125, + "eval_Qnli-dev_dot_accuracy_threshold": 422.4071044921875, + "eval_Qnli-dev_dot_ap": 0.6272854075784636, + "eval_Qnli-dev_dot_f1": 0.6666666666666665, + "eval_Qnli-dev_dot_f1_threshold": 336.7034912109375, + "eval_Qnli-dev_dot_precision": 0.5408970976253298, + "eval_Qnli-dev_dot_recall": 0.8686440677966102, + "eval_Qnli-dev_euclidean_accuracy": 0.669921875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.211593627929688, + "eval_Qnli-dev_euclidean_ap": 0.715624677536188, + "eval_Qnli-dev_euclidean_f1": 0.68630849220104, + "eval_Qnli-dev_euclidean_f1_threshold": 15.768988609313965, + "eval_Qnli-dev_euclidean_precision": 0.5806451612903226, + "eval_Qnli-dev_euclidean_recall": 0.8389830508474576, + "eval_Qnli-dev_manhattan_accuracy": 0.671875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 243.31283569335938, + "eval_Qnli-dev_manhattan_ap": 0.717001328296482, + "eval_Qnli-dev_manhattan_f1": 0.6766169154228855, + "eval_Qnli-dev_manhattan_f1_threshold": 297.87896728515625, + "eval_Qnli-dev_manhattan_precision": 0.555858310626703, + "eval_Qnli-dev_manhattan_recall": 0.864406779661017, + "eval_Qnli-dev_max_accuracy": 0.673828125, + "eval_Qnli-dev_max_accuracy_threshold": 422.4071044921875, + "eval_Qnli-dev_max_ap": 0.717001328296482, + "eval_Qnli-dev_max_f1": 0.68630849220104, + "eval_Qnli-dev_max_f1_threshold": 336.7034912109375, + "eval_Qnli-dev_max_precision": 0.5806451612903226, + "eval_Qnli-dev_max_recall": 0.8686440677966102, + "eval_allNLI-dev_cosine_accuracy": 0.728515625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8661868572235107, + "eval_allNLI-dev_cosine_ap": 0.5983464129460488, + "eval_allNLI-dev_cosine_f1": 0.6079295154185022, + "eval_allNLI-dev_cosine_f1_threshold": 0.7829838991165161, + "eval_allNLI-dev_cosine_precision": 0.49110320284697506, + "eval_allNLI-dev_cosine_recall": 0.7976878612716763, + "eval_allNLI-dev_dot_accuracy": 0.681640625, + "eval_allNLI-dev_dot_accuracy_threshold": 397.853759765625, + "eval_allNLI-dev_dot_ap": 0.5078987601358806, + "eval_allNLI-dev_dot_f1": 0.5784114052953157, + "eval_allNLI-dev_dot_f1_threshold": 330.011962890625, + "eval_allNLI-dev_dot_precision": 0.44654088050314467, + "eval_allNLI-dev_dot_recall": 0.8208092485549133, + "eval_allNLI-dev_euclidean_accuracy": 0.73046875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.946388244628906, + "eval_allNLI-dev_euclidean_ap": 0.60366313556962, + "eval_allNLI-dev_euclidean_f1": 0.6153846153846154, + "eval_allNLI-dev_euclidean_f1_threshold": 12.059853553771973, + "eval_allNLI-dev_euclidean_precision": 0.5863874345549738, + "eval_allNLI-dev_euclidean_recall": 0.6473988439306358, + "eval_allNLI-dev_manhattan_accuracy": 0.73828125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 204.42019653320312, + "eval_allNLI-dev_manhattan_ap": 0.6067915176037065, + "eval_allNLI-dev_manhattan_f1": 0.6140724946695097, + "eval_allNLI-dev_manhattan_f1_threshold": 262.2166748046875, + "eval_allNLI-dev_manhattan_precision": 0.4864864864864865, + "eval_allNLI-dev_manhattan_recall": 0.8323699421965318, + "eval_allNLI-dev_max_accuracy": 0.73828125, + "eval_allNLI-dev_max_accuracy_threshold": 397.853759765625, + "eval_allNLI-dev_max_ap": 0.6067915176037065, + "eval_allNLI-dev_max_f1": 0.6153846153846154, + "eval_allNLI-dev_max_f1_threshold": 330.011962890625, + "eval_allNLI-dev_max_precision": 0.5863874345549738, + "eval_allNLI-dev_max_recall": 0.8323699421965318, + "eval_sequential_score": 0.717001328296482, + "eval_sts-test_pearson_cosine": 0.830052147827965, + "eval_sts-test_pearson_dot": 0.7858756759883921, + "eval_sts-test_pearson_euclidean": 0.8660243031877457, + "eval_sts-test_pearson_manhattan": 0.8640129261644363, + "eval_sts-test_pearson_max": 0.8660243031877457, + "eval_sts-test_spearman_cosine": 0.8708098280252287, + "eval_sts-test_spearman_dot": 0.776653174057724, + "eval_sts-test_spearman_euclidean": 0.8679540262795364, + "eval_sts-test_spearman_manhattan": 0.8668826810818263, + "eval_sts-test_spearman_max": 0.8708098280252287, + "eval_vitaminc-pairs_loss": 3.771179676055908, + "eval_vitaminc-pairs_runtime": 5.4759, + "eval_vitaminc-pairs_samples_per_second": 23.375, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 1420 + }, + { + "epoch": 1.1016291698991467, + "eval_negation-triplets_loss": 1.7074307203292847, + "eval_negation-triplets_runtime": 1.0745, + "eval_negation-triplets_samples_per_second": 119.125, + "eval_negation-triplets_steps_per_second": 1.861, + "step": 1420 + }, + { + "epoch": 1.1016291698991467, + "eval_scitail-pairs-pos_loss": 0.13299213349819183, + "eval_scitail-pairs-pos_runtime": 1.2994, + "eval_scitail-pairs-pos_samples_per_second": 98.503, + "eval_scitail-pairs-pos_steps_per_second": 1.539, + "step": 1420 + }, + { + "epoch": 1.1016291698991467, + "eval_scitail-pairs-qa_loss": 0.007115138694643974, + "eval_scitail-pairs-qa_runtime": 0.7851, + "eval_scitail-pairs-qa_samples_per_second": 163.041, + "eval_scitail-pairs-qa_steps_per_second": 2.548, + "step": 1420 + }, + { + "epoch": 1.1016291698991467, + "eval_xsum-pairs_loss": 0.08819232881069183, + "eval_xsum-pairs_runtime": 3.9492, + "eval_xsum-pairs_samples_per_second": 32.412, + "eval_xsum-pairs_steps_per_second": 0.506, + "step": 1420 + }, + { + "epoch": 1.1016291698991467, + "eval_sciq_pairs_loss": 0.061290107667446136, + "eval_sciq_pairs_runtime": 6.1657, + "eval_sciq_pairs_samples_per_second": 20.76, + "eval_sciq_pairs_steps_per_second": 0.324, + "step": 1420 + }, + { + "epoch": 1.1016291698991467, + "eval_qasc_pairs_loss": 0.3105396628379822, + "eval_qasc_pairs_runtime": 0.9363, + "eval_qasc_pairs_samples_per_second": 136.702, + "eval_qasc_pairs_steps_per_second": 2.136, + "step": 1420 + }, + { + "epoch": 1.1016291698991467, + "eval_openbookqa_pairs_loss": 1.1044814586639404, + "eval_openbookqa_pairs_runtime": 0.8801, + "eval_openbookqa_pairs_samples_per_second": 145.432, + "eval_openbookqa_pairs_steps_per_second": 2.272, + "step": 1420 + }, + { + "epoch": 1.1016291698991467, + "eval_msmarco_pairs_loss": 0.4150286614894867, + "eval_msmarco_pairs_runtime": 1.9764, + "eval_msmarco_pairs_samples_per_second": 64.766, + "eval_msmarco_pairs_steps_per_second": 1.012, + "step": 1420 + }, + { + "epoch": 1.1016291698991467, + "eval_nq_pairs_loss": 0.6082783341407776, + "eval_nq_pairs_runtime": 3.3469, + "eval_nq_pairs_samples_per_second": 38.245, + "eval_nq_pairs_steps_per_second": 0.598, + "step": 1420 + }, + { + "epoch": 1.1016291698991467, + "eval_trivia_pairs_loss": 0.9493303894996643, + "eval_trivia_pairs_runtime": 4.4795, + "eval_trivia_pairs_samples_per_second": 28.575, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 1420 + }, + { + "epoch": 1.1016291698991467, + "eval_gooaq_pairs_loss": 0.6915829181671143, + "eval_gooaq_pairs_runtime": 1.4668, + "eval_gooaq_pairs_samples_per_second": 87.262, + "eval_gooaq_pairs_steps_per_second": 1.363, + "step": 1420 + }, + { + "epoch": 1.1016291698991467, + "eval_paws-pos_loss": 0.03125006705522537, + "eval_paws-pos_runtime": 1.0349, + "eval_paws-pos_samples_per_second": 123.68, + "eval_paws-pos_steps_per_second": 1.933, + "step": 1420 + }, + { + "epoch": 1.1016291698991467, + "eval_global_dataset_loss": 0.4164344072341919, + "eval_global_dataset_runtime": 19.4091, + "eval_global_dataset_samples_per_second": 21.433, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1420 + }, + { + "epoch": 1.1024049650892165, + "grad_norm": 8.372739791870117, + "learning_rate": 2.7339053210021757e-05, + "loss": 0.4228, + "step": 1421 + }, + { + "epoch": 1.1031807602792862, + "grad_norm": 0.7265329957008362, + "learning_rate": 2.731445636542022e-05, + "loss": 0.0158, + "step": 1422 + }, + { + "epoch": 1.103956555469356, + "grad_norm": 0.950513482093811, + "learning_rate": 2.7289832834415265e-05, + "loss": 0.058, + "step": 1423 + }, + { + "epoch": 1.104732350659426, + "grad_norm": 7.69651985168457, + "learning_rate": 2.7265182698489574e-05, + "loss": 0.2635, + "step": 1424 + }, + { + "epoch": 1.1055081458494957, + "grad_norm": 7.246150493621826, + "learning_rate": 2.7240506039213862e-05, + "loss": 0.3355, + "step": 1425 + }, + { + "epoch": 1.1062839410395655, + "grad_norm": 6.487372875213623, + "learning_rate": 2.7215802938246605e-05, + "loss": 0.2386, + "step": 1426 + }, + { + "epoch": 1.1070597362296355, + "grad_norm": 8.316450119018555, + "learning_rate": 2.71910734773338e-05, + "loss": 0.4069, + "step": 1427 + }, + { + "epoch": 1.1078355314197053, + "grad_norm": 7.702548503875732, + "learning_rate": 2.716631773830865e-05, + "loss": 0.3289, + "step": 1428 + }, + { + "epoch": 1.108611326609775, + "grad_norm": 10.460127830505371, + "learning_rate": 2.7141535803091324e-05, + "loss": 0.7937, + "step": 1429 + }, + { + "epoch": 1.1093871217998448, + "grad_norm": 10.539872169494629, + "learning_rate": 2.711672775368868e-05, + "loss": 0.7856, + "step": 1430 + }, + { + "epoch": 1.1101629169899148, + "grad_norm": 5.7495198249816895, + "learning_rate": 2.7091893672193998e-05, + "loss": 0.2626, + "step": 1431 + }, + { + "epoch": 1.1109387121799845, + "grad_norm": 10.13537311553955, + "learning_rate": 2.7067033640786686e-05, + "loss": 0.8485, + "step": 1432 + }, + { + "epoch": 1.1117145073700543, + "grad_norm": 0.844298779964447, + "learning_rate": 2.704214774173204e-05, + "loss": 0.019, + "step": 1433 + }, + { + "epoch": 1.112490302560124, + "grad_norm": 8.26998233795166, + "learning_rate": 2.701723605738094e-05, + "loss": 0.6568, + "step": 1434 + }, + { + "epoch": 1.113266097750194, + "grad_norm": 8.675110816955566, + "learning_rate": 2.6992298670169602e-05, + "loss": 0.5833, + "step": 1435 + }, + { + "epoch": 1.1140418929402638, + "grad_norm": 8.054617881774902, + "learning_rate": 2.6967335662619307e-05, + "loss": 0.442, + "step": 1436 + }, + { + "epoch": 1.1148176881303336, + "grad_norm": 6.371271133422852, + "learning_rate": 2.6942347117336086e-05, + "loss": 0.3428, + "step": 1437 + }, + { + "epoch": 1.1155934833204033, + "grad_norm": 5.796209812164307, + "learning_rate": 2.6917333117010508e-05, + "loss": 0.2306, + "step": 1438 + }, + { + "epoch": 1.1163692785104733, + "grad_norm": 8.226943016052246, + "learning_rate": 2.689229374441737e-05, + "loss": 0.5905, + "step": 1439 + }, + { + "epoch": 1.117145073700543, + "grad_norm": 5.173073768615723, + "learning_rate": 2.6867229082415413e-05, + "loss": 0.226, + "step": 1440 + }, + { + "epoch": 1.117145073700543, + "eval_Qnli-dev_cosine_accuracy": 0.6796875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8125748634338379, + "eval_Qnli-dev_cosine_ap": 0.7198942617466971, + "eval_Qnli-dev_cosine_f1": 0.6898954703832751, + "eval_Qnli-dev_cosine_f1_threshold": 0.7642830014228821, + "eval_Qnli-dev_cosine_precision": 0.5857988165680473, + "eval_Qnli-dev_cosine_recall": 0.8389830508474576, + "eval_Qnli-dev_dot_accuracy": 0.64453125, + "eval_Qnli-dev_dot_accuracy_threshold": 402.282958984375, + "eval_Qnli-dev_dot_ap": 0.6256403192422778, + "eval_Qnli-dev_dot_f1": 0.6616766467065869, + "eval_Qnli-dev_dot_f1_threshold": 326.68896484375, + "eval_Qnli-dev_dot_precision": 0.5115740740740741, + "eval_Qnli-dev_dot_recall": 0.9364406779661016, + "eval_Qnli-dev_euclidean_accuracy": 0.68359375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.980140686035156, + "eval_Qnli-dev_euclidean_ap": 0.7244630317872427, + "eval_Qnli-dev_euclidean_f1": 0.6928327645051194, + "eval_Qnli-dev_euclidean_f1_threshold": 15.435157775878906, + "eval_Qnli-dev_euclidean_precision": 0.58, + "eval_Qnli-dev_euclidean_recall": 0.8601694915254238, + "eval_Qnli-dev_manhattan_accuracy": 0.689453125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 249.00460815429688, + "eval_Qnli-dev_manhattan_ap": 0.7223725900544616, + "eval_Qnli-dev_manhattan_f1": 0.6862745098039217, + "eval_Qnli-dev_manhattan_f1_threshold": 252.96636962890625, + "eval_Qnli-dev_manhattan_precision": 0.6386861313868614, + "eval_Qnli-dev_manhattan_recall": 0.7415254237288136, + "eval_Qnli-dev_max_accuracy": 0.689453125, + "eval_Qnli-dev_max_accuracy_threshold": 402.282958984375, + "eval_Qnli-dev_max_ap": 0.7244630317872427, + "eval_Qnli-dev_max_f1": 0.6928327645051194, + "eval_Qnli-dev_max_f1_threshold": 326.68896484375, + "eval_Qnli-dev_max_precision": 0.6386861313868614, + "eval_Qnli-dev_max_recall": 0.9364406779661016, + "eval_allNLI-dev_cosine_accuracy": 0.72265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8902768492698669, + "eval_allNLI-dev_cosine_ap": 0.6050046208045297, + "eval_allNLI-dev_cosine_f1": 0.6125, + "eval_allNLI-dev_cosine_f1_threshold": 0.787812352180481, + "eval_allNLI-dev_cosine_precision": 0.4788273615635179, + "eval_allNLI-dev_cosine_recall": 0.8497109826589595, + "eval_allNLI-dev_dot_accuracy": 0.693359375, + "eval_allNLI-dev_dot_accuracy_threshold": 410.6229248046875, + "eval_allNLI-dev_dot_ap": 0.5129505393979772, + "eval_allNLI-dev_dot_f1": 0.5828092243186582, + "eval_allNLI-dev_dot_f1_threshold": 357.3798828125, + "eval_allNLI-dev_dot_precision": 0.45723684210526316, + "eval_allNLI-dev_dot_recall": 0.8034682080924855, + "eval_allNLI-dev_euclidean_accuracy": 0.7265625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.413698196411133, + "eval_allNLI-dev_euclidean_ap": 0.608900829281882, + "eval_allNLI-dev_euclidean_f1": 0.6126482213438735, + "eval_allNLI-dev_euclidean_f1_threshold": 14.72088623046875, + "eval_allNLI-dev_euclidean_precision": 0.46546546546546547, + "eval_allNLI-dev_euclidean_recall": 0.8959537572254336, + "eval_allNLI-dev_manhattan_accuracy": 0.73828125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 197.18508911132812, + "eval_allNLI-dev_manhattan_ap": 0.6101816036146291, + "eval_allNLI-dev_manhattan_f1": 0.6147186147186148, + "eval_allNLI-dev_manhattan_f1_threshold": 247.49366760253906, + "eval_allNLI-dev_manhattan_precision": 0.4913494809688581, + "eval_allNLI-dev_manhattan_recall": 0.8208092485549133, + "eval_allNLI-dev_max_accuracy": 0.73828125, + "eval_allNLI-dev_max_accuracy_threshold": 410.6229248046875, + "eval_allNLI-dev_max_ap": 0.6101816036146291, + "eval_allNLI-dev_max_f1": 0.6147186147186148, + "eval_allNLI-dev_max_f1_threshold": 357.3798828125, + "eval_allNLI-dev_max_precision": 0.4913494809688581, + "eval_allNLI-dev_max_recall": 0.8959537572254336, + "eval_sequential_score": 0.7244630317872427, + "eval_sts-test_pearson_cosine": 0.8318146673658309, + "eval_sts-test_pearson_dot": 0.7735849614058687, + "eval_sts-test_pearson_euclidean": 0.86685209086377, + "eval_sts-test_pearson_manhattan": 0.8656695472460445, + "eval_sts-test_pearson_max": 0.86685209086377, + "eval_sts-test_spearman_cosine": 0.8683412589443563, + "eval_sts-test_spearman_dot": 0.7632718756327749, + "eval_sts-test_spearman_euclidean": 0.8675143003851048, + "eval_sts-test_spearman_manhattan": 0.8661032931134337, + "eval_sts-test_spearman_max": 0.8683412589443563, + "eval_vitaminc-pairs_loss": 3.6102519035339355, + "eval_vitaminc-pairs_runtime": 5.4599, + "eval_vitaminc-pairs_samples_per_second": 23.444, + "eval_vitaminc-pairs_steps_per_second": 0.366, + "step": 1440 + }, + { + "epoch": 1.117145073700543, + "eval_negation-triplets_loss": 1.6761524677276611, + "eval_negation-triplets_runtime": 1.0696, + "eval_negation-triplets_samples_per_second": 119.666, + "eval_negation-triplets_steps_per_second": 1.87, + "step": 1440 + }, + { + "epoch": 1.117145073700543, + "eval_scitail-pairs-pos_loss": 0.13355915248394012, + "eval_scitail-pairs-pos_runtime": 1.265, + "eval_scitail-pairs-pos_samples_per_second": 101.184, + "eval_scitail-pairs-pos_steps_per_second": 1.581, + "step": 1440 + }, + { + "epoch": 1.117145073700543, + "eval_scitail-pairs-qa_loss": 0.008299741894006729, + "eval_scitail-pairs-qa_runtime": 0.769, + "eval_scitail-pairs-qa_samples_per_second": 166.454, + "eval_scitail-pairs-qa_steps_per_second": 2.601, + "step": 1440 + }, + { + "epoch": 1.117145073700543, + "eval_xsum-pairs_loss": 0.13131901621818542, + "eval_xsum-pairs_runtime": 3.917, + "eval_xsum-pairs_samples_per_second": 32.678, + "eval_xsum-pairs_steps_per_second": 0.511, + "step": 1440 + }, + { + "epoch": 1.117145073700543, + "eval_sciq_pairs_loss": 0.05616779252886772, + "eval_sciq_pairs_runtime": 6.0522, + "eval_sciq_pairs_samples_per_second": 21.149, + "eval_sciq_pairs_steps_per_second": 0.33, + "step": 1440 + }, + { + "epoch": 1.117145073700543, + "eval_qasc_pairs_loss": 0.2950361967086792, + "eval_qasc_pairs_runtime": 0.9451, + "eval_qasc_pairs_samples_per_second": 135.435, + "eval_qasc_pairs_steps_per_second": 2.116, + "step": 1440 + }, + { + "epoch": 1.117145073700543, + "eval_openbookqa_pairs_loss": 1.1163709163665771, + "eval_openbookqa_pairs_runtime": 0.8792, + "eval_openbookqa_pairs_samples_per_second": 145.589, + "eval_openbookqa_pairs_steps_per_second": 2.275, + "step": 1440 + }, + { + "epoch": 1.117145073700543, + "eval_msmarco_pairs_loss": 0.3954874277114868, + "eval_msmarco_pairs_runtime": 1.9667, + "eval_msmarco_pairs_samples_per_second": 65.084, + "eval_msmarco_pairs_steps_per_second": 1.017, + "step": 1440 + }, + { + "epoch": 1.117145073700543, + "eval_nq_pairs_loss": 0.5966031551361084, + "eval_nq_pairs_runtime": 3.3305, + "eval_nq_pairs_samples_per_second": 38.432, + "eval_nq_pairs_steps_per_second": 0.601, + "step": 1440 + }, + { + "epoch": 1.117145073700543, + "eval_trivia_pairs_loss": 0.9561909437179565, + "eval_trivia_pairs_runtime": 4.5041, + "eval_trivia_pairs_samples_per_second": 28.418, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 1440 + }, + { + "epoch": 1.117145073700543, + "eval_gooaq_pairs_loss": 0.6156429052352905, + "eval_gooaq_pairs_runtime": 1.4778, + "eval_gooaq_pairs_samples_per_second": 86.614, + "eval_gooaq_pairs_steps_per_second": 1.353, + "step": 1440 + }, + { + "epoch": 1.117145073700543, + "eval_paws-pos_loss": 0.0324188657104969, + "eval_paws-pos_runtime": 1.0363, + "eval_paws-pos_samples_per_second": 123.521, + "eval_paws-pos_steps_per_second": 1.93, + "step": 1440 + }, + { + "epoch": 1.117145073700543, + "eval_global_dataset_loss": 0.4049144387245178, + "eval_global_dataset_runtime": 19.4393, + "eval_global_dataset_samples_per_second": 21.4, + "eval_global_dataset_steps_per_second": 0.36, + "step": 1440 + }, + { + "epoch": 1.1179208688906128, + "grad_norm": 16.199684143066406, + "learning_rate": 2.6842139213947084e-05, + "loss": 2.5883, + "step": 1441 + }, + { + "epoch": 1.1186966640806828, + "grad_norm": 8.538520812988281, + "learning_rate": 2.681702422203824e-05, + "loss": 0.6197, + "step": 1442 + }, + { + "epoch": 1.1194724592707526, + "grad_norm": 1.6734013557434082, + "learning_rate": 2.6791884189797854e-05, + "loss": 0.0428, + "step": 1443 + }, + { + "epoch": 1.1202482544608223, + "grad_norm": 8.074939727783203, + "learning_rate": 2.676671920041779e-05, + "loss": 0.4969, + "step": 1444 + }, + { + "epoch": 1.121024049650892, + "grad_norm": 8.805115699768066, + "learning_rate": 2.6741529337172477e-05, + "loss": 0.7399, + "step": 1445 + }, + { + "epoch": 1.121799844840962, + "grad_norm": 7.280085563659668, + "learning_rate": 2.6716314683418662e-05, + "loss": 0.3279, + "step": 1446 + }, + { + "epoch": 1.1225756400310318, + "grad_norm": 5.1095147132873535, + "learning_rate": 2.6691075322595128e-05, + "loss": 0.2188, + "step": 1447 + }, + { + "epoch": 1.1233514352211016, + "grad_norm": 10.979981422424316, + "learning_rate": 2.6665811338222422e-05, + "loss": 1.0943, + "step": 1448 + }, + { + "epoch": 1.1241272304111714, + "grad_norm": 6.22988224029541, + "learning_rate": 2.6640522813902558e-05, + "loss": 0.3005, + "step": 1449 + }, + { + "epoch": 1.1249030256012413, + "grad_norm": 6.932554721832275, + "learning_rate": 2.661520983331878e-05, + "loss": 0.4593, + "step": 1450 + }, + { + "epoch": 1.125678820791311, + "grad_norm": 6.561832427978516, + "learning_rate": 2.6589872480235234e-05, + "loss": 0.2467, + "step": 1451 + }, + { + "epoch": 1.1264546159813809, + "grad_norm": 7.054421424865723, + "learning_rate": 2.6564510838496743e-05, + "loss": 0.392, + "step": 1452 + }, + { + "epoch": 1.1272304111714506, + "grad_norm": 3.8438827991485596, + "learning_rate": 2.653912499202849e-05, + "loss": 0.1271, + "step": 1453 + }, + { + "epoch": 1.1280062063615206, + "grad_norm": 7.819214820861816, + "learning_rate": 2.651371502483575e-05, + "loss": 0.6121, + "step": 1454 + }, + { + "epoch": 1.1287820015515904, + "grad_norm": 6.482337951660156, + "learning_rate": 2.648828102100364e-05, + "loss": 0.2908, + "step": 1455 + }, + { + "epoch": 1.1295577967416601, + "grad_norm": 7.898482799530029, + "learning_rate": 2.646282306469679e-05, + "loss": 0.483, + "step": 1456 + }, + { + "epoch": 1.13033359193173, + "grad_norm": 8.907037734985352, + "learning_rate": 2.643734124015911e-05, + "loss": 0.7697, + "step": 1457 + }, + { + "epoch": 1.1311093871217999, + "grad_norm": 5.938931465148926, + "learning_rate": 2.6411835631713485e-05, + "loss": 0.2909, + "step": 1458 + }, + { + "epoch": 1.1318851823118696, + "grad_norm": 9.068387031555176, + "learning_rate": 2.6386306323761513e-05, + "loss": 0.6871, + "step": 1459 + }, + { + "epoch": 1.1326609775019394, + "grad_norm": 6.06436824798584, + "learning_rate": 2.63607534007832e-05, + "loss": 0.3756, + "step": 1460 + }, + { + "epoch": 1.1326609775019394, + "eval_Qnli-dev_cosine_accuracy": 0.681640625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8425674438476562, + "eval_Qnli-dev_cosine_ap": 0.7163981219155651, + "eval_Qnli-dev_cosine_f1": 0.6807131280388979, + "eval_Qnli-dev_cosine_f1_threshold": 0.7558989524841309, + "eval_Qnli-dev_cosine_precision": 0.5511811023622047, + "eval_Qnli-dev_cosine_recall": 0.8898305084745762, + "eval_Qnli-dev_dot_accuracy": 0.6484375, + "eval_Qnli-dev_dot_accuracy_threshold": 427.9423828125, + "eval_Qnli-dev_dot_ap": 0.6292059966433998, + "eval_Qnli-dev_dot_f1": 0.6627043090638931, + "eval_Qnli-dev_dot_f1_threshold": 329.7099304199219, + "eval_Qnli-dev_dot_precision": 0.5102974828375286, + "eval_Qnli-dev_dot_recall": 0.9449152542372882, + "eval_Qnli-dev_euclidean_accuracy": 0.673828125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.458602905273438, + "eval_Qnli-dev_euclidean_ap": 0.7147378575580118, + "eval_Qnli-dev_euclidean_f1": 0.6862068965517241, + "eval_Qnli-dev_euclidean_f1_threshold": 14.913080215454102, + "eval_Qnli-dev_euclidean_precision": 0.5784883720930233, + "eval_Qnli-dev_euclidean_recall": 0.8432203389830508, + "eval_Qnli-dev_manhattan_accuracy": 0.68359375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 239.92735290527344, + "eval_Qnli-dev_manhattan_ap": 0.7171023986170371, + "eval_Qnli-dev_manhattan_f1": 0.6825396825396826, + "eval_Qnli-dev_manhattan_f1_threshold": 286.45257568359375, + "eval_Qnli-dev_manhattan_precision": 0.5456852791878173, + "eval_Qnli-dev_manhattan_recall": 0.9110169491525424, + "eval_Qnli-dev_max_accuracy": 0.68359375, + "eval_Qnli-dev_max_accuracy_threshold": 427.9423828125, + "eval_Qnli-dev_max_ap": 0.7171023986170371, + "eval_Qnli-dev_max_f1": 0.6862068965517241, + "eval_Qnli-dev_max_f1_threshold": 329.7099304199219, + "eval_Qnli-dev_max_precision": 0.5784883720930233, + "eval_Qnli-dev_max_recall": 0.9449152542372882, + "eval_allNLI-dev_cosine_accuracy": 0.7265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8857695460319519, + "eval_allNLI-dev_cosine_ap": 0.6133304602423284, + "eval_allNLI-dev_cosine_f1": 0.6109890109890109, + "eval_allNLI-dev_cosine_f1_threshold": 0.8157030344009399, + "eval_allNLI-dev_cosine_precision": 0.4929078014184397, + "eval_allNLI-dev_cosine_recall": 0.8034682080924855, + "eval_allNLI-dev_dot_accuracy": 0.6875, + "eval_allNLI-dev_dot_accuracy_threshold": 426.5369873046875, + "eval_allNLI-dev_dot_ap": 0.5183834456851001, + "eval_allNLI-dev_dot_f1": 0.5810810810810811, + "eval_allNLI-dev_dot_f1_threshold": 379.0350341796875, + "eval_allNLI-dev_dot_precision": 0.47601476014760147, + "eval_allNLI-dev_dot_recall": 0.7456647398843931, + "eval_allNLI-dev_euclidean_accuracy": 0.724609375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.410016059875488, + "eval_allNLI-dev_euclidean_ap": 0.6134842114515171, + "eval_allNLI-dev_euclidean_f1": 0.6120689655172413, + "eval_allNLI-dev_euclidean_f1_threshold": 13.430139541625977, + "eval_allNLI-dev_euclidean_precision": 0.4879725085910653, + "eval_allNLI-dev_euclidean_recall": 0.8208092485549133, + "eval_allNLI-dev_manhattan_accuracy": 0.73828125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 193.06277465820312, + "eval_allNLI-dev_manhattan_ap": 0.6161928632866415, + "eval_allNLI-dev_manhattan_f1": 0.6136865342163355, + "eval_allNLI-dev_manhattan_f1_threshold": 237.98532104492188, + "eval_allNLI-dev_manhattan_precision": 0.49642857142857144, + "eval_allNLI-dev_manhattan_recall": 0.8034682080924855, + "eval_allNLI-dev_max_accuracy": 0.73828125, + "eval_allNLI-dev_max_accuracy_threshold": 426.5369873046875, + "eval_allNLI-dev_max_ap": 0.6161928632866415, + "eval_allNLI-dev_max_f1": 0.6136865342163355, + "eval_allNLI-dev_max_f1_threshold": 379.0350341796875, + "eval_allNLI-dev_max_precision": 0.49642857142857144, + "eval_allNLI-dev_max_recall": 0.8208092485549133, + "eval_sequential_score": 0.7171023986170371, + "eval_sts-test_pearson_cosine": 0.8311052852276188, + "eval_sts-test_pearson_dot": 0.770136132980644, + "eval_sts-test_pearson_euclidean": 0.8657450718755565, + "eval_sts-test_pearson_manhattan": 0.8633827683247665, + "eval_sts-test_pearson_max": 0.8657450718755565, + "eval_sts-test_spearman_cosine": 0.8675187755112438, + "eval_sts-test_spearman_dot": 0.758241654847336, + "eval_sts-test_spearman_euclidean": 0.8647464796193078, + "eval_sts-test_spearman_manhattan": 0.8630040445057724, + "eval_sts-test_spearman_max": 0.8675187755112438, + "eval_vitaminc-pairs_loss": 3.4783616065979004, + "eval_vitaminc-pairs_runtime": 5.4726, + "eval_vitaminc-pairs_samples_per_second": 23.389, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 1460 + }, + { + "epoch": 1.1326609775019394, + "eval_negation-triplets_loss": 1.6732555627822876, + "eval_negation-triplets_runtime": 1.0776, + "eval_negation-triplets_samples_per_second": 118.786, + "eval_negation-triplets_steps_per_second": 1.856, + "step": 1460 + }, + { + "epoch": 1.1326609775019394, + "eval_scitail-pairs-pos_loss": 0.14788450300693512, + "eval_scitail-pairs-pos_runtime": 1.2656, + "eval_scitail-pairs-pos_samples_per_second": 101.139, + "eval_scitail-pairs-pos_steps_per_second": 1.58, + "step": 1460 + }, + { + "epoch": 1.1326609775019394, + "eval_scitail-pairs-qa_loss": 0.011658175848424435, + "eval_scitail-pairs-qa_runtime": 0.7705, + "eval_scitail-pairs-qa_samples_per_second": 166.132, + "eval_scitail-pairs-qa_steps_per_second": 2.596, + "step": 1460 + }, + { + "epoch": 1.1326609775019394, + "eval_xsum-pairs_loss": 0.13660183548927307, + "eval_xsum-pairs_runtime": 3.9255, + "eval_xsum-pairs_samples_per_second": 32.608, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 1460 + }, + { + "epoch": 1.1326609775019394, + "eval_sciq_pairs_loss": 0.05248516798019409, + "eval_sciq_pairs_runtime": 6.1157, + "eval_sciq_pairs_samples_per_second": 20.93, + "eval_sciq_pairs_steps_per_second": 0.327, + "step": 1460 + }, + { + "epoch": 1.1326609775019394, + "eval_qasc_pairs_loss": 0.2878599762916565, + "eval_qasc_pairs_runtime": 0.9333, + "eval_qasc_pairs_samples_per_second": 137.151, + "eval_qasc_pairs_steps_per_second": 2.143, + "step": 1460 + }, + { + "epoch": 1.1326609775019394, + "eval_openbookqa_pairs_loss": 1.0870670080184937, + "eval_openbookqa_pairs_runtime": 0.8785, + "eval_openbookqa_pairs_samples_per_second": 145.705, + "eval_openbookqa_pairs_steps_per_second": 2.277, + "step": 1460 + }, + { + "epoch": 1.1326609775019394, + "eval_msmarco_pairs_loss": 0.37255820631980896, + "eval_msmarco_pairs_runtime": 1.9673, + "eval_msmarco_pairs_samples_per_second": 65.064, + "eval_msmarco_pairs_steps_per_second": 1.017, + "step": 1460 + }, + { + "epoch": 1.1326609775019394, + "eval_nq_pairs_loss": 0.6457039713859558, + "eval_nq_pairs_runtime": 3.3335, + "eval_nq_pairs_samples_per_second": 38.398, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 1460 + }, + { + "epoch": 1.1326609775019394, + "eval_trivia_pairs_loss": 0.9141166806221008, + "eval_trivia_pairs_runtime": 4.4898, + "eval_trivia_pairs_samples_per_second": 28.509, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 1460 + }, + { + "epoch": 1.1326609775019394, + "eval_gooaq_pairs_loss": 0.567143440246582, + "eval_gooaq_pairs_runtime": 1.4872, + "eval_gooaq_pairs_samples_per_second": 86.07, + "eval_gooaq_pairs_steps_per_second": 1.345, + "step": 1460 + }, + { + "epoch": 1.1326609775019394, + "eval_paws-pos_loss": 0.03394191339612007, + "eval_paws-pos_runtime": 1.0575, + "eval_paws-pos_samples_per_second": 121.041, + "eval_paws-pos_steps_per_second": 1.891, + "step": 1460 + }, + { + "epoch": 1.1326609775019394, + "eval_global_dataset_loss": 0.42527249455451965, + "eval_global_dataset_runtime": 19.476, + "eval_global_dataset_samples_per_second": 21.36, + "eval_global_dataset_steps_per_second": 0.359, + "step": 1460 + }, + { + "epoch": 1.1334367726920094, + "grad_norm": 5.224971771240234, + "learning_rate": 2.6335176947336718e-05, + "loss": 0.1857, + "step": 1461 + }, + { + "epoch": 1.1342125678820791, + "grad_norm": 0.24938639998435974, + "learning_rate": 2.630957704805809e-05, + "loss": 0.0048, + "step": 1462 + }, + { + "epoch": 1.134988363072149, + "grad_norm": 7.411340236663818, + "learning_rate": 2.6283953787660934e-05, + "loss": 0.392, + "step": 1463 + }, + { + "epoch": 1.1357641582622189, + "grad_norm": 7.127811431884766, + "learning_rate": 2.625830725093617e-05, + "loss": 0.3265, + "step": 1464 + }, + { + "epoch": 1.1365399534522886, + "grad_norm": 7.088199615478516, + "learning_rate": 2.623263752275173e-05, + "loss": 0.2754, + "step": 1465 + }, + { + "epoch": 1.1373157486423584, + "grad_norm": 9.551557540893555, + "learning_rate": 2.6206944688052315e-05, + "loss": 0.6548, + "step": 1466 + }, + { + "epoch": 1.1380915438324282, + "grad_norm": 7.111812591552734, + "learning_rate": 2.6181228831859063e-05, + "loss": 0.2368, + "step": 1467 + }, + { + "epoch": 1.1388673390224981, + "grad_norm": 10.905604362487793, + "learning_rate": 2.615549003926931e-05, + "loss": 0.7451, + "step": 1468 + }, + { + "epoch": 1.139643134212568, + "grad_norm": 7.544840335845947, + "learning_rate": 2.612972839545629e-05, + "loss": 0.4304, + "step": 1469 + }, + { + "epoch": 1.1404189294026377, + "grad_norm": 5.333651542663574, + "learning_rate": 2.6103943985668853e-05, + "loss": 0.3603, + "step": 1470 + }, + { + "epoch": 1.1411947245927074, + "grad_norm": 9.735820770263672, + "learning_rate": 2.607813689523117e-05, + "loss": 0.8947, + "step": 1471 + }, + { + "epoch": 1.1419705197827774, + "grad_norm": 4.950901985168457, + "learning_rate": 2.605230720954249e-05, + "loss": 0.1603, + "step": 1472 + }, + { + "epoch": 1.1427463149728472, + "grad_norm": 5.766415596008301, + "learning_rate": 2.602645501407682e-05, + "loss": 0.2344, + "step": 1473 + }, + { + "epoch": 1.143522110162917, + "grad_norm": 2.239945650100708, + "learning_rate": 2.6000580394382655e-05, + "loss": 0.0523, + "step": 1474 + }, + { + "epoch": 1.1442979053529867, + "grad_norm": 7.647494316101074, + "learning_rate": 2.59746834360827e-05, + "loss": 0.2711, + "step": 1475 + }, + { + "epoch": 1.1450737005430567, + "grad_norm": 7.284696578979492, + "learning_rate": 2.5948764224873566e-05, + "loss": 0.5757, + "step": 1476 + }, + { + "epoch": 1.1458494957331264, + "grad_norm": 7.674700736999512, + "learning_rate": 2.592282284652553e-05, + "loss": 0.2832, + "step": 1477 + }, + { + "epoch": 1.1466252909231962, + "grad_norm": 10.532242774963379, + "learning_rate": 2.5896859386882198e-05, + "loss": 0.6805, + "step": 1478 + }, + { + "epoch": 1.1474010861132662, + "grad_norm": 7.4618682861328125, + "learning_rate": 2.5870873931860257e-05, + "loss": 0.2861, + "step": 1479 + }, + { + "epoch": 1.148176881303336, + "grad_norm": 5.702700138092041, + "learning_rate": 2.5844866567449178e-05, + "loss": 0.2553, + "step": 1480 + }, + { + "epoch": 1.148176881303336, + "eval_Qnli-dev_cosine_accuracy": 0.658203125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8395674228668213, + "eval_Qnli-dev_cosine_ap": 0.6958314766705993, + "eval_Qnli-dev_cosine_f1": 0.6740858505564389, + "eval_Qnli-dev_cosine_f1_threshold": 0.7243646383285522, + "eval_Qnli-dev_cosine_precision": 0.539440203562341, + "eval_Qnli-dev_cosine_recall": 0.8983050847457628, + "eval_Qnli-dev_dot_accuracy": 0.64453125, + "eval_Qnli-dev_dot_accuracy_threshold": 415.52166748046875, + "eval_Qnli-dev_dot_ap": 0.6167987996206776, + "eval_Qnli-dev_dot_f1": 0.6728682170542635, + "eval_Qnli-dev_dot_f1_threshold": 334.96197509765625, + "eval_Qnli-dev_dot_precision": 0.530562347188264, + "eval_Qnli-dev_dot_recall": 0.9194915254237288, + "eval_Qnli-dev_euclidean_accuracy": 0.666015625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 11.931020736694336, + "eval_Qnli-dev_euclidean_ap": 0.6964878543754244, + "eval_Qnli-dev_euclidean_f1": 0.6752411575562701, + "eval_Qnli-dev_euclidean_f1_threshold": 16.25579833984375, + "eval_Qnli-dev_euclidean_precision": 0.5440414507772021, + "eval_Qnli-dev_euclidean_recall": 0.8898305084745762, + "eval_Qnli-dev_manhattan_accuracy": 0.66015625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 210.19219970703125, + "eval_Qnli-dev_manhattan_ap": 0.6988306616607896, + "eval_Qnli-dev_manhattan_f1": 0.6718750000000001, + "eval_Qnli-dev_manhattan_f1_threshold": 302.6781311035156, + "eval_Qnli-dev_manhattan_precision": 0.5321782178217822, + "eval_Qnli-dev_manhattan_recall": 0.9110169491525424, + "eval_Qnli-dev_max_accuracy": 0.666015625, + "eval_Qnli-dev_max_accuracy_threshold": 415.52166748046875, + "eval_Qnli-dev_max_ap": 0.6988306616607896, + "eval_Qnli-dev_max_f1": 0.6752411575562701, + "eval_Qnli-dev_max_f1_threshold": 334.96197509765625, + "eval_Qnli-dev_max_precision": 0.5440414507772021, + "eval_Qnli-dev_max_recall": 0.9194915254237288, + "eval_allNLI-dev_cosine_accuracy": 0.71875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8620715737342834, + "eval_allNLI-dev_cosine_ap": 0.5938865602030686, + "eval_allNLI-dev_cosine_f1": 0.5983935742971888, + "eval_allNLI-dev_cosine_f1_threshold": 0.7492684125900269, + "eval_allNLI-dev_cosine_precision": 0.4584615384615385, + "eval_allNLI-dev_cosine_recall": 0.861271676300578, + "eval_allNLI-dev_dot_accuracy": 0.6796875, + "eval_allNLI-dev_dot_accuracy_threshold": 393.6879577636719, + "eval_allNLI-dev_dot_ap": 0.49909312411536505, + "eval_allNLI-dev_dot_f1": 0.5725190839694657, + "eval_allNLI-dev_dot_f1_threshold": 330.7287902832031, + "eval_allNLI-dev_dot_precision": 0.42735042735042733, + "eval_allNLI-dev_dot_recall": 0.8670520231213873, + "eval_allNLI-dev_euclidean_accuracy": 0.716796875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.306994438171387, + "eval_allNLI-dev_euclidean_ap": 0.5976689557235553, + "eval_allNLI-dev_euclidean_f1": 0.6073752711496746, + "eval_allNLI-dev_euclidean_f1_threshold": 14.391624450683594, + "eval_allNLI-dev_euclidean_precision": 0.4861111111111111, + "eval_allNLI-dev_euclidean_recall": 0.8092485549132948, + "eval_allNLI-dev_manhattan_accuracy": 0.73046875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 203.8892822265625, + "eval_allNLI-dev_manhattan_ap": 0.6002741732903955, + "eval_allNLI-dev_manhattan_f1": 0.6004140786749482, + "eval_allNLI-dev_manhattan_f1_threshold": 269.5694580078125, + "eval_allNLI-dev_manhattan_precision": 0.46774193548387094, + "eval_allNLI-dev_manhattan_recall": 0.838150289017341, + "eval_allNLI-dev_max_accuracy": 0.73046875, + "eval_allNLI-dev_max_accuracy_threshold": 393.6879577636719, + "eval_allNLI-dev_max_ap": 0.6002741732903955, + "eval_allNLI-dev_max_f1": 0.6073752711496746, + "eval_allNLI-dev_max_f1_threshold": 330.7287902832031, + "eval_allNLI-dev_max_precision": 0.4861111111111111, + "eval_allNLI-dev_max_recall": 0.8670520231213873, + "eval_sequential_score": 0.6988306616607896, + "eval_sts-test_pearson_cosine": 0.8350004326497902, + "eval_sts-test_pearson_dot": 0.8086822165058274, + "eval_sts-test_pearson_euclidean": 0.8695129520911423, + "eval_sts-test_pearson_manhattan": 0.8671401728898951, + "eval_sts-test_pearson_max": 0.8695129520911423, + "eval_sts-test_spearman_cosine": 0.8744849806157359, + "eval_sts-test_spearman_dot": 0.8075704780152003, + "eval_sts-test_spearman_euclidean": 0.8692574067675609, + "eval_sts-test_spearman_manhattan": 0.8677102661587376, + "eval_sts-test_spearman_max": 0.8744849806157359, + "eval_vitaminc-pairs_loss": 3.6984663009643555, + "eval_vitaminc-pairs_runtime": 5.5322, + "eval_vitaminc-pairs_samples_per_second": 23.137, + "eval_vitaminc-pairs_steps_per_second": 0.362, + "step": 1480 + }, + { + "epoch": 1.148176881303336, + "eval_negation-triplets_loss": 1.7224050760269165, + "eval_negation-triplets_runtime": 1.0944, + "eval_negation-triplets_samples_per_second": 116.964, + "eval_negation-triplets_steps_per_second": 1.828, + "step": 1480 + }, + { + "epoch": 1.148176881303336, + "eval_scitail-pairs-pos_loss": 0.11527465283870697, + "eval_scitail-pairs-pos_runtime": 1.3198, + "eval_scitail-pairs-pos_samples_per_second": 96.986, + "eval_scitail-pairs-pos_steps_per_second": 1.515, + "step": 1480 + }, + { + "epoch": 1.148176881303336, + "eval_scitail-pairs-qa_loss": 0.010091670788824558, + "eval_scitail-pairs-qa_runtime": 0.7916, + "eval_scitail-pairs-qa_samples_per_second": 161.707, + "eval_scitail-pairs-qa_steps_per_second": 2.527, + "step": 1480 + }, + { + "epoch": 1.148176881303336, + "eval_xsum-pairs_loss": 0.1435304582118988, + "eval_xsum-pairs_runtime": 3.9417, + "eval_xsum-pairs_samples_per_second": 32.474, + "eval_xsum-pairs_steps_per_second": 0.507, + "step": 1480 + }, + { + "epoch": 1.148176881303336, + "eval_sciq_pairs_loss": 0.047609906643629074, + "eval_sciq_pairs_runtime": 6.1411, + "eval_sciq_pairs_samples_per_second": 20.843, + "eval_sciq_pairs_steps_per_second": 0.326, + "step": 1480 + }, + { + "epoch": 1.148176881303336, + "eval_qasc_pairs_loss": 0.2893540859222412, + "eval_qasc_pairs_runtime": 0.9339, + "eval_qasc_pairs_samples_per_second": 137.052, + "eval_qasc_pairs_steps_per_second": 2.141, + "step": 1480 + }, + { + "epoch": 1.148176881303336, + "eval_openbookqa_pairs_loss": 1.0783970355987549, + "eval_openbookqa_pairs_runtime": 0.9029, + "eval_openbookqa_pairs_samples_per_second": 141.772, + "eval_openbookqa_pairs_steps_per_second": 2.215, + "step": 1480 + }, + { + "epoch": 1.148176881303336, + "eval_msmarco_pairs_loss": 0.3920026421546936, + "eval_msmarco_pairs_runtime": 1.9909, + "eval_msmarco_pairs_samples_per_second": 64.292, + "eval_msmarco_pairs_steps_per_second": 1.005, + "step": 1480 + }, + { + "epoch": 1.148176881303336, + "eval_nq_pairs_loss": 0.7322936058044434, + "eval_nq_pairs_runtime": 3.3629, + "eval_nq_pairs_samples_per_second": 38.063, + "eval_nq_pairs_steps_per_second": 0.595, + "step": 1480 + }, + { + "epoch": 1.148176881303336, + "eval_trivia_pairs_loss": 0.8554942607879639, + "eval_trivia_pairs_runtime": 4.491, + "eval_trivia_pairs_samples_per_second": 28.502, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 1480 + }, + { + "epoch": 1.148176881303336, + "eval_gooaq_pairs_loss": 0.6243541240692139, + "eval_gooaq_pairs_runtime": 1.481, + "eval_gooaq_pairs_samples_per_second": 86.428, + "eval_gooaq_pairs_steps_per_second": 1.35, + "step": 1480 + }, + { + "epoch": 1.148176881303336, + "eval_paws-pos_loss": 0.03402430936694145, + "eval_paws-pos_runtime": 1.0476, + "eval_paws-pos_samples_per_second": 122.185, + "eval_paws-pos_steps_per_second": 1.909, + "step": 1480 + }, + { + "epoch": 1.148176881303336, + "eval_global_dataset_loss": 0.44348788261413574, + "eval_global_dataset_runtime": 19.4559, + "eval_global_dataset_samples_per_second": 21.382, + "eval_global_dataset_steps_per_second": 0.36, + "step": 1480 + }, + { + "epoch": 1.1489526764934057, + "grad_norm": 9.776398658752441, + "learning_rate": 2.5818837379710934e-05, + "loss": 0.7319, + "step": 1481 + }, + { + "epoch": 1.1497284716834755, + "grad_norm": 12.09112548828125, + "learning_rate": 2.579278645477971e-05, + "loss": 0.9576, + "step": 1482 + }, + { + "epoch": 1.1505042668735455, + "grad_norm": 13.346002578735352, + "learning_rate": 2.576671387886163e-05, + "loss": 1.112, + "step": 1483 + }, + { + "epoch": 1.1512800620636152, + "grad_norm": 6.117687702178955, + "learning_rate": 2.5740619738234456e-05, + "loss": 0.2992, + "step": 1484 + }, + { + "epoch": 1.152055857253685, + "grad_norm": 6.321808815002441, + "learning_rate": 2.5714504119247318e-05, + "loss": 0.18, + "step": 1485 + }, + { + "epoch": 1.152831652443755, + "grad_norm": 0.8148465752601624, + "learning_rate": 2.5688367108320412e-05, + "loss": 0.0175, + "step": 1486 + }, + { + "epoch": 1.1536074476338247, + "grad_norm": 8.619049072265625, + "learning_rate": 2.5662208791944738e-05, + "loss": 0.573, + "step": 1487 + }, + { + "epoch": 1.1543832428238945, + "grad_norm": 7.5657453536987305, + "learning_rate": 2.5636029256681784e-05, + "loss": 0.3565, + "step": 1488 + }, + { + "epoch": 1.1551590380139642, + "grad_norm": 2.8809728622436523, + "learning_rate": 2.5609828589163253e-05, + "loss": 0.0861, + "step": 1489 + }, + { + "epoch": 1.1559348332040342, + "grad_norm": 12.367385864257812, + "learning_rate": 2.5583606876090792e-05, + "loss": 0.7502, + "step": 1490 + }, + { + "epoch": 1.156710628394104, + "grad_norm": 6.0324225425720215, + "learning_rate": 2.555736420423568e-05, + "loss": 0.2423, + "step": 1491 + }, + { + "epoch": 1.1574864235841738, + "grad_norm": 10.391775131225586, + "learning_rate": 2.5531100660438562e-05, + "loss": 0.7839, + "step": 1492 + }, + { + "epoch": 1.1582622187742435, + "grad_norm": 5.627542018890381, + "learning_rate": 2.5504816331609127e-05, + "loss": 0.2386, + "step": 1493 + }, + { + "epoch": 1.1590380139643135, + "grad_norm": 0.9938138723373413, + "learning_rate": 2.547851130472587e-05, + "loss": 0.0202, + "step": 1494 + }, + { + "epoch": 1.1598138091543833, + "grad_norm": 8.644150733947754, + "learning_rate": 2.5452185666835767e-05, + "loss": 0.4935, + "step": 1495 + }, + { + "epoch": 1.160589604344453, + "grad_norm": 7.171272277832031, + "learning_rate": 2.5425839505054003e-05, + "loss": 0.5567, + "step": 1496 + }, + { + "epoch": 1.1613653995345228, + "grad_norm": 7.482605457305908, + "learning_rate": 2.539947290656368e-05, + "loss": 0.5883, + "step": 1497 + }, + { + "epoch": 1.1621411947245928, + "grad_norm": 1.2809475660324097, + "learning_rate": 2.5373085958615516e-05, + "loss": 0.0451, + "step": 1498 + }, + { + "epoch": 1.1629169899146625, + "grad_norm": 13.362834930419922, + "learning_rate": 2.5346678748527592e-05, + "loss": 1.0393, + "step": 1499 + }, + { + "epoch": 1.1636927851047323, + "grad_norm": 3.649599552154541, + "learning_rate": 2.5320251363685015e-05, + "loss": 0.0354, + "step": 1500 + }, + { + "epoch": 1.1636927851047323, + "eval_Qnli-dev_cosine_accuracy": 0.6796875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8156468272209167, + "eval_Qnli-dev_cosine_ap": 0.7215111297791443, + "eval_Qnli-dev_cosine_f1": 0.6845425867507886, + "eval_Qnli-dev_cosine_f1_threshold": 0.6985333561897278, + "eval_Qnli-dev_cosine_precision": 0.5452261306532663, + "eval_Qnli-dev_cosine_recall": 0.9194915254237288, + "eval_Qnli-dev_dot_accuracy": 0.66015625, + "eval_Qnli-dev_dot_accuracy_threshold": 382.33575439453125, + "eval_Qnli-dev_dot_ap": 0.6520922502973784, + "eval_Qnli-dev_dot_f1": 0.6689419795221844, + "eval_Qnli-dev_dot_f1_threshold": 340.53900146484375, + "eval_Qnli-dev_dot_precision": 0.56, + "eval_Qnli-dev_dot_recall": 0.8305084745762712, + "eval_Qnli-dev_euclidean_accuracy": 0.6796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.274478912353516, + "eval_Qnli-dev_euclidean_ap": 0.7203306446181923, + "eval_Qnli-dev_euclidean_f1": 0.686084142394822, + "eval_Qnli-dev_euclidean_f1_threshold": 16.68572425842285, + "eval_Qnli-dev_euclidean_precision": 0.5549738219895288, + "eval_Qnli-dev_euclidean_recall": 0.8983050847457628, + "eval_Qnli-dev_manhattan_accuracy": 0.677734375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 251.43592834472656, + "eval_Qnli-dev_manhattan_ap": 0.720124356868738, + "eval_Qnli-dev_manhattan_f1": 0.6843853820598007, + "eval_Qnli-dev_manhattan_f1_threshold": 295.993408203125, + "eval_Qnli-dev_manhattan_precision": 0.5628415300546448, + "eval_Qnli-dev_manhattan_recall": 0.8728813559322034, + "eval_Qnli-dev_max_accuracy": 0.6796875, + "eval_Qnli-dev_max_accuracy_threshold": 382.33575439453125, + "eval_Qnli-dev_max_ap": 0.7215111297791443, + "eval_Qnli-dev_max_f1": 0.686084142394822, + "eval_Qnli-dev_max_f1_threshold": 340.53900146484375, + "eval_Qnli-dev_max_precision": 0.5628415300546448, + "eval_Qnli-dev_max_recall": 0.9194915254237288, + "eval_allNLI-dev_cosine_accuracy": 0.72265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8557730317115784, + "eval_allNLI-dev_cosine_ap": 0.6045129745109319, + "eval_allNLI-dev_cosine_f1": 0.6084905660377359, + "eval_allNLI-dev_cosine_f1_threshold": 0.8040892481803894, + "eval_allNLI-dev_cosine_precision": 0.5139442231075697, + "eval_allNLI-dev_cosine_recall": 0.7456647398843931, + "eval_allNLI-dev_dot_accuracy": 0.69140625, + "eval_allNLI-dev_dot_accuracy_threshold": 411.4944152832031, + "eval_allNLI-dev_dot_ap": 0.5231371322974107, + "eval_allNLI-dev_dot_f1": 0.5867768595041323, + "eval_allNLI-dev_dot_f1_threshold": 337.33502197265625, + "eval_allNLI-dev_dot_precision": 0.4565916398713826, + "eval_allNLI-dev_dot_recall": 0.8208092485549133, + "eval_allNLI-dev_euclidean_accuracy": 0.72265625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.557563781738281, + "eval_allNLI-dev_euclidean_ap": 0.6089537045105125, + "eval_allNLI-dev_euclidean_f1": 0.611336032388664, + "eval_allNLI-dev_euclidean_f1_threshold": 15.133037567138672, + "eval_allNLI-dev_euclidean_precision": 0.470404984423676, + "eval_allNLI-dev_euclidean_recall": 0.8728323699421965, + "eval_allNLI-dev_manhattan_accuracy": 0.734375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 204.24826049804688, + "eval_allNLI-dev_manhattan_ap": 0.613309725046712, + "eval_allNLI-dev_manhattan_f1": 0.6187363834422658, + "eval_allNLI-dev_manhattan_f1_threshold": 258.6210632324219, + "eval_allNLI-dev_manhattan_precision": 0.4965034965034965, + "eval_allNLI-dev_manhattan_recall": 0.8208092485549133, + "eval_allNLI-dev_max_accuracy": 0.734375, + "eval_allNLI-dev_max_accuracy_threshold": 411.4944152832031, + "eval_allNLI-dev_max_ap": 0.613309725046712, + "eval_allNLI-dev_max_f1": 0.6187363834422658, + "eval_allNLI-dev_max_f1_threshold": 337.33502197265625, + "eval_allNLI-dev_max_precision": 0.5139442231075697, + "eval_allNLI-dev_max_recall": 0.8728323699421965, + "eval_sequential_score": 0.7215111297791443, + "eval_sts-test_pearson_cosine": 0.8336829975221669, + "eval_sts-test_pearson_dot": 0.80471416625464, + "eval_sts-test_pearson_euclidean": 0.8684453347367653, + "eval_sts-test_pearson_manhattan": 0.865929842678619, + "eval_sts-test_pearson_max": 0.8684453347367653, + "eval_sts-test_spearman_cosine": 0.8728263198635252, + "eval_sts-test_spearman_dot": 0.8026715574306836, + "eval_sts-test_spearman_euclidean": 0.8682778464069646, + "eval_sts-test_spearman_manhattan": 0.8666767805281644, + "eval_sts-test_spearman_max": 0.8728263198635252, + "eval_vitaminc-pairs_loss": 3.5097579956054688, + "eval_vitaminc-pairs_runtime": 5.5283, + "eval_vitaminc-pairs_samples_per_second": 23.153, + "eval_vitaminc-pairs_steps_per_second": 0.362, + "step": 1500 + }, + { + "epoch": 1.1636927851047323, + "eval_negation-triplets_loss": 1.6857094764709473, + "eval_negation-triplets_runtime": 1.0933, + "eval_negation-triplets_samples_per_second": 117.074, + "eval_negation-triplets_steps_per_second": 1.829, + "step": 1500 + }, + { + "epoch": 1.1636927851047323, + "eval_scitail-pairs-pos_loss": 0.13186798989772797, + "eval_scitail-pairs-pos_runtime": 1.3278, + "eval_scitail-pairs-pos_samples_per_second": 96.399, + "eval_scitail-pairs-pos_steps_per_second": 1.506, + "step": 1500 + }, + { + "epoch": 1.1636927851047323, + "eval_scitail-pairs-qa_loss": 0.007198065519332886, + "eval_scitail-pairs-qa_runtime": 0.7895, + "eval_scitail-pairs-qa_samples_per_second": 162.133, + "eval_scitail-pairs-qa_steps_per_second": 2.533, + "step": 1500 + }, + { + "epoch": 1.1636927851047323, + "eval_xsum-pairs_loss": 0.1924620121717453, + "eval_xsum-pairs_runtime": 3.946, + "eval_xsum-pairs_samples_per_second": 32.438, + "eval_xsum-pairs_steps_per_second": 0.507, + "step": 1500 + }, + { + "epoch": 1.1636927851047323, + "eval_sciq_pairs_loss": 0.057192038744688034, + "eval_sciq_pairs_runtime": 6.0852, + "eval_sciq_pairs_samples_per_second": 21.035, + "eval_sciq_pairs_steps_per_second": 0.329, + "step": 1500 + }, + { + "epoch": 1.1636927851047323, + "eval_qasc_pairs_loss": 0.26513028144836426, + "eval_qasc_pairs_runtime": 0.9464, + "eval_qasc_pairs_samples_per_second": 135.245, + "eval_qasc_pairs_steps_per_second": 2.113, + "step": 1500 + }, + { + "epoch": 1.1636927851047323, + "eval_openbookqa_pairs_loss": 1.0588297843933105, + "eval_openbookqa_pairs_runtime": 0.8788, + "eval_openbookqa_pairs_samples_per_second": 145.661, + "eval_openbookqa_pairs_steps_per_second": 2.276, + "step": 1500 + }, + { + "epoch": 1.1636927851047323, + "eval_msmarco_pairs_loss": 0.48213279247283936, + "eval_msmarco_pairs_runtime": 1.9895, + "eval_msmarco_pairs_samples_per_second": 64.338, + "eval_msmarco_pairs_steps_per_second": 1.005, + "step": 1500 + }, + { + "epoch": 1.1636927851047323, + "eval_nq_pairs_loss": 0.6814334392547607, + "eval_nq_pairs_runtime": 3.3543, + "eval_nq_pairs_samples_per_second": 38.16, + "eval_nq_pairs_steps_per_second": 0.596, + "step": 1500 + }, + { + "epoch": 1.1636927851047323, + "eval_trivia_pairs_loss": 0.8786530494689941, + "eval_trivia_pairs_runtime": 4.4996, + "eval_trivia_pairs_samples_per_second": 28.447, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 1500 + }, + { + "epoch": 1.1636927851047323, + "eval_gooaq_pairs_loss": 0.6245259046554565, + "eval_gooaq_pairs_runtime": 1.4807, + "eval_gooaq_pairs_samples_per_second": 86.447, + "eval_gooaq_pairs_steps_per_second": 1.351, + "step": 1500 + }, + { + "epoch": 1.1636927851047323, + "eval_paws-pos_loss": 0.032582592219114304, + "eval_paws-pos_runtime": 1.0467, + "eval_paws-pos_samples_per_second": 122.288, + "eval_paws-pos_steps_per_second": 1.911, + "step": 1500 + }, + { + "epoch": 1.1636927851047323, + "eval_global_dataset_loss": 0.4620685577392578, + "eval_global_dataset_runtime": 19.4955, + "eval_global_dataset_samples_per_second": 21.338, + "eval_global_dataset_steps_per_second": 0.359, + "step": 1500 + }, + { + "epoch": 1.1644685802948023, + "grad_norm": 1.1272958517074585, + "learning_rate": 2.529380389153966e-05, + "loss": 0.0372, + "step": 1501 + }, + { + "epoch": 1.165244375484872, + "grad_norm": 6.593901634216309, + "learning_rate": 2.5267336419609883e-05, + "loss": 0.3066, + "step": 1502 + }, + { + "epoch": 1.1660201706749418, + "grad_norm": 0.2677835524082184, + "learning_rate": 2.5240849035480216e-05, + "loss": 0.0036, + "step": 1503 + }, + { + "epoch": 1.1667959658650116, + "grad_norm": 9.186358451843262, + "learning_rate": 2.5214341826801076e-05, + "loss": 0.9254, + "step": 1504 + }, + { + "epoch": 1.1675717610550815, + "grad_norm": 7.8784356117248535, + "learning_rate": 2.5187814881288496e-05, + "loss": 0.4806, + "step": 1505 + }, + { + "epoch": 1.1683475562451513, + "grad_norm": 11.509079933166504, + "learning_rate": 2.5161268286723812e-05, + "loss": 0.9052, + "step": 1506 + }, + { + "epoch": 1.169123351435221, + "grad_norm": 9.891919136047363, + "learning_rate": 2.5134702130953385e-05, + "loss": 0.6403, + "step": 1507 + }, + { + "epoch": 1.169899146625291, + "grad_norm": 5.768681526184082, + "learning_rate": 2.51081165018883e-05, + "loss": 0.1765, + "step": 1508 + }, + { + "epoch": 1.1706749418153608, + "grad_norm": 0.4727028012275696, + "learning_rate": 2.5081511487504087e-05, + "loss": 0.0081, + "step": 1509 + }, + { + "epoch": 1.1714507370054306, + "grad_norm": 6.101867198944092, + "learning_rate": 2.505488717584043e-05, + "loss": 0.2318, + "step": 1510 + }, + { + "epoch": 1.1722265321955003, + "grad_norm": 7.482995986938477, + "learning_rate": 2.5028243655000864e-05, + "loss": 0.6451, + "step": 1511 + }, + { + "epoch": 1.17300232738557, + "grad_norm": 6.727688789367676, + "learning_rate": 2.5001581013152486e-05, + "loss": 0.3813, + "step": 1512 + }, + { + "epoch": 1.17377812257564, + "grad_norm": 6.3394060134887695, + "learning_rate": 2.497489933852568e-05, + "loss": 0.3024, + "step": 1513 + }, + { + "epoch": 1.1745539177657098, + "grad_norm": 9.04686450958252, + "learning_rate": 2.494819871941379e-05, + "loss": 0.7217, + "step": 1514 + }, + { + "epoch": 1.1753297129557796, + "grad_norm": 10.873148918151855, + "learning_rate": 2.492147924417288e-05, + "loss": 1.0036, + "step": 1515 + }, + { + "epoch": 1.1761055081458496, + "grad_norm": 9.66885757446289, + "learning_rate": 2.4894741001221387e-05, + "loss": 0.748, + "step": 1516 + }, + { + "epoch": 1.1768813033359193, + "grad_norm": 13.938639640808105, + "learning_rate": 2.4867984079039862e-05, + "loss": 2.2849, + "step": 1517 + }, + { + "epoch": 1.177657098525989, + "grad_norm": 6.365065574645996, + "learning_rate": 2.484120856617067e-05, + "loss": 0.5299, + "step": 1518 + }, + { + "epoch": 1.1784328937160589, + "grad_norm": 0.2980465590953827, + "learning_rate": 2.4814414551217694e-05, + "loss": 0.0054, + "step": 1519 + }, + { + "epoch": 1.1792086889061288, + "grad_norm": 7.215659141540527, + "learning_rate": 2.4787602122846034e-05, + "loss": 0.3577, + "step": 1520 + }, + { + "epoch": 1.1792086889061288, + "eval_Qnli-dev_cosine_accuracy": 0.689453125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8502223491668701, + "eval_Qnli-dev_cosine_ap": 0.7232187179234506, + "eval_Qnli-dev_cosine_f1": 0.6872852233676976, + "eval_Qnli-dev_cosine_f1_threshold": 0.7783228158950806, + "eval_Qnli-dev_cosine_precision": 0.5780346820809249, + "eval_Qnli-dev_cosine_recall": 0.847457627118644, + "eval_Qnli-dev_dot_accuracy": 0.65625, + "eval_Qnli-dev_dot_accuracy_threshold": 424.6481018066406, + "eval_Qnli-dev_dot_ap": 0.6362800978462658, + "eval_Qnli-dev_dot_f1": 0.6718750000000001, + "eval_Qnli-dev_dot_f1_threshold": 345.9131774902344, + "eval_Qnli-dev_dot_precision": 0.5321782178217822, + "eval_Qnli-dev_dot_recall": 0.9110169491525424, + "eval_Qnli-dev_euclidean_accuracy": 0.6953125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.58099365234375, + "eval_Qnli-dev_euclidean_ap": 0.72091265090783, + "eval_Qnli-dev_euclidean_f1": 0.6823956442831216, + "eval_Qnli-dev_euclidean_f1_threshold": 14.020771026611328, + "eval_Qnli-dev_euclidean_precision": 0.5968253968253968, + "eval_Qnli-dev_euclidean_recall": 0.7966101694915254, + "eval_Qnli-dev_manhattan_accuracy": 0.677734375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 215.9249267578125, + "eval_Qnli-dev_manhattan_ap": 0.7227062394443693, + "eval_Qnli-dev_manhattan_f1": 0.686832740213523, + "eval_Qnli-dev_manhattan_f1_threshold": 257.152099609375, + "eval_Qnli-dev_manhattan_precision": 0.5920245398773006, + "eval_Qnli-dev_manhattan_recall": 0.8177966101694916, + "eval_Qnli-dev_max_accuracy": 0.6953125, + "eval_Qnli-dev_max_accuracy_threshold": 424.6481018066406, + "eval_Qnli-dev_max_ap": 0.7232187179234506, + "eval_Qnli-dev_max_f1": 0.6872852233676976, + "eval_Qnli-dev_max_f1_threshold": 345.9131774902344, + "eval_Qnli-dev_max_precision": 0.5968253968253968, + "eval_Qnli-dev_max_recall": 0.9110169491525424, + "eval_allNLI-dev_cosine_accuracy": 0.72265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9029145240783691, + "eval_allNLI-dev_cosine_ap": 0.6158450221982292, + "eval_allNLI-dev_cosine_f1": 0.6288659793814432, + "eval_allNLI-dev_cosine_f1_threshold": 0.8434971570968628, + "eval_allNLI-dev_cosine_precision": 0.5674418604651162, + "eval_allNLI-dev_cosine_recall": 0.7052023121387283, + "eval_allNLI-dev_dot_accuracy": 0.685546875, + "eval_allNLI-dev_dot_accuracy_threshold": 423.7976379394531, + "eval_allNLI-dev_dot_ap": 0.5218741390384131, + "eval_allNLI-dev_dot_f1": 0.5783582089552239, + "eval_allNLI-dev_dot_f1_threshold": 351.46209716796875, + "eval_allNLI-dev_dot_precision": 0.42699724517906334, + "eval_allNLI-dev_dot_recall": 0.8959537572254336, + "eval_allNLI-dev_euclidean_accuracy": 0.728515625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.645256042480469, + "eval_allNLI-dev_euclidean_ap": 0.6200723503440493, + "eval_allNLI-dev_euclidean_f1": 0.6219512195121951, + "eval_allNLI-dev_euclidean_f1_threshold": 14.377790451049805, + "eval_allNLI-dev_euclidean_precision": 0.47962382445141066, + "eval_allNLI-dev_euclidean_recall": 0.884393063583815, + "eval_allNLI-dev_manhattan_accuracy": 0.73046875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 202.22116088867188, + "eval_allNLI-dev_manhattan_ap": 0.619739343485595, + "eval_allNLI-dev_manhattan_f1": 0.6338028169014085, + "eval_allNLI-dev_manhattan_f1_threshold": 234.70828247070312, + "eval_allNLI-dev_manhattan_precision": 0.5335968379446641, + "eval_allNLI-dev_manhattan_recall": 0.7803468208092486, + "eval_allNLI-dev_max_accuracy": 0.73046875, + "eval_allNLI-dev_max_accuracy_threshold": 423.7976379394531, + "eval_allNLI-dev_max_ap": 0.6200723503440493, + "eval_allNLI-dev_max_f1": 0.6338028169014085, + "eval_allNLI-dev_max_f1_threshold": 351.46209716796875, + "eval_allNLI-dev_max_precision": 0.5674418604651162, + "eval_allNLI-dev_max_recall": 0.8959537572254336, + "eval_sequential_score": 0.7232187179234506, + "eval_sts-test_pearson_cosine": 0.8257327226654304, + "eval_sts-test_pearson_dot": 0.775742733669371, + "eval_sts-test_pearson_euclidean": 0.864829908276067, + "eval_sts-test_pearson_manhattan": 0.8630152077980378, + "eval_sts-test_pearson_max": 0.864829908276067, + "eval_sts-test_spearman_cosine": 0.8697459115369096, + "eval_sts-test_spearman_dot": 0.773258746129925, + "eval_sts-test_spearman_euclidean": 0.8676068459936622, + "eval_sts-test_spearman_manhattan": 0.8659502885507366, + "eval_sts-test_spearman_max": 0.8697459115369096, + "eval_vitaminc-pairs_loss": 3.5805532932281494, + "eval_vitaminc-pairs_runtime": 5.5397, + "eval_vitaminc-pairs_samples_per_second": 23.106, + "eval_vitaminc-pairs_steps_per_second": 0.361, + "step": 1520 + }, + { + "epoch": 1.1792086889061288, + "eval_negation-triplets_loss": 1.6359732151031494, + "eval_negation-triplets_runtime": 1.1011, + "eval_negation-triplets_samples_per_second": 116.249, + "eval_negation-triplets_steps_per_second": 1.816, + "step": 1520 + }, + { + "epoch": 1.1792086889061288, + "eval_scitail-pairs-pos_loss": 0.14347651600837708, + "eval_scitail-pairs-pos_runtime": 1.3435, + "eval_scitail-pairs-pos_samples_per_second": 95.271, + "eval_scitail-pairs-pos_steps_per_second": 1.489, + "step": 1520 + }, + { + "epoch": 1.1792086889061288, + "eval_scitail-pairs-qa_loss": 0.009301604703068733, + "eval_scitail-pairs-qa_runtime": 0.7927, + "eval_scitail-pairs-qa_samples_per_second": 161.48, + "eval_scitail-pairs-qa_steps_per_second": 2.523, + "step": 1520 + }, + { + "epoch": 1.1792086889061288, + "eval_xsum-pairs_loss": 0.14715328812599182, + "eval_xsum-pairs_runtime": 3.9503, + "eval_xsum-pairs_samples_per_second": 32.403, + "eval_xsum-pairs_steps_per_second": 0.506, + "step": 1520 + }, + { + "epoch": 1.1792086889061288, + "eval_sciq_pairs_loss": 0.04745934158563614, + "eval_sciq_pairs_runtime": 6.1999, + "eval_sciq_pairs_samples_per_second": 20.646, + "eval_sciq_pairs_steps_per_second": 0.323, + "step": 1520 + }, + { + "epoch": 1.1792086889061288, + "eval_qasc_pairs_loss": 0.24222904443740845, + "eval_qasc_pairs_runtime": 0.9763, + "eval_qasc_pairs_samples_per_second": 131.111, + "eval_qasc_pairs_steps_per_second": 2.049, + "step": 1520 + }, + { + "epoch": 1.1792086889061288, + "eval_openbookqa_pairs_loss": 1.0575793981552124, + "eval_openbookqa_pairs_runtime": 0.8926, + "eval_openbookqa_pairs_samples_per_second": 143.408, + "eval_openbookqa_pairs_steps_per_second": 2.241, + "step": 1520 + }, + { + "epoch": 1.1792086889061288, + "eval_msmarco_pairs_loss": 0.4181935787200928, + "eval_msmarco_pairs_runtime": 1.9828, + "eval_msmarco_pairs_samples_per_second": 64.555, + "eval_msmarco_pairs_steps_per_second": 1.009, + "step": 1520 + }, + { + "epoch": 1.1792086889061288, + "eval_nq_pairs_loss": 0.6532794833183289, + "eval_nq_pairs_runtime": 3.3558, + "eval_nq_pairs_samples_per_second": 38.143, + "eval_nq_pairs_steps_per_second": 0.596, + "step": 1520 + }, + { + "epoch": 1.1792086889061288, + "eval_trivia_pairs_loss": 0.7690937519073486, + "eval_trivia_pairs_runtime": 4.5109, + "eval_trivia_pairs_samples_per_second": 28.376, + "eval_trivia_pairs_steps_per_second": 0.443, + "step": 1520 + }, + { + "epoch": 1.1792086889061288, + "eval_gooaq_pairs_loss": 0.6132965087890625, + "eval_gooaq_pairs_runtime": 1.4786, + "eval_gooaq_pairs_samples_per_second": 86.566, + "eval_gooaq_pairs_steps_per_second": 1.353, + "step": 1520 + }, + { + "epoch": 1.1792086889061288, + "eval_paws-pos_loss": 0.03435618802905083, + "eval_paws-pos_runtime": 1.0657, + "eval_paws-pos_samples_per_second": 120.111, + "eval_paws-pos_steps_per_second": 1.877, + "step": 1520 + }, + { + "epoch": 1.1792086889061288, + "eval_global_dataset_loss": 0.41063812375068665, + "eval_global_dataset_runtime": 19.5142, + "eval_global_dataset_samples_per_second": 21.318, + "eval_global_dataset_steps_per_second": 0.359, + "step": 1520 + }, + { + "epoch": 1.1799844840961986, + "grad_norm": 8.976807594299316, + "learning_rate": 2.4760771369781736e-05, + "loss": 0.6189, + "step": 1521 + }, + { + "epoch": 1.1807602792862684, + "grad_norm": 6.717977046966553, + "learning_rate": 2.4733922380811482e-05, + "loss": 0.3631, + "step": 1522 + }, + { + "epoch": 1.1815360744763383, + "grad_norm": 7.200489044189453, + "learning_rate": 2.4707055244782286e-05, + "loss": 0.3755, + "step": 1523 + }, + { + "epoch": 1.182311869666408, + "grad_norm": 4.432051658630371, + "learning_rate": 2.4680170050601236e-05, + "loss": 0.132, + "step": 1524 + }, + { + "epoch": 1.1830876648564779, + "grad_norm": 7.068164348602295, + "learning_rate": 2.4653266887235154e-05, + "loss": 0.6194, + "step": 1525 + }, + { + "epoch": 1.1838634600465476, + "grad_norm": 6.1961822509765625, + "learning_rate": 2.462634584371034e-05, + "loss": 0.2404, + "step": 1526 + }, + { + "epoch": 1.1846392552366176, + "grad_norm": 0.18275730311870575, + "learning_rate": 2.4599407009112258e-05, + "loss": 0.0034, + "step": 1527 + }, + { + "epoch": 1.1854150504266874, + "grad_norm": 5.256531238555908, + "learning_rate": 2.457245047258524e-05, + "loss": 0.239, + "step": 1528 + }, + { + "epoch": 1.1861908456167571, + "grad_norm": 0.22063124179840088, + "learning_rate": 2.4545476323332203e-05, + "loss": 0.0041, + "step": 1529 + }, + { + "epoch": 1.1869666408068271, + "grad_norm": 8.477676391601562, + "learning_rate": 2.4518484650614347e-05, + "loss": 0.6449, + "step": 1530 + }, + { + "epoch": 1.1877424359968969, + "grad_norm": 6.831202507019043, + "learning_rate": 2.4491475543750848e-05, + "loss": 0.445, + "step": 1531 + }, + { + "epoch": 1.1885182311869666, + "grad_norm": 5.1271562576293945, + "learning_rate": 2.446444909211859e-05, + "loss": 0.216, + "step": 1532 + }, + { + "epoch": 1.1892940263770364, + "grad_norm": 7.095730304718018, + "learning_rate": 2.4437405385151845e-05, + "loss": 0.3011, + "step": 1533 + }, + { + "epoch": 1.1900698215671062, + "grad_norm": 0.38344866037368774, + "learning_rate": 2.441034451234198e-05, + "loss": 0.0109, + "step": 1534 + }, + { + "epoch": 1.1908456167571762, + "grad_norm": 6.971592426300049, + "learning_rate": 2.4383266563237182e-05, + "loss": 0.2959, + "step": 1535 + }, + { + "epoch": 1.191621411947246, + "grad_norm": 3.200517177581787, + "learning_rate": 2.435617162744213e-05, + "loss": 0.072, + "step": 1536 + }, + { + "epoch": 1.1923972071373157, + "grad_norm": 6.572030067443848, + "learning_rate": 2.4329059794617717e-05, + "loss": 0.3534, + "step": 1537 + }, + { + "epoch": 1.1931730023273857, + "grad_norm": 6.895688056945801, + "learning_rate": 2.430193115448076e-05, + "loss": 0.4648, + "step": 1538 + }, + { + "epoch": 1.1939487975174554, + "grad_norm": 12.778672218322754, + "learning_rate": 2.4274785796803693e-05, + "loss": 0.8411, + "step": 1539 + }, + { + "epoch": 1.1947245927075252, + "grad_norm": 7.5648884773254395, + "learning_rate": 2.424762381141425e-05, + "loss": 0.4505, + "step": 1540 + }, + { + "epoch": 1.1947245927075252, + "eval_Qnli-dev_cosine_accuracy": 0.681640625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8129110336303711, + "eval_Qnli-dev_cosine_ap": 0.7179927178188497, + "eval_Qnli-dev_cosine_f1": 0.6751188589540412, + "eval_Qnli-dev_cosine_f1_threshold": 0.7016874551773071, + "eval_Qnli-dev_cosine_precision": 0.5392405063291139, + "eval_Qnli-dev_cosine_recall": 0.902542372881356, + "eval_Qnli-dev_dot_accuracy": 0.6640625, + "eval_Qnli-dev_dot_accuracy_threshold": 400.9525146484375, + "eval_Qnli-dev_dot_ap": 0.6510286030043275, + "eval_Qnli-dev_dot_f1": 0.675968992248062, + "eval_Qnli-dev_dot_f1_threshold": 317.58575439453125, + "eval_Qnli-dev_dot_precision": 0.5330073349633252, + "eval_Qnli-dev_dot_recall": 0.923728813559322, + "eval_Qnli-dev_euclidean_accuracy": 0.677734375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.371767044067383, + "eval_Qnli-dev_euclidean_ap": 0.7165828087172217, + "eval_Qnli-dev_euclidean_f1": 0.6794258373205743, + "eval_Qnli-dev_euclidean_f1_threshold": 16.988174438476562, + "eval_Qnli-dev_euclidean_precision": 0.5447570332480819, + "eval_Qnli-dev_euclidean_recall": 0.902542372881356, + "eval_Qnli-dev_manhattan_accuracy": 0.673828125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 221.26708984375, + "eval_Qnli-dev_manhattan_ap": 0.7184904772518774, + "eval_Qnli-dev_manhattan_f1": 0.6775244299674267, + "eval_Qnli-dev_manhattan_f1_threshold": 304.7618408203125, + "eval_Qnli-dev_manhattan_precision": 0.5502645502645502, + "eval_Qnli-dev_manhattan_recall": 0.8813559322033898, + "eval_Qnli-dev_max_accuracy": 0.681640625, + "eval_Qnli-dev_max_accuracy_threshold": 400.9525146484375, + "eval_Qnli-dev_max_ap": 0.7184904772518774, + "eval_Qnli-dev_max_f1": 0.6794258373205743, + "eval_Qnli-dev_max_f1_threshold": 317.58575439453125, + "eval_Qnli-dev_max_precision": 0.5502645502645502, + "eval_Qnli-dev_max_recall": 0.923728813559322, + "eval_allNLI-dev_cosine_accuracy": 0.72265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8484312295913696, + "eval_allNLI-dev_cosine_ap": 0.6156221836498927, + "eval_allNLI-dev_cosine_f1": 0.6206896551724138, + "eval_allNLI-dev_cosine_f1_threshold": 0.7896684408187866, + "eval_allNLI-dev_cosine_precision": 0.5152671755725191, + "eval_allNLI-dev_cosine_recall": 0.7803468208092486, + "eval_allNLI-dev_dot_accuracy": 0.685546875, + "eval_allNLI-dev_dot_accuracy_threshold": 413.1653747558594, + "eval_allNLI-dev_dot_ap": 0.5179773656749745, + "eval_allNLI-dev_dot_f1": 0.5780590717299579, + "eval_allNLI-dev_dot_f1_threshold": 346.2738037109375, + "eval_allNLI-dev_dot_precision": 0.45514950166112955, + "eval_allNLI-dev_dot_recall": 0.791907514450867, + "eval_allNLI-dev_euclidean_accuracy": 0.728515625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 12.101924896240234, + "eval_allNLI-dev_euclidean_ap": 0.6183578354407625, + "eval_allNLI-dev_euclidean_f1": 0.6282722513089005, + "eval_allNLI-dev_euclidean_f1_threshold": 12.668548583984375, + "eval_allNLI-dev_euclidean_precision": 0.5741626794258373, + "eval_allNLI-dev_euclidean_recall": 0.6936416184971098, + "eval_allNLI-dev_manhattan_accuracy": 0.73046875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 210.4146728515625, + "eval_allNLI-dev_manhattan_ap": 0.621398722427615, + "eval_allNLI-dev_manhattan_f1": 0.632603406326034, + "eval_allNLI-dev_manhattan_f1_threshold": 241.35394287109375, + "eval_allNLI-dev_manhattan_precision": 0.5462184873949579, + "eval_allNLI-dev_manhattan_recall": 0.7514450867052023, + "eval_allNLI-dev_max_accuracy": 0.73046875, + "eval_allNLI-dev_max_accuracy_threshold": 413.1653747558594, + "eval_allNLI-dev_max_ap": 0.621398722427615, + "eval_allNLI-dev_max_f1": 0.632603406326034, + "eval_allNLI-dev_max_f1_threshold": 346.2738037109375, + "eval_allNLI-dev_max_precision": 0.5741626794258373, + "eval_allNLI-dev_max_recall": 0.791907514450867, + "eval_sequential_score": 0.7184904772518774, + "eval_sts-test_pearson_cosine": 0.8287358545291583, + "eval_sts-test_pearson_dot": 0.8043536787086222, + "eval_sts-test_pearson_euclidean": 0.8620388347133311, + "eval_sts-test_pearson_manhattan": 0.8594997340383432, + "eval_sts-test_pearson_max": 0.8620388347133311, + "eval_sts-test_spearman_cosine": 0.8684227509913501, + "eval_sts-test_spearman_dot": 0.8001078024168318, + "eval_sts-test_spearman_euclidean": 0.8641730369558385, + "eval_sts-test_spearman_manhattan": 0.8616584188269943, + "eval_sts-test_spearman_max": 0.8684227509913501, + "eval_vitaminc-pairs_loss": 3.519922971725464, + "eval_vitaminc-pairs_runtime": 5.5492, + "eval_vitaminc-pairs_samples_per_second": 23.066, + "eval_vitaminc-pairs_steps_per_second": 0.36, + "step": 1540 + }, + { + "epoch": 1.1947245927075252, + "eval_negation-triplets_loss": 1.6979248523712158, + "eval_negation-triplets_runtime": 1.0971, + "eval_negation-triplets_samples_per_second": 116.671, + "eval_negation-triplets_steps_per_second": 1.823, + "step": 1540 + }, + { + "epoch": 1.1947245927075252, + "eval_scitail-pairs-pos_loss": 0.148373544216156, + "eval_scitail-pairs-pos_runtime": 1.3255, + "eval_scitail-pairs-pos_samples_per_second": 96.568, + "eval_scitail-pairs-pos_steps_per_second": 1.509, + "step": 1540 + }, + { + "epoch": 1.1947245927075252, + "eval_scitail-pairs-qa_loss": 0.01045343279838562, + "eval_scitail-pairs-qa_runtime": 0.8073, + "eval_scitail-pairs-qa_samples_per_second": 158.56, + "eval_scitail-pairs-qa_steps_per_second": 2.478, + "step": 1540 + }, + { + "epoch": 1.1947245927075252, + "eval_xsum-pairs_loss": 0.13372820615768433, + "eval_xsum-pairs_runtime": 3.9479, + "eval_xsum-pairs_samples_per_second": 32.423, + "eval_xsum-pairs_steps_per_second": 0.507, + "step": 1540 + }, + { + "epoch": 1.1947245927075252, + "eval_sciq_pairs_loss": 0.046155717223882675, + "eval_sciq_pairs_runtime": 6.1736, + "eval_sciq_pairs_samples_per_second": 20.733, + "eval_sciq_pairs_steps_per_second": 0.324, + "step": 1540 + }, + { + "epoch": 1.1947245927075252, + "eval_qasc_pairs_loss": 0.2676868438720703, + "eval_qasc_pairs_runtime": 0.9376, + "eval_qasc_pairs_samples_per_second": 136.512, + "eval_qasc_pairs_steps_per_second": 2.133, + "step": 1540 + }, + { + "epoch": 1.1947245927075252, + "eval_openbookqa_pairs_loss": 1.0789381265640259, + "eval_openbookqa_pairs_runtime": 0.8926, + "eval_openbookqa_pairs_samples_per_second": 143.394, + "eval_openbookqa_pairs_steps_per_second": 2.241, + "step": 1540 + }, + { + "epoch": 1.1947245927075252, + "eval_msmarco_pairs_loss": 0.36950698494911194, + "eval_msmarco_pairs_runtime": 1.9931, + "eval_msmarco_pairs_samples_per_second": 64.221, + "eval_msmarco_pairs_steps_per_second": 1.003, + "step": 1540 + }, + { + "epoch": 1.1947245927075252, + "eval_nq_pairs_loss": 0.7016381621360779, + "eval_nq_pairs_runtime": 3.364, + "eval_nq_pairs_samples_per_second": 38.049, + "eval_nq_pairs_steps_per_second": 0.595, + "step": 1540 + }, + { + "epoch": 1.1947245927075252, + "eval_trivia_pairs_loss": 0.8353316783905029, + "eval_trivia_pairs_runtime": 4.5048, + "eval_trivia_pairs_samples_per_second": 28.414, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 1540 + }, + { + "epoch": 1.1947245927075252, + "eval_gooaq_pairs_loss": 0.6751132011413574, + "eval_gooaq_pairs_runtime": 1.4892, + "eval_gooaq_pairs_samples_per_second": 85.952, + "eval_gooaq_pairs_steps_per_second": 1.343, + "step": 1540 + }, + { + "epoch": 1.1947245927075252, + "eval_paws-pos_loss": 0.03093648888170719, + "eval_paws-pos_runtime": 1.0541, + "eval_paws-pos_samples_per_second": 121.426, + "eval_paws-pos_steps_per_second": 1.897, + "step": 1540 + }, + { + "epoch": 1.1947245927075252, + "eval_global_dataset_loss": 0.40634652972221375, + "eval_global_dataset_runtime": 19.4825, + "eval_global_dataset_samples_per_second": 21.353, + "eval_global_dataset_steps_per_second": 0.359, + "step": 1540 + }, + { + "epoch": 1.195500387897595, + "grad_norm": 8.158822059631348, + "learning_rate": 2.4220445288195218e-05, + "loss": 0.4777, + "step": 1541 + }, + { + "epoch": 1.196276183087665, + "grad_norm": 3.8686323165893555, + "learning_rate": 2.4193250317084088e-05, + "loss": 0.2208, + "step": 1542 + }, + { + "epoch": 1.1970519782777347, + "grad_norm": 14.706791877746582, + "learning_rate": 2.4166038988072792e-05, + "loss": 1.8666, + "step": 1543 + }, + { + "epoch": 1.1978277734678044, + "grad_norm": 7.085853099822998, + "learning_rate": 2.413881139120738e-05, + "loss": 0.2432, + "step": 1544 + }, + { + "epoch": 1.1986035686578744, + "grad_norm": 6.740437984466553, + "learning_rate": 2.4111567616587752e-05, + "loss": 0.4298, + "step": 1545 + }, + { + "epoch": 1.1993793638479442, + "grad_norm": 9.433274269104004, + "learning_rate": 2.4084307754367334e-05, + "loss": 0.7261, + "step": 1546 + }, + { + "epoch": 1.200155159038014, + "grad_norm": 1.2580301761627197, + "learning_rate": 2.4057031894752766e-05, + "loss": 0.0249, + "step": 1547 + }, + { + "epoch": 1.2009309542280837, + "grad_norm": 5.576623439788818, + "learning_rate": 2.4029740128003674e-05, + "loss": 0.1261, + "step": 1548 + }, + { + "epoch": 1.2017067494181537, + "grad_norm": 12.782208442687988, + "learning_rate": 2.4002432544432284e-05, + "loss": 0.8455, + "step": 1549 + }, + { + "epoch": 1.2024825446082235, + "grad_norm": 7.493494033813477, + "learning_rate": 2.3975109234403163e-05, + "loss": 0.2049, + "step": 1550 + }, + { + "epoch": 1.2032583397982932, + "grad_norm": 5.4548211097717285, + "learning_rate": 2.3947770288332945e-05, + "loss": 0.1597, + "step": 1551 + }, + { + "epoch": 1.204034134988363, + "grad_norm": 13.153420448303223, + "learning_rate": 2.3920415796689988e-05, + "loss": 1.1437, + "step": 1552 + }, + { + "epoch": 1.204809930178433, + "grad_norm": 7.789492607116699, + "learning_rate": 2.389304584999408e-05, + "loss": 0.2887, + "step": 1553 + }, + { + "epoch": 1.2055857253685027, + "grad_norm": 9.63109016418457, + "learning_rate": 2.3865660538816186e-05, + "loss": 0.6308, + "step": 1554 + }, + { + "epoch": 1.2063615205585725, + "grad_norm": 10.215594291687012, + "learning_rate": 2.383825995377808e-05, + "loss": 0.5854, + "step": 1555 + }, + { + "epoch": 1.2071373157486422, + "grad_norm": 4.493361473083496, + "learning_rate": 2.3810844185552104e-05, + "loss": 0.1006, + "step": 1556 + }, + { + "epoch": 1.2079131109387122, + "grad_norm": 9.264915466308594, + "learning_rate": 2.3783413324860834e-05, + "loss": 0.5408, + "step": 1557 + }, + { + "epoch": 1.208688906128782, + "grad_norm": 0.4795230031013489, + "learning_rate": 2.3755967462476774e-05, + "loss": 0.0179, + "step": 1558 + }, + { + "epoch": 1.2094647013188518, + "grad_norm": 10.787131309509277, + "learning_rate": 2.3728506689222105e-05, + "loss": 0.7118, + "step": 1559 + }, + { + "epoch": 1.2102404965089217, + "grad_norm": 7.773983478546143, + "learning_rate": 2.370103109596832e-05, + "loss": 0.3176, + "step": 1560 + }, + { + "epoch": 1.2102404965089217, + "eval_Qnli-dev_cosine_accuracy": 0.681640625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8107146620750427, + "eval_Qnli-dev_cosine_ap": 0.7137156711469417, + "eval_Qnli-dev_cosine_f1": 0.68824531516184, + "eval_Qnli-dev_cosine_f1_threshold": 0.7436014413833618, + "eval_Qnli-dev_cosine_precision": 0.5754985754985755, + "eval_Qnli-dev_cosine_recall": 0.8559322033898306, + "eval_Qnli-dev_dot_accuracy": 0.6640625, + "eval_Qnli-dev_dot_accuracy_threshold": 388.3597717285156, + "eval_Qnli-dev_dot_ap": 0.6282772573604471, + "eval_Qnli-dev_dot_f1": 0.6707692307692309, + "eval_Qnli-dev_dot_f1_threshold": 319.99456787109375, + "eval_Qnli-dev_dot_precision": 0.5265700483091788, + "eval_Qnli-dev_dot_recall": 0.923728813559322, + "eval_Qnli-dev_euclidean_accuracy": 0.6796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.618818283081055, + "eval_Qnli-dev_euclidean_ap": 0.7120975703737914, + "eval_Qnli-dev_euclidean_f1": 0.6761565836298933, + "eval_Qnli-dev_euclidean_f1_threshold": 15.161880493164062, + "eval_Qnli-dev_euclidean_precision": 0.5828220858895705, + "eval_Qnli-dev_euclidean_recall": 0.8050847457627118, + "eval_Qnli-dev_manhattan_accuracy": 0.68359375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 227.99832153320312, + "eval_Qnli-dev_manhattan_ap": 0.7167196607222883, + "eval_Qnli-dev_manhattan_f1": 0.6799999999999999, + "eval_Qnli-dev_manhattan_f1_threshold": 268.1365661621094, + "eval_Qnli-dev_manhattan_precision": 0.5955414012738853, + "eval_Qnli-dev_manhattan_recall": 0.7923728813559322, + "eval_Qnli-dev_max_accuracy": 0.68359375, + "eval_Qnli-dev_max_accuracy_threshold": 388.3597717285156, + "eval_Qnli-dev_max_ap": 0.7167196607222883, + "eval_Qnli-dev_max_f1": 0.68824531516184, + "eval_Qnli-dev_max_f1_threshold": 319.99456787109375, + "eval_Qnli-dev_max_precision": 0.5955414012738853, + "eval_Qnli-dev_max_recall": 0.923728813559322, + "eval_allNLI-dev_cosine_accuracy": 0.734375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.849510669708252, + "eval_allNLI-dev_cosine_ap": 0.617201268548783, + "eval_allNLI-dev_cosine_f1": 0.6284403669724771, + "eval_allNLI-dev_cosine_f1_threshold": 0.8078023195266724, + "eval_allNLI-dev_cosine_precision": 0.5209125475285171, + "eval_allNLI-dev_cosine_recall": 0.791907514450867, + "eval_allNLI-dev_dot_accuracy": 0.6875, + "eval_allNLI-dev_dot_accuracy_threshold": 411.1774597167969, + "eval_allNLI-dev_dot_ap": 0.5148398997702295, + "eval_allNLI-dev_dot_f1": 0.5828571428571427, + "eval_allNLI-dev_dot_f1_threshold": 333.9784240722656, + "eval_allNLI-dev_dot_precision": 0.4346590909090909, + "eval_allNLI-dev_dot_recall": 0.884393063583815, + "eval_allNLI-dev_euclidean_accuracy": 0.736328125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 12.117086410522461, + "eval_allNLI-dev_euclidean_ap": 0.6195431051448221, + "eval_allNLI-dev_euclidean_f1": 0.6399999999999999, + "eval_allNLI-dev_euclidean_f1_threshold": 12.117086410522461, + "eval_allNLI-dev_euclidean_precision": 0.594059405940594, + "eval_allNLI-dev_euclidean_recall": 0.6936416184971098, + "eval_allNLI-dev_manhattan_accuracy": 0.744140625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 214.7247314453125, + "eval_allNLI-dev_manhattan_ap": 0.620953158827849, + "eval_allNLI-dev_manhattan_f1": 0.6469002695417789, + "eval_allNLI-dev_manhattan_f1_threshold": 214.7247314453125, + "eval_allNLI-dev_manhattan_precision": 0.6060606060606061, + "eval_allNLI-dev_manhattan_recall": 0.6936416184971098, + "eval_allNLI-dev_max_accuracy": 0.744140625, + "eval_allNLI-dev_max_accuracy_threshold": 411.1774597167969, + "eval_allNLI-dev_max_ap": 0.620953158827849, + "eval_allNLI-dev_max_f1": 0.6469002695417789, + "eval_allNLI-dev_max_f1_threshold": 333.9784240722656, + "eval_allNLI-dev_max_precision": 0.6060606060606061, + "eval_allNLI-dev_max_recall": 0.884393063583815, + "eval_sequential_score": 0.7167196607222883, + "eval_sts-test_pearson_cosine": 0.8359021538437803, + "eval_sts-test_pearson_dot": 0.7917786144294416, + "eval_sts-test_pearson_euclidean": 0.8714581166329882, + "eval_sts-test_pearson_manhattan": 0.8701583411855128, + "eval_sts-test_pearson_max": 0.8714581166329882, + "eval_sts-test_spearman_cosine": 0.8758165096471757, + "eval_sts-test_spearman_dot": 0.7892398243341464, + "eval_sts-test_spearman_euclidean": 0.8721260521252726, + "eval_sts-test_spearman_manhattan": 0.8710030192206554, + "eval_sts-test_spearman_max": 0.8758165096471757, + "eval_vitaminc-pairs_loss": 3.4827423095703125, + "eval_vitaminc-pairs_runtime": 5.5588, + "eval_vitaminc-pairs_samples_per_second": 23.027, + "eval_vitaminc-pairs_steps_per_second": 0.36, + "step": 1560 + }, + { + "epoch": 1.2102404965089217, + "eval_negation-triplets_loss": 1.653562307357788, + "eval_negation-triplets_runtime": 1.1369, + "eval_negation-triplets_samples_per_second": 112.586, + "eval_negation-triplets_steps_per_second": 1.759, + "step": 1560 + }, + { + "epoch": 1.2102404965089217, + "eval_scitail-pairs-pos_loss": 0.1651732176542282, + "eval_scitail-pairs-pos_runtime": 1.3833, + "eval_scitail-pairs-pos_samples_per_second": 92.53, + "eval_scitail-pairs-pos_steps_per_second": 1.446, + "step": 1560 + }, + { + "epoch": 1.2102404965089217, + "eval_scitail-pairs-qa_loss": 0.007786516100168228, + "eval_scitail-pairs-qa_runtime": 0.8005, + "eval_scitail-pairs-qa_samples_per_second": 159.893, + "eval_scitail-pairs-qa_steps_per_second": 2.498, + "step": 1560 + }, + { + "epoch": 1.2102404965089217, + "eval_xsum-pairs_loss": 0.1304883360862732, + "eval_xsum-pairs_runtime": 3.9611, + "eval_xsum-pairs_samples_per_second": 32.314, + "eval_xsum-pairs_steps_per_second": 0.505, + "step": 1560 + }, + { + "epoch": 1.2102404965089217, + "eval_sciq_pairs_loss": 0.05155816674232483, + "eval_sciq_pairs_runtime": 6.15, + "eval_sciq_pairs_samples_per_second": 20.813, + "eval_sciq_pairs_steps_per_second": 0.325, + "step": 1560 + }, + { + "epoch": 1.2102404965089217, + "eval_qasc_pairs_loss": 0.22321471571922302, + "eval_qasc_pairs_runtime": 0.9431, + "eval_qasc_pairs_samples_per_second": 135.72, + "eval_qasc_pairs_steps_per_second": 2.121, + "step": 1560 + }, + { + "epoch": 1.2102404965089217, + "eval_openbookqa_pairs_loss": 1.0748668909072876, + "eval_openbookqa_pairs_runtime": 0.8933, + "eval_openbookqa_pairs_samples_per_second": 143.296, + "eval_openbookqa_pairs_steps_per_second": 2.239, + "step": 1560 + }, + { + "epoch": 1.2102404965089217, + "eval_msmarco_pairs_loss": 0.362208753824234, + "eval_msmarco_pairs_runtime": 1.9945, + "eval_msmarco_pairs_samples_per_second": 64.175, + "eval_msmarco_pairs_steps_per_second": 1.003, + "step": 1560 + }, + { + "epoch": 1.2102404965089217, + "eval_nq_pairs_loss": 0.6640698313713074, + "eval_nq_pairs_runtime": 3.366, + "eval_nq_pairs_samples_per_second": 38.027, + "eval_nq_pairs_steps_per_second": 0.594, + "step": 1560 + }, + { + "epoch": 1.2102404965089217, + "eval_trivia_pairs_loss": 0.8185591697692871, + "eval_trivia_pairs_runtime": 4.5121, + "eval_trivia_pairs_samples_per_second": 28.368, + "eval_trivia_pairs_steps_per_second": 0.443, + "step": 1560 + }, + { + "epoch": 1.2102404965089217, + "eval_gooaq_pairs_loss": 0.6577709913253784, + "eval_gooaq_pairs_runtime": 1.4818, + "eval_gooaq_pairs_samples_per_second": 86.382, + "eval_gooaq_pairs_steps_per_second": 1.35, + "step": 1560 + }, + { + "epoch": 1.2102404965089217, + "eval_paws-pos_loss": 0.029750881716609, + "eval_paws-pos_runtime": 1.0539, + "eval_paws-pos_samples_per_second": 121.455, + "eval_paws-pos_steps_per_second": 1.898, + "step": 1560 + }, + { + "epoch": 1.2102404965089217, + "eval_global_dataset_loss": 0.4120141565799713, + "eval_global_dataset_runtime": 19.4571, + "eval_global_dataset_samples_per_second": 21.38, + "eval_global_dataset_steps_per_second": 0.36, + "step": 1560 + }, + { + "epoch": 1.2110162916989915, + "grad_norm": 4.9396867752075195, + "learning_rate": 2.3673540773635956e-05, + "loss": 0.2553, + "step": 1561 + }, + { + "epoch": 1.2117920868890613, + "grad_norm": 6.797117233276367, + "learning_rate": 2.3646035813194312e-05, + "loss": 0.5118, + "step": 1562 + }, + { + "epoch": 1.212567882079131, + "grad_norm": 12.328256607055664, + "learning_rate": 2.3618516305661114e-05, + "loss": 0.9595, + "step": 1563 + }, + { + "epoch": 1.213343677269201, + "grad_norm": 10.576864242553711, + "learning_rate": 2.3590982342102207e-05, + "loss": 0.7337, + "step": 1564 + }, + { + "epoch": 1.2141194724592708, + "grad_norm": 6.28556489944458, + "learning_rate": 2.356343401363131e-05, + "loss": 0.1941, + "step": 1565 + }, + { + "epoch": 1.2148952676493405, + "grad_norm": 4.195956707000732, + "learning_rate": 2.3535871411409645e-05, + "loss": 0.1899, + "step": 1566 + }, + { + "epoch": 1.2156710628394105, + "grad_norm": 13.180573463439941, + "learning_rate": 2.3508294626645698e-05, + "loss": 1.8494, + "step": 1567 + }, + { + "epoch": 1.2164468580294803, + "grad_norm": 10.092689514160156, + "learning_rate": 2.3480703750594857e-05, + "loss": 0.6463, + "step": 1568 + }, + { + "epoch": 1.21722265321955, + "grad_norm": 6.918635845184326, + "learning_rate": 2.3453098874559155e-05, + "loss": 0.5506, + "step": 1569 + }, + { + "epoch": 1.2179984484096198, + "grad_norm": 12.60912799835205, + "learning_rate": 2.342548008988695e-05, + "loss": 0.8244, + "step": 1570 + }, + { + "epoch": 1.2187742435996898, + "grad_norm": 1.5747973918914795, + "learning_rate": 2.3397847487972635e-05, + "loss": 0.0437, + "step": 1571 + }, + { + "epoch": 1.2195500387897595, + "grad_norm": 6.859606742858887, + "learning_rate": 2.337020116025631e-05, + "loss": 0.5794, + "step": 1572 + }, + { + "epoch": 1.2203258339798293, + "grad_norm": 0.7837724089622498, + "learning_rate": 2.3342541198223505e-05, + "loss": 0.0483, + "step": 1573 + }, + { + "epoch": 1.221101629169899, + "grad_norm": 7.525580406188965, + "learning_rate": 2.3314867693404876e-05, + "loss": 0.2608, + "step": 1574 + }, + { + "epoch": 1.221877424359969, + "grad_norm": 6.473620414733887, + "learning_rate": 2.3287180737375874e-05, + "loss": 0.3172, + "step": 1575 + }, + { + "epoch": 1.2226532195500388, + "grad_norm": 12.195878028869629, + "learning_rate": 2.325948042175647e-05, + "loss": 1.0814, + "step": 1576 + }, + { + "epoch": 1.2234290147401086, + "grad_norm": 3.658269166946411, + "learning_rate": 2.323176683821086e-05, + "loss": 0.1237, + "step": 1577 + }, + { + "epoch": 1.2242048099301783, + "grad_norm": 10.743631362915039, + "learning_rate": 2.3204040078447116e-05, + "loss": 0.8407, + "step": 1578 + }, + { + "epoch": 1.2249806051202483, + "grad_norm": 5.95540714263916, + "learning_rate": 2.3176300234216936e-05, + "loss": 0.1473, + "step": 1579 + }, + { + "epoch": 1.225756400310318, + "grad_norm": 5.932210922241211, + "learning_rate": 2.3148547397315307e-05, + "loss": 0.2965, + "step": 1580 + }, + { + "epoch": 1.225756400310318, + "eval_Qnli-dev_cosine_accuracy": 0.681640625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.857793927192688, + "eval_Qnli-dev_cosine_ap": 0.7089311667601617, + "eval_Qnli-dev_cosine_f1": 0.6852459016393443, + "eval_Qnli-dev_cosine_f1_threshold": 0.7522059679031372, + "eval_Qnli-dev_cosine_precision": 0.5588235294117647, + "eval_Qnli-dev_cosine_recall": 0.885593220338983, + "eval_Qnli-dev_dot_accuracy": 0.66015625, + "eval_Qnli-dev_dot_accuracy_threshold": 415.0566101074219, + "eval_Qnli-dev_dot_ap": 0.6431455672431684, + "eval_Qnli-dev_dot_f1": 0.6677265500794912, + "eval_Qnli-dev_dot_f1_threshold": 357.79400634765625, + "eval_Qnli-dev_dot_precision": 0.5343511450381679, + "eval_Qnli-dev_dot_recall": 0.8898305084745762, + "eval_Qnli-dev_euclidean_accuracy": 0.67578125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.450175285339355, + "eval_Qnli-dev_euclidean_ap": 0.7034961306085394, + "eval_Qnli-dev_euclidean_f1": 0.6819672131147542, + "eval_Qnli-dev_euclidean_f1_threshold": 16.000988006591797, + "eval_Qnli-dev_euclidean_precision": 0.5561497326203209, + "eval_Qnli-dev_euclidean_recall": 0.8813559322033898, + "eval_Qnli-dev_manhattan_accuracy": 0.67578125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 234.38916015625, + "eval_Qnli-dev_manhattan_ap": 0.7059056264127279, + "eval_Qnli-dev_manhattan_f1": 0.6822742474916388, + "eval_Qnli-dev_manhattan_f1_threshold": 284.732421875, + "eval_Qnli-dev_manhattan_precision": 0.56353591160221, + "eval_Qnli-dev_manhattan_recall": 0.864406779661017, + "eval_Qnli-dev_max_accuracy": 0.681640625, + "eval_Qnli-dev_max_accuracy_threshold": 415.0566101074219, + "eval_Qnli-dev_max_ap": 0.7089311667601617, + "eval_Qnli-dev_max_f1": 0.6852459016393443, + "eval_Qnli-dev_max_f1_threshold": 357.79400634765625, + "eval_Qnli-dev_max_precision": 0.56353591160221, + "eval_Qnli-dev_max_recall": 0.8898305084745762, + "eval_allNLI-dev_cosine_accuracy": 0.73046875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8626295328140259, + "eval_allNLI-dev_cosine_ap": 0.612243316101513, + "eval_allNLI-dev_cosine_f1": 0.6313465783664459, + "eval_allNLI-dev_cosine_f1_threshold": 0.7991390228271484, + "eval_allNLI-dev_cosine_precision": 0.5107142857142857, + "eval_allNLI-dev_cosine_recall": 0.8265895953757225, + "eval_allNLI-dev_dot_accuracy": 0.677734375, + "eval_allNLI-dev_dot_accuracy_threshold": 489.10205078125, + "eval_allNLI-dev_dot_ap": 0.4969245380948055, + "eval_allNLI-dev_dot_f1": 0.5815602836879432, + "eval_allNLI-dev_dot_f1_threshold": 387.7714538574219, + "eval_allNLI-dev_dot_precision": 0.492, + "eval_allNLI-dev_dot_recall": 0.7109826589595376, + "eval_allNLI-dev_euclidean_accuracy": 0.734375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.293333053588867, + "eval_allNLI-dev_euclidean_ap": 0.6187619041059933, + "eval_allNLI-dev_euclidean_f1": 0.6380090497737557, + "eval_allNLI-dev_euclidean_f1_threshold": 13.743589401245117, + "eval_allNLI-dev_euclidean_precision": 0.5241635687732342, + "eval_allNLI-dev_euclidean_recall": 0.815028901734104, + "eval_allNLI-dev_manhattan_accuracy": 0.740234375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 216.93431091308594, + "eval_allNLI-dev_manhattan_ap": 0.6192147814409138, + "eval_allNLI-dev_manhattan_f1": 0.638095238095238, + "eval_allNLI-dev_manhattan_f1_threshold": 240.5207977294922, + "eval_allNLI-dev_manhattan_precision": 0.5425101214574899, + "eval_allNLI-dev_manhattan_recall": 0.7745664739884393, + "eval_allNLI-dev_max_accuracy": 0.740234375, + "eval_allNLI-dev_max_accuracy_threshold": 489.10205078125, + "eval_allNLI-dev_max_ap": 0.6192147814409138, + "eval_allNLI-dev_max_f1": 0.638095238095238, + "eval_allNLI-dev_max_f1_threshold": 387.7714538574219, + "eval_allNLI-dev_max_precision": 0.5425101214574899, + "eval_allNLI-dev_max_recall": 0.8265895953757225, + "eval_sequential_score": 0.7089311667601617, + "eval_sts-test_pearson_cosine": 0.8264897776358879, + "eval_sts-test_pearson_dot": 0.783374956457153, + "eval_sts-test_pearson_euclidean": 0.860394912663312, + "eval_sts-test_pearson_manhattan": 0.8589731838156086, + "eval_sts-test_pearson_max": 0.860394912663312, + "eval_sts-test_spearman_cosine": 0.8678902109807923, + "eval_sts-test_spearman_dot": 0.7806161667615009, + "eval_sts-test_spearman_euclidean": 0.8627395197996879, + "eval_sts-test_spearman_manhattan": 0.8617376285596572, + "eval_sts-test_spearman_max": 0.8678902109807923, + "eval_vitaminc-pairs_loss": 3.6752569675445557, + "eval_vitaminc-pairs_runtime": 5.4596, + "eval_vitaminc-pairs_samples_per_second": 23.445, + "eval_vitaminc-pairs_steps_per_second": 0.366, + "step": 1580 + }, + { + "epoch": 1.225756400310318, + "eval_negation-triplets_loss": 1.5785763263702393, + "eval_negation-triplets_runtime": 1.1062, + "eval_negation-triplets_samples_per_second": 115.713, + "eval_negation-triplets_steps_per_second": 1.808, + "step": 1580 + }, + { + "epoch": 1.225756400310318, + "eval_scitail-pairs-pos_loss": 0.15904644131660461, + "eval_scitail-pairs-pos_runtime": 1.34, + "eval_scitail-pairs-pos_samples_per_second": 95.523, + "eval_scitail-pairs-pos_steps_per_second": 1.493, + "step": 1580 + }, + { + "epoch": 1.225756400310318, + "eval_scitail-pairs-qa_loss": 0.005985472351312637, + "eval_scitail-pairs-qa_runtime": 0.7899, + "eval_scitail-pairs-qa_samples_per_second": 162.041, + "eval_scitail-pairs-qa_steps_per_second": 2.532, + "step": 1580 + }, + { + "epoch": 1.225756400310318, + "eval_xsum-pairs_loss": 0.1399654746055603, + "eval_xsum-pairs_runtime": 3.9128, + "eval_xsum-pairs_samples_per_second": 32.713, + "eval_xsum-pairs_steps_per_second": 0.511, + "step": 1580 + }, + { + "epoch": 1.225756400310318, + "eval_sciq_pairs_loss": 0.05025120824575424, + "eval_sciq_pairs_runtime": 6.0756, + "eval_sciq_pairs_samples_per_second": 21.068, + "eval_sciq_pairs_steps_per_second": 0.329, + "step": 1580 + }, + { + "epoch": 1.225756400310318, + "eval_qasc_pairs_loss": 0.25028979778289795, + "eval_qasc_pairs_runtime": 0.9192, + "eval_qasc_pairs_samples_per_second": 139.255, + "eval_qasc_pairs_steps_per_second": 2.176, + "step": 1580 + }, + { + "epoch": 1.225756400310318, + "eval_openbookqa_pairs_loss": 1.0688073635101318, + "eval_openbookqa_pairs_runtime": 0.8796, + "eval_openbookqa_pairs_samples_per_second": 145.517, + "eval_openbookqa_pairs_steps_per_second": 2.274, + "step": 1580 + }, + { + "epoch": 1.225756400310318, + "eval_msmarco_pairs_loss": 0.42507630586624146, + "eval_msmarco_pairs_runtime": 1.9686, + "eval_msmarco_pairs_samples_per_second": 65.021, + "eval_msmarco_pairs_steps_per_second": 1.016, + "step": 1580 + }, + { + "epoch": 1.225756400310318, + "eval_nq_pairs_loss": 0.6445481777191162, + "eval_nq_pairs_runtime": 3.3393, + "eval_nq_pairs_samples_per_second": 38.332, + "eval_nq_pairs_steps_per_second": 0.599, + "step": 1580 + }, + { + "epoch": 1.225756400310318, + "eval_trivia_pairs_loss": 0.7414390444755554, + "eval_trivia_pairs_runtime": 4.4879, + "eval_trivia_pairs_samples_per_second": 28.521, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 1580 + }, + { + "epoch": 1.225756400310318, + "eval_gooaq_pairs_loss": 0.5973804593086243, + "eval_gooaq_pairs_runtime": 1.4676, + "eval_gooaq_pairs_samples_per_second": 87.219, + "eval_gooaq_pairs_steps_per_second": 1.363, + "step": 1580 + }, + { + "epoch": 1.225756400310318, + "eval_paws-pos_loss": 0.031936317682266235, + "eval_paws-pos_runtime": 1.0437, + "eval_paws-pos_samples_per_second": 122.635, + "eval_paws-pos_steps_per_second": 1.916, + "step": 1580 + }, + { + "epoch": 1.225756400310318, + "eval_global_dataset_loss": 0.4217734634876251, + "eval_global_dataset_runtime": 19.4138, + "eval_global_dataset_samples_per_second": 21.428, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1580 + }, + { + "epoch": 1.2265321955003878, + "grad_norm": 4.737555980682373, + "learning_rate": 2.3120781659580218e-05, + "loss": 0.2391, + "step": 1581 + }, + { + "epoch": 1.2273079906904578, + "grad_norm": 4.559937953948975, + "learning_rate": 2.309300311289233e-05, + "loss": 0.1864, + "step": 1582 + }, + { + "epoch": 1.2280837858805276, + "grad_norm": 2.631824254989624, + "learning_rate": 2.3065211849174705e-05, + "loss": 0.0535, + "step": 1583 + }, + { + "epoch": 1.2288595810705973, + "grad_norm": 6.840622425079346, + "learning_rate": 2.303740796039249e-05, + "loss": 0.442, + "step": 1584 + }, + { + "epoch": 1.229635376260667, + "grad_norm": 5.308975696563721, + "learning_rate": 2.3009591538552597e-05, + "loss": 0.1903, + "step": 1585 + }, + { + "epoch": 1.230411171450737, + "grad_norm": 14.915870666503906, + "learning_rate": 2.2981762675703434e-05, + "loss": 2.3485, + "step": 1586 + }, + { + "epoch": 1.2311869666408068, + "grad_norm": 5.644608497619629, + "learning_rate": 2.2953921463934553e-05, + "loss": 0.2044, + "step": 1587 + }, + { + "epoch": 1.2319627618308766, + "grad_norm": 6.448950290679932, + "learning_rate": 2.292606799537638e-05, + "loss": 0.582, + "step": 1588 + }, + { + "epoch": 1.2327385570209466, + "grad_norm": 6.304661273956299, + "learning_rate": 2.2898202362199907e-05, + "loss": 0.387, + "step": 1589 + }, + { + "epoch": 1.2335143522110164, + "grad_norm": 11.64503288269043, + "learning_rate": 2.2870324656616376e-05, + "loss": 0.7817, + "step": 1590 + }, + { + "epoch": 1.2342901474010861, + "grad_norm": 6.087871074676514, + "learning_rate": 2.2842434970876967e-05, + "loss": 0.185, + "step": 1591 + }, + { + "epoch": 1.2350659425911559, + "grad_norm": 0.4097508490085602, + "learning_rate": 2.281453339727252e-05, + "loss": 0.0325, + "step": 1592 + }, + { + "epoch": 1.2358417377812256, + "grad_norm": 10.933095932006836, + "learning_rate": 2.278662002813321e-05, + "loss": 0.9596, + "step": 1593 + }, + { + "epoch": 1.2366175329712956, + "grad_norm": 7.441780090332031, + "learning_rate": 2.2758694955828244e-05, + "loss": 0.3216, + "step": 1594 + }, + { + "epoch": 1.2373933281613654, + "grad_norm": 0.9908187389373779, + "learning_rate": 2.2730758272765552e-05, + "loss": 0.0246, + "step": 1595 + }, + { + "epoch": 1.2381691233514351, + "grad_norm": 5.871782302856445, + "learning_rate": 2.2702810071391484e-05, + "loss": 0.452, + "step": 1596 + }, + { + "epoch": 1.2389449185415051, + "grad_norm": 0.3802882134914398, + "learning_rate": 2.2674850444190524e-05, + "loss": 0.0113, + "step": 1597 + }, + { + "epoch": 1.2397207137315749, + "grad_norm": 6.073673248291016, + "learning_rate": 2.264687948368494e-05, + "loss": 0.214, + "step": 1598 + }, + { + "epoch": 1.2404965089216446, + "grad_norm": 5.662391185760498, + "learning_rate": 2.2618897282434517e-05, + "loss": 0.1592, + "step": 1599 + }, + { + "epoch": 1.2412723041117144, + "grad_norm": 7.3329339027404785, + "learning_rate": 2.2590903933036237e-05, + "loss": 0.2291, + "step": 1600 + }, + { + "epoch": 1.2412723041117144, + "eval_Qnli-dev_cosine_accuracy": 0.68359375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8140503168106079, + "eval_Qnli-dev_cosine_ap": 0.7202980627220541, + "eval_Qnli-dev_cosine_f1": 0.6893039049235994, + "eval_Qnli-dev_cosine_f1_threshold": 0.7504552006721497, + "eval_Qnli-dev_cosine_precision": 0.5750708215297451, + "eval_Qnli-dev_cosine_recall": 0.8601694915254238, + "eval_Qnli-dev_dot_accuracy": 0.6484375, + "eval_Qnli-dev_dot_accuracy_threshold": 429.31719970703125, + "eval_Qnli-dev_dot_ap": 0.6359286383364835, + "eval_Qnli-dev_dot_f1": 0.6720516962843297, + "eval_Qnli-dev_dot_f1_threshold": 352.7978515625, + "eval_Qnli-dev_dot_precision": 0.5430809399477807, + "eval_Qnli-dev_dot_recall": 0.8813559322033898, + "eval_Qnli-dev_euclidean_accuracy": 0.68359375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 14.442277908325195, + "eval_Qnli-dev_euclidean_ap": 0.7222410397276917, + "eval_Qnli-dev_euclidean_f1": 0.6912280701754385, + "eval_Qnli-dev_euclidean_f1_threshold": 15.659895896911621, + "eval_Qnli-dev_euclidean_precision": 0.5898203592814372, + "eval_Qnli-dev_euclidean_recall": 0.8347457627118644, + "eval_Qnli-dev_manhattan_accuracy": 0.681640625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 231.46694946289062, + "eval_Qnli-dev_manhattan_ap": 0.726070142697283, + "eval_Qnli-dev_manhattan_f1": 0.6912280701754385, + "eval_Qnli-dev_manhattan_f1_threshold": 283.40924072265625, + "eval_Qnli-dev_manhattan_precision": 0.5898203592814372, + "eval_Qnli-dev_manhattan_recall": 0.8347457627118644, + "eval_Qnli-dev_max_accuracy": 0.68359375, + "eval_Qnli-dev_max_accuracy_threshold": 429.31719970703125, + "eval_Qnli-dev_max_ap": 0.726070142697283, + "eval_Qnli-dev_max_f1": 0.6912280701754385, + "eval_Qnli-dev_max_f1_threshold": 352.7978515625, + "eval_Qnli-dev_max_precision": 0.5898203592814372, + "eval_Qnli-dev_max_recall": 0.8813559322033898, + "eval_allNLI-dev_cosine_accuracy": 0.720703125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8938466310501099, + "eval_allNLI-dev_cosine_ap": 0.6119584583042442, + "eval_allNLI-dev_cosine_f1": 0.6212121212121212, + "eval_allNLI-dev_cosine_f1_threshold": 0.8255913257598877, + "eval_allNLI-dev_cosine_precision": 0.5515695067264574, + "eval_allNLI-dev_cosine_recall": 0.7109826589595376, + "eval_allNLI-dev_dot_accuracy": 0.6875, + "eval_allNLI-dev_dot_accuracy_threshold": 425.5631408691406, + "eval_allNLI-dev_dot_ap": 0.5157616082754238, + "eval_allNLI-dev_dot_f1": 0.5775193798449613, + "eval_allNLI-dev_dot_f1_threshold": 348.901123046875, + "eval_allNLI-dev_dot_precision": 0.43440233236151604, + "eval_allNLI-dev_dot_recall": 0.861271676300578, + "eval_allNLI-dev_euclidean_accuracy": 0.728515625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 12.375224113464355, + "eval_allNLI-dev_euclidean_ap": 0.6167007575456817, + "eval_allNLI-dev_euclidean_f1": 0.6340326340326341, + "eval_allNLI-dev_euclidean_f1_threshold": 13.584259033203125, + "eval_allNLI-dev_euclidean_precision": 0.53125, + "eval_allNLI-dev_euclidean_recall": 0.7861271676300579, + "eval_allNLI-dev_manhattan_accuracy": 0.736328125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 211.34193420410156, + "eval_allNLI-dev_manhattan_ap": 0.6184757557836061, + "eval_allNLI-dev_manhattan_f1": 0.6417910447761194, + "eval_allNLI-dev_manhattan_f1_threshold": 237.83590698242188, + "eval_allNLI-dev_manhattan_precision": 0.5633187772925764, + "eval_allNLI-dev_manhattan_recall": 0.7456647398843931, + "eval_allNLI-dev_max_accuracy": 0.736328125, + "eval_allNLI-dev_max_accuracy_threshold": 425.5631408691406, + "eval_allNLI-dev_max_ap": 0.6184757557836061, + "eval_allNLI-dev_max_f1": 0.6417910447761194, + "eval_allNLI-dev_max_f1_threshold": 348.901123046875, + "eval_allNLI-dev_max_precision": 0.5633187772925764, + "eval_allNLI-dev_max_recall": 0.861271676300578, + "eval_sequential_score": 0.726070142697283, + "eval_sts-test_pearson_cosine": 0.8304270525447578, + "eval_sts-test_pearson_dot": 0.7958604481048973, + "eval_sts-test_pearson_euclidean": 0.867181169291225, + "eval_sts-test_pearson_manhattan": 0.8652502630873753, + "eval_sts-test_pearson_max": 0.867181169291225, + "eval_sts-test_spearman_cosine": 0.8744124388210205, + "eval_sts-test_spearman_dot": 0.7914709879733374, + "eval_sts-test_spearman_euclidean": 0.8684974408466123, + "eval_sts-test_spearman_manhattan": 0.8678822005050033, + "eval_sts-test_spearman_max": 0.8744124388210205, + "eval_vitaminc-pairs_loss": 3.768341064453125, + "eval_vitaminc-pairs_runtime": 5.4661, + "eval_vitaminc-pairs_samples_per_second": 23.417, + "eval_vitaminc-pairs_steps_per_second": 0.366, + "step": 1600 + }, + { + "epoch": 1.2412723041117144, + "eval_negation-triplets_loss": 1.688029170036316, + "eval_negation-triplets_runtime": 1.0868, + "eval_negation-triplets_samples_per_second": 117.782, + "eval_negation-triplets_steps_per_second": 1.84, + "step": 1600 + }, + { + "epoch": 1.2412723041117144, + "eval_scitail-pairs-pos_loss": 0.1675368696451187, + "eval_scitail-pairs-pos_runtime": 1.3393, + "eval_scitail-pairs-pos_samples_per_second": 95.57, + "eval_scitail-pairs-pos_steps_per_second": 1.493, + "step": 1600 + }, + { + "epoch": 1.2412723041117144, + "eval_scitail-pairs-qa_loss": 0.004187453538179398, + "eval_scitail-pairs-qa_runtime": 0.7913, + "eval_scitail-pairs-qa_samples_per_second": 161.759, + "eval_scitail-pairs-qa_steps_per_second": 2.527, + "step": 1600 + }, + { + "epoch": 1.2412723041117144, + "eval_xsum-pairs_loss": 0.14741384983062744, + "eval_xsum-pairs_runtime": 3.9287, + "eval_xsum-pairs_samples_per_second": 32.581, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 1600 + }, + { + "epoch": 1.2412723041117144, + "eval_sciq_pairs_loss": 0.04968512803316116, + "eval_sciq_pairs_runtime": 5.9555, + "eval_sciq_pairs_samples_per_second": 21.493, + "eval_sciq_pairs_steps_per_second": 0.336, + "step": 1600 + }, + { + "epoch": 1.2412723041117144, + "eval_qasc_pairs_loss": 0.2350328415632248, + "eval_qasc_pairs_runtime": 0.9169, + "eval_qasc_pairs_samples_per_second": 139.601, + "eval_qasc_pairs_steps_per_second": 2.181, + "step": 1600 + }, + { + "epoch": 1.2412723041117144, + "eval_openbookqa_pairs_loss": 1.1051180362701416, + "eval_openbookqa_pairs_runtime": 0.8675, + "eval_openbookqa_pairs_samples_per_second": 147.556, + "eval_openbookqa_pairs_steps_per_second": 2.306, + "step": 1600 + }, + { + "epoch": 1.2412723041117144, + "eval_msmarco_pairs_loss": 0.43787407875061035, + "eval_msmarco_pairs_runtime": 1.9712, + "eval_msmarco_pairs_samples_per_second": 64.936, + "eval_msmarco_pairs_steps_per_second": 1.015, + "step": 1600 + }, + { + "epoch": 1.2412723041117144, + "eval_nq_pairs_loss": 0.6179434061050415, + "eval_nq_pairs_runtime": 3.346, + "eval_nq_pairs_samples_per_second": 38.255, + "eval_nq_pairs_steps_per_second": 0.598, + "step": 1600 + }, + { + "epoch": 1.2412723041117144, + "eval_trivia_pairs_loss": 0.7854765057563782, + "eval_trivia_pairs_runtime": 4.4876, + "eval_trivia_pairs_samples_per_second": 28.523, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 1600 + }, + { + "epoch": 1.2412723041117144, + "eval_gooaq_pairs_loss": 0.5935078263282776, + "eval_gooaq_pairs_runtime": 1.4686, + "eval_gooaq_pairs_samples_per_second": 87.158, + "eval_gooaq_pairs_steps_per_second": 1.362, + "step": 1600 + }, + { + "epoch": 1.2412723041117144, + "eval_paws-pos_loss": 0.031957607716321945, + "eval_paws-pos_runtime": 1.0337, + "eval_paws-pos_samples_per_second": 123.824, + "eval_paws-pos_steps_per_second": 1.935, + "step": 1600 + }, + { + "epoch": 1.2412723041117144, + "eval_global_dataset_loss": 0.4228430390357971, + "eval_global_dataset_runtime": 19.4113, + "eval_global_dataset_samples_per_second": 21.431, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1600 + }, + { + "epoch": 1.2420480993017844, + "grad_norm": 12.785367012023926, + "learning_rate": 2.2562899528123976e-05, + "loss": 1.1211, + "step": 1601 + }, + { + "epoch": 1.2428238944918542, + "grad_norm": 9.290999412536621, + "learning_rate": 2.253488416036819e-05, + "loss": 0.6048, + "step": 1602 + }, + { + "epoch": 1.243599689681924, + "grad_norm": 7.840250015258789, + "learning_rate": 2.2506857922475602e-05, + "loss": 0.2928, + "step": 1603 + }, + { + "epoch": 1.244375484871994, + "grad_norm": 0.6756582856178284, + "learning_rate": 2.2478820907188926e-05, + "loss": 0.0314, + "step": 1604 + }, + { + "epoch": 1.2451512800620637, + "grad_norm": 3.825641393661499, + "learning_rate": 2.2450773207286524e-05, + "loss": 0.0852, + "step": 1605 + }, + { + "epoch": 1.2459270752521334, + "grad_norm": 7.228362083435059, + "learning_rate": 2.2422714915582127e-05, + "loss": 0.2909, + "step": 1606 + }, + { + "epoch": 1.2467028704422032, + "grad_norm": 4.394501686096191, + "learning_rate": 2.2394646124924505e-05, + "loss": 0.1415, + "step": 1607 + }, + { + "epoch": 1.2474786656322732, + "grad_norm": 4.166509628295898, + "learning_rate": 2.2366566928197176e-05, + "loss": 0.1048, + "step": 1608 + }, + { + "epoch": 1.248254460822343, + "grad_norm": 7.84224796295166, + "learning_rate": 2.2338477418318097e-05, + "loss": 0.3505, + "step": 1609 + }, + { + "epoch": 1.2490302560124127, + "grad_norm": 9.903580665588379, + "learning_rate": 2.2310377688239334e-05, + "loss": 0.6822, + "step": 1610 + }, + { + "epoch": 1.2498060512024824, + "grad_norm": 11.826993942260742, + "learning_rate": 2.2282267830946804e-05, + "loss": 0.9485, + "step": 1611 + }, + { + "epoch": 1.2505818463925524, + "grad_norm": 8.946475982666016, + "learning_rate": 2.225414793945991e-05, + "loss": 0.5965, + "step": 1612 + }, + { + "epoch": 1.2513576415826222, + "grad_norm": 5.708798408508301, + "learning_rate": 2.222601810683127e-05, + "loss": 0.209, + "step": 1613 + }, + { + "epoch": 1.252133436772692, + "grad_norm": 5.977161407470703, + "learning_rate": 2.2197878426146402e-05, + "loss": 0.3293, + "step": 1614 + }, + { + "epoch": 1.2529092319627617, + "grad_norm": 6.413185119628906, + "learning_rate": 2.2169728990523404e-05, + "loss": 0.3497, + "step": 1615 + }, + { + "epoch": 1.2536850271528317, + "grad_norm": 6.827978134155273, + "learning_rate": 2.214156989311265e-05, + "loss": 0.2358, + "step": 1616 + }, + { + "epoch": 1.2544608223429015, + "grad_norm": 10.524563789367676, + "learning_rate": 2.211340122709651e-05, + "loss": 0.7111, + "step": 1617 + }, + { + "epoch": 1.2552366175329712, + "grad_norm": 9.276382446289062, + "learning_rate": 2.2085223085688993e-05, + "loss": 0.4747, + "step": 1618 + }, + { + "epoch": 1.2560124127230412, + "grad_norm": 7.415008544921875, + "learning_rate": 2.205703556213548e-05, + "loss": 0.337, + "step": 1619 + }, + { + "epoch": 1.256788207913111, + "grad_norm": 0.38036254048347473, + "learning_rate": 2.2028838749712387e-05, + "loss": 0.0072, + "step": 1620 + }, + { + "epoch": 1.256788207913111, + "eval_Qnli-dev_cosine_accuracy": 0.673828125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8317668437957764, + "eval_Qnli-dev_cosine_ap": 0.7014318283089398, + "eval_Qnli-dev_cosine_f1": 0.6782334384858044, + "eval_Qnli-dev_cosine_f1_threshold": 0.7380469441413879, + "eval_Qnli-dev_cosine_precision": 0.5402010050251256, + "eval_Qnli-dev_cosine_recall": 0.9110169491525424, + "eval_Qnli-dev_dot_accuracy": 0.658203125, + "eval_Qnli-dev_dot_accuracy_threshold": 406.4397277832031, + "eval_Qnli-dev_dot_ap": 0.6254911926168678, + "eval_Qnli-dev_dot_f1": 0.6697530864197531, + "eval_Qnli-dev_dot_f1_threshold": 332.8388366699219, + "eval_Qnli-dev_dot_precision": 0.5266990291262136, + "eval_Qnli-dev_dot_recall": 0.9194915254237288, + "eval_Qnli-dev_euclidean_accuracy": 0.673828125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.04751205444336, + "eval_Qnli-dev_euclidean_ap": 0.6971807612256411, + "eval_Qnli-dev_euclidean_f1": 0.6708463949843261, + "eval_Qnli-dev_euclidean_f1_threshold": 16.19192123413086, + "eval_Qnli-dev_euclidean_precision": 0.5323383084577115, + "eval_Qnli-dev_euclidean_recall": 0.9067796610169492, + "eval_Qnli-dev_manhattan_accuracy": 0.66796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 242.59906005859375, + "eval_Qnli-dev_manhattan_ap": 0.7031373767951252, + "eval_Qnli-dev_manhattan_f1": 0.6691588785046729, + "eval_Qnli-dev_manhattan_f1_threshold": 255.60610961914062, + "eval_Qnli-dev_manhattan_precision": 0.5986622073578596, + "eval_Qnli-dev_manhattan_recall": 0.7584745762711864, + "eval_Qnli-dev_max_accuracy": 0.673828125, + "eval_Qnli-dev_max_accuracy_threshold": 406.4397277832031, + "eval_Qnli-dev_max_ap": 0.7031373767951252, + "eval_Qnli-dev_max_f1": 0.6782334384858044, + "eval_Qnli-dev_max_f1_threshold": 332.8388366699219, + "eval_Qnli-dev_max_precision": 0.5986622073578596, + "eval_Qnli-dev_max_recall": 0.9194915254237288, + "eval_allNLI-dev_cosine_accuracy": 0.724609375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8553310632705688, + "eval_allNLI-dev_cosine_ap": 0.6087050752595364, + "eval_allNLI-dev_cosine_f1": 0.6206896551724138, + "eval_allNLI-dev_cosine_f1_threshold": 0.8054232001304626, + "eval_allNLI-dev_cosine_precision": 0.5152671755725191, + "eval_allNLI-dev_cosine_recall": 0.7803468208092486, + "eval_allNLI-dev_dot_accuracy": 0.697265625, + "eval_allNLI-dev_dot_accuracy_threshold": 401.2228698730469, + "eval_allNLI-dev_dot_ap": 0.5260708632219723, + "eval_allNLI-dev_dot_f1": 0.5843621399176955, + "eval_allNLI-dev_dot_f1_threshold": 355.6851806640625, + "eval_allNLI-dev_dot_precision": 0.4536741214057508, + "eval_allNLI-dev_dot_recall": 0.8208092485549133, + "eval_allNLI-dev_euclidean_accuracy": 0.724609375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.614076614379883, + "eval_allNLI-dev_euclidean_ap": 0.6099943892410107, + "eval_allNLI-dev_euclidean_f1": 0.6200873362445415, + "eval_allNLI-dev_euclidean_f1_threshold": 14.068817138671875, + "eval_allNLI-dev_euclidean_precision": 0.4982456140350877, + "eval_allNLI-dev_euclidean_recall": 0.8208092485549133, + "eval_allNLI-dev_manhattan_accuracy": 0.73046875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 217.39959716796875, + "eval_allNLI-dev_manhattan_ap": 0.6102756238935482, + "eval_allNLI-dev_manhattan_f1": 0.6259946949602122, + "eval_allNLI-dev_manhattan_f1_threshold": 223.59832763671875, + "eval_allNLI-dev_manhattan_precision": 0.5784313725490197, + "eval_allNLI-dev_manhattan_recall": 0.6820809248554913, + "eval_allNLI-dev_max_accuracy": 0.73046875, + "eval_allNLI-dev_max_accuracy_threshold": 401.2228698730469, + "eval_allNLI-dev_max_ap": 0.6102756238935482, + "eval_allNLI-dev_max_f1": 0.6259946949602122, + "eval_allNLI-dev_max_f1_threshold": 355.6851806640625, + "eval_allNLI-dev_max_precision": 0.5784313725490197, + "eval_allNLI-dev_max_recall": 0.8208092485549133, + "eval_sequential_score": 0.7031373767951252, + "eval_sts-test_pearson_cosine": 0.8299253739696071, + "eval_sts-test_pearson_dot": 0.7899912016762258, + "eval_sts-test_pearson_euclidean": 0.8675595131994968, + "eval_sts-test_pearson_manhattan": 0.8663890878770465, + "eval_sts-test_pearson_max": 0.8675595131994968, + "eval_sts-test_spearman_cosine": 0.8731698752972269, + "eval_sts-test_spearman_dot": 0.7861697087976507, + "eval_sts-test_spearman_euclidean": 0.8698085633028574, + "eval_sts-test_spearman_manhattan": 0.8698949779886044, + "eval_sts-test_spearman_max": 0.8731698752972269, + "eval_vitaminc-pairs_loss": 3.929328203201294, + "eval_vitaminc-pairs_runtime": 5.4635, + "eval_vitaminc-pairs_samples_per_second": 23.428, + "eval_vitaminc-pairs_steps_per_second": 0.366, + "step": 1620 + }, + { + "epoch": 1.256788207913111, + "eval_negation-triplets_loss": 1.6538103818893433, + "eval_negation-triplets_runtime": 1.0822, + "eval_negation-triplets_samples_per_second": 118.274, + "eval_negation-triplets_steps_per_second": 1.848, + "step": 1620 + }, + { + "epoch": 1.256788207913111, + "eval_scitail-pairs-pos_loss": 0.14308230578899384, + "eval_scitail-pairs-pos_runtime": 1.2969, + "eval_scitail-pairs-pos_samples_per_second": 98.698, + "eval_scitail-pairs-pos_steps_per_second": 1.542, + "step": 1620 + }, + { + "epoch": 1.256788207913111, + "eval_scitail-pairs-qa_loss": 0.0069819046184420586, + "eval_scitail-pairs-qa_runtime": 0.7754, + "eval_scitail-pairs-qa_samples_per_second": 165.08, + "eval_scitail-pairs-qa_steps_per_second": 2.579, + "step": 1620 + }, + { + "epoch": 1.256788207913111, + "eval_xsum-pairs_loss": 0.11357403546571732, + "eval_xsum-pairs_runtime": 3.9172, + "eval_xsum-pairs_samples_per_second": 32.677, + "eval_xsum-pairs_steps_per_second": 0.511, + "step": 1620 + }, + { + "epoch": 1.256788207913111, + "eval_sciq_pairs_loss": 0.047332148998975754, + "eval_sciq_pairs_runtime": 6.1653, + "eval_sciq_pairs_samples_per_second": 20.762, + "eval_sciq_pairs_steps_per_second": 0.324, + "step": 1620 + }, + { + "epoch": 1.256788207913111, + "eval_qasc_pairs_loss": 0.23769663274288177, + "eval_qasc_pairs_runtime": 0.9284, + "eval_qasc_pairs_samples_per_second": 137.872, + "eval_qasc_pairs_steps_per_second": 2.154, + "step": 1620 + }, + { + "epoch": 1.256788207913111, + "eval_openbookqa_pairs_loss": 1.092557668685913, + "eval_openbookqa_pairs_runtime": 0.8809, + "eval_openbookqa_pairs_samples_per_second": 145.301, + "eval_openbookqa_pairs_steps_per_second": 2.27, + "step": 1620 + }, + { + "epoch": 1.256788207913111, + "eval_msmarco_pairs_loss": 0.38742268085479736, + "eval_msmarco_pairs_runtime": 1.9706, + "eval_msmarco_pairs_samples_per_second": 64.954, + "eval_msmarco_pairs_steps_per_second": 1.015, + "step": 1620 + }, + { + "epoch": 1.256788207913111, + "eval_nq_pairs_loss": 0.5972813963890076, + "eval_nq_pairs_runtime": 3.3355, + "eval_nq_pairs_samples_per_second": 38.375, + "eval_nq_pairs_steps_per_second": 0.6, + "step": 1620 + }, + { + "epoch": 1.256788207913111, + "eval_trivia_pairs_loss": 0.726773738861084, + "eval_trivia_pairs_runtime": 4.4938, + "eval_trivia_pairs_samples_per_second": 28.484, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 1620 + }, + { + "epoch": 1.256788207913111, + "eval_gooaq_pairs_loss": 0.5606580376625061, + "eval_gooaq_pairs_runtime": 1.4685, + "eval_gooaq_pairs_samples_per_second": 87.165, + "eval_gooaq_pairs_steps_per_second": 1.362, + "step": 1620 + }, + { + "epoch": 1.256788207913111, + "eval_paws-pos_loss": 0.032778069376945496, + "eval_paws-pos_runtime": 1.0409, + "eval_paws-pos_samples_per_second": 122.972, + "eval_paws-pos_steps_per_second": 1.921, + "step": 1620 + }, + { + "epoch": 1.256788207913111, + "eval_global_dataset_loss": 0.42444390058517456, + "eval_global_dataset_runtime": 19.436, + "eval_global_dataset_samples_per_second": 21.404, + "eval_global_dataset_steps_per_second": 0.36, + "step": 1620 + }, + { + "epoch": 1.2575640031031807, + "grad_norm": 0.6430074572563171, + "learning_rate": 2.2000632741726872e-05, + "loss": 0.0153, + "step": 1621 + }, + { + "epoch": 1.2583397982932505, + "grad_norm": 5.4249267578125, + "learning_rate": 2.197241763151652e-05, + "loss": 0.147, + "step": 1622 + }, + { + "epoch": 1.2591155934833205, + "grad_norm": 9.37018871307373, + "learning_rate": 2.1944193512449055e-05, + "loss": 0.789, + "step": 1623 + }, + { + "epoch": 1.2598913886733902, + "grad_norm": 4.03061580657959, + "learning_rate": 2.1915960477921975e-05, + "loss": 0.1587, + "step": 1624 + }, + { + "epoch": 1.26066718386346, + "grad_norm": 5.177279949188232, + "learning_rate": 2.188771862136232e-05, + "loss": 0.1313, + "step": 1625 + }, + { + "epoch": 1.26144297905353, + "grad_norm": 6.746555805206299, + "learning_rate": 2.18594680362263e-05, + "loss": 0.2186, + "step": 1626 + }, + { + "epoch": 1.2622187742435997, + "grad_norm": 4.807129859924316, + "learning_rate": 2.183120881599902e-05, + "loss": 0.1917, + "step": 1627 + }, + { + "epoch": 1.2629945694336695, + "grad_norm": 0.7335848808288574, + "learning_rate": 2.1802941054194143e-05, + "loss": 0.0142, + "step": 1628 + }, + { + "epoch": 1.2637703646237393, + "grad_norm": 10.189105033874512, + "learning_rate": 2.1774664844353618e-05, + "loss": 0.5545, + "step": 1629 + }, + { + "epoch": 1.264546159813809, + "grad_norm": 6.414772033691406, + "learning_rate": 2.174638028004735e-05, + "loss": 0.159, + "step": 1630 + }, + { + "epoch": 1.265321955003879, + "grad_norm": 0.6904690265655518, + "learning_rate": 2.1718087454872863e-05, + "loss": 0.0224, + "step": 1631 + }, + { + "epoch": 1.2660977501939488, + "grad_norm": 0.6443296670913696, + "learning_rate": 2.168978646245504e-05, + "loss": 0.0089, + "step": 1632 + }, + { + "epoch": 1.2668735453840188, + "grad_norm": 6.721205711364746, + "learning_rate": 2.16614773964458e-05, + "loss": 0.2194, + "step": 1633 + }, + { + "epoch": 1.2676493405740885, + "grad_norm": 9.069674491882324, + "learning_rate": 2.1633160350523744e-05, + "loss": 0.4266, + "step": 1634 + }, + { + "epoch": 1.2684251357641583, + "grad_norm": 12.271734237670898, + "learning_rate": 2.160483541839391e-05, + "loss": 0.8429, + "step": 1635 + }, + { + "epoch": 1.269200930954228, + "grad_norm": 6.057973861694336, + "learning_rate": 2.157650269378743e-05, + "loss": 0.1864, + "step": 1636 + }, + { + "epoch": 1.2699767261442978, + "grad_norm": 0.7014800310134888, + "learning_rate": 2.1548162270461202e-05, + "loss": 0.0146, + "step": 1637 + }, + { + "epoch": 1.2707525213343678, + "grad_norm": 0.4622633755207062, + "learning_rate": 2.151981424219762e-05, + "loss": 0.0058, + "step": 1638 + }, + { + "epoch": 1.2715283165244375, + "grad_norm": 7.871875286102295, + "learning_rate": 2.1491458702804244e-05, + "loss": 0.2992, + "step": 1639 + }, + { + "epoch": 1.2723041117145073, + "grad_norm": 7.959434986114502, + "learning_rate": 2.1463095746113465e-05, + "loss": 0.3364, + "step": 1640 + }, + { + "epoch": 1.2723041117145073, + "eval_Qnli-dev_cosine_accuracy": 0.6875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.7844488024711609, + "eval_Qnli-dev_cosine_ap": 0.7144696302453439, + "eval_Qnli-dev_cosine_f1": 0.6865671641791046, + "eval_Qnli-dev_cosine_f1_threshold": 0.685598611831665, + "eval_Qnli-dev_cosine_precision": 0.5640326975476839, + "eval_Qnli-dev_cosine_recall": 0.8771186440677966, + "eval_Qnli-dev_dot_accuracy": 0.666015625, + "eval_Qnli-dev_dot_accuracy_threshold": 351.26055908203125, + "eval_Qnli-dev_dot_ap": 0.6379710624634651, + "eval_Qnli-dev_dot_f1": 0.6759259259259259, + "eval_Qnli-dev_dot_f1_threshold": 294.90838623046875, + "eval_Qnli-dev_dot_precision": 0.5315533980582524, + "eval_Qnli-dev_dot_recall": 0.9279661016949152, + "eval_Qnli-dev_euclidean_accuracy": 0.68359375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 14.158502578735352, + "eval_Qnli-dev_euclidean_ap": 0.7184214116924033, + "eval_Qnli-dev_euclidean_f1": 0.6885245901639343, + "eval_Qnli-dev_euclidean_f1_threshold": 15.842201232910156, + "eval_Qnli-dev_euclidean_precision": 0.6038338658146964, + "eval_Qnli-dev_euclidean_recall": 0.8008474576271186, + "eval_Qnli-dev_manhattan_accuracy": 0.6875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 256.33404541015625, + "eval_Qnli-dev_manhattan_ap": 0.7194716517037396, + "eval_Qnli-dev_manhattan_f1": 0.6788685524126457, + "eval_Qnli-dev_manhattan_f1_threshold": 311.5912170410156, + "eval_Qnli-dev_manhattan_precision": 0.5589041095890411, + "eval_Qnli-dev_manhattan_recall": 0.864406779661017, + "eval_Qnli-dev_max_accuracy": 0.6875, + "eval_Qnli-dev_max_accuracy_threshold": 351.26055908203125, + "eval_Qnli-dev_max_ap": 0.7194716517037396, + "eval_Qnli-dev_max_f1": 0.6885245901639343, + "eval_Qnli-dev_max_f1_threshold": 311.5912170410156, + "eval_Qnli-dev_max_precision": 0.6038338658146964, + "eval_Qnli-dev_max_recall": 0.9279661016949152, + "eval_allNLI-dev_cosine_accuracy": 0.73046875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8605382442474365, + "eval_allNLI-dev_cosine_ap": 0.6031263949390056, + "eval_allNLI-dev_cosine_f1": 0.6066350710900473, + "eval_allNLI-dev_cosine_f1_threshold": 0.8008022308349609, + "eval_allNLI-dev_cosine_precision": 0.5140562248995983, + "eval_allNLI-dev_cosine_recall": 0.7398843930635838, + "eval_allNLI-dev_dot_accuracy": 0.685546875, + "eval_allNLI-dev_dot_accuracy_threshold": 397.16717529296875, + "eval_allNLI-dev_dot_ap": 0.5169942060670281, + "eval_allNLI-dev_dot_f1": 0.5754527162977867, + "eval_allNLI-dev_dot_f1_threshold": 328.2814025878906, + "eval_allNLI-dev_dot_precision": 0.44135802469135804, + "eval_allNLI-dev_dot_recall": 0.8265895953757225, + "eval_allNLI-dev_euclidean_accuracy": 0.732421875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.462906837463379, + "eval_allNLI-dev_euclidean_ap": 0.6068722679181434, + "eval_allNLI-dev_euclidean_f1": 0.620985010706638, + "eval_allNLI-dev_euclidean_f1_threshold": 14.440187454223633, + "eval_allNLI-dev_euclidean_precision": 0.4931972789115646, + "eval_allNLI-dev_euclidean_recall": 0.838150289017341, + "eval_allNLI-dev_manhattan_accuracy": 0.7265625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 202.6051788330078, + "eval_allNLI-dev_manhattan_ap": 0.6072831653578705, + "eval_allNLI-dev_manhattan_f1": 0.619047619047619, + "eval_allNLI-dev_manhattan_f1_threshold": 243.4591827392578, + "eval_allNLI-dev_manhattan_precision": 0.5263157894736842, + "eval_allNLI-dev_manhattan_recall": 0.7514450867052023, + "eval_allNLI-dev_max_accuracy": 0.732421875, + "eval_allNLI-dev_max_accuracy_threshold": 397.16717529296875, + "eval_allNLI-dev_max_ap": 0.6072831653578705, + "eval_allNLI-dev_max_f1": 0.620985010706638, + "eval_allNLI-dev_max_f1_threshold": 328.2814025878906, + "eval_allNLI-dev_max_precision": 0.5263157894736842, + "eval_allNLI-dev_max_recall": 0.838150289017341, + "eval_sequential_score": 0.7194716517037396, + "eval_sts-test_pearson_cosine": 0.8365480314211637, + "eval_sts-test_pearson_dot": 0.7930315027368717, + "eval_sts-test_pearson_euclidean": 0.8715333460921924, + "eval_sts-test_pearson_manhattan": 0.8693715361792872, + "eval_sts-test_pearson_max": 0.8715333460921924, + "eval_sts-test_spearman_cosine": 0.8755303253305776, + "eval_sts-test_spearman_dot": 0.7862886576504292, + "eval_sts-test_spearman_euclidean": 0.8712485246406485, + "eval_sts-test_spearman_manhattan": 0.8707917037643651, + "eval_sts-test_spearman_max": 0.8755303253305776, + "eval_vitaminc-pairs_loss": 3.9037399291992188, + "eval_vitaminc-pairs_runtime": 5.4533, + "eval_vitaminc-pairs_samples_per_second": 23.472, + "eval_vitaminc-pairs_steps_per_second": 0.367, + "step": 1640 + }, + { + "epoch": 1.2723041117145073, + "eval_negation-triplets_loss": 1.679166316986084, + "eval_negation-triplets_runtime": 1.0937, + "eval_negation-triplets_samples_per_second": 117.029, + "eval_negation-triplets_steps_per_second": 1.829, + "step": 1640 + }, + { + "epoch": 1.2723041117145073, + "eval_scitail-pairs-pos_loss": 0.13054484128952026, + "eval_scitail-pairs-pos_runtime": 1.2953, + "eval_scitail-pairs-pos_samples_per_second": 98.815, + "eval_scitail-pairs-pos_steps_per_second": 1.544, + "step": 1640 + }, + { + "epoch": 1.2723041117145073, + "eval_scitail-pairs-qa_loss": 0.003328065387904644, + "eval_scitail-pairs-qa_runtime": 0.7765, + "eval_scitail-pairs-qa_samples_per_second": 164.85, + "eval_scitail-pairs-qa_steps_per_second": 2.576, + "step": 1640 + }, + { + "epoch": 1.2723041117145073, + "eval_xsum-pairs_loss": 0.11773895472288132, + "eval_xsum-pairs_runtime": 3.9132, + "eval_xsum-pairs_samples_per_second": 32.71, + "eval_xsum-pairs_steps_per_second": 0.511, + "step": 1640 + }, + { + "epoch": 1.2723041117145073, + "eval_sciq_pairs_loss": 0.0565481036901474, + "eval_sciq_pairs_runtime": 6.1128, + "eval_sciq_pairs_samples_per_second": 20.94, + "eval_sciq_pairs_steps_per_second": 0.327, + "step": 1640 + }, + { + "epoch": 1.2723041117145073, + "eval_qasc_pairs_loss": 0.23977893590927124, + "eval_qasc_pairs_runtime": 0.9251, + "eval_qasc_pairs_samples_per_second": 138.367, + "eval_qasc_pairs_steps_per_second": 2.162, + "step": 1640 + }, + { + "epoch": 1.2723041117145073, + "eval_openbookqa_pairs_loss": 1.1295928955078125, + "eval_openbookqa_pairs_runtime": 0.8783, + "eval_openbookqa_pairs_samples_per_second": 145.737, + "eval_openbookqa_pairs_steps_per_second": 2.277, + "step": 1640 + }, + { + "epoch": 1.2723041117145073, + "eval_msmarco_pairs_loss": 0.38614338636398315, + "eval_msmarco_pairs_runtime": 1.971, + "eval_msmarco_pairs_samples_per_second": 64.942, + "eval_msmarco_pairs_steps_per_second": 1.015, + "step": 1640 + }, + { + "epoch": 1.2723041117145073, + "eval_nq_pairs_loss": 0.6011468172073364, + "eval_nq_pairs_runtime": 3.3396, + "eval_nq_pairs_samples_per_second": 38.328, + "eval_nq_pairs_steps_per_second": 0.599, + "step": 1640 + }, + { + "epoch": 1.2723041117145073, + "eval_trivia_pairs_loss": 0.8389471173286438, + "eval_trivia_pairs_runtime": 4.5009, + "eval_trivia_pairs_samples_per_second": 28.439, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 1640 + }, + { + "epoch": 1.2723041117145073, + "eval_gooaq_pairs_loss": 0.6536858081817627, + "eval_gooaq_pairs_runtime": 1.4742, + "eval_gooaq_pairs_samples_per_second": 86.829, + "eval_gooaq_pairs_steps_per_second": 1.357, + "step": 1640 + }, + { + "epoch": 1.2723041117145073, + "eval_paws-pos_loss": 0.031681545078754425, + "eval_paws-pos_runtime": 1.0438, + "eval_paws-pos_samples_per_second": 122.634, + "eval_paws-pos_steps_per_second": 1.916, + "step": 1640 + }, + { + "epoch": 1.2723041117145073, + "eval_global_dataset_loss": 0.4345376491546631, + "eval_global_dataset_runtime": 19.4239, + "eval_global_dataset_samples_per_second": 21.417, + "eval_global_dataset_steps_per_second": 0.36, + "step": 1640 + }, + { + "epoch": 1.2730799069045773, + "grad_norm": 6.540362358093262, + "learning_rate": 2.1434725465982252e-05, + "loss": 0.2258, + "step": 1641 + }, + { + "epoch": 1.273855702094647, + "grad_norm": 1.0891000032424927, + "learning_rate": 2.140634795629178e-05, + "loss": 0.0175, + "step": 1642 + }, + { + "epoch": 1.2746314972847168, + "grad_norm": 0.13839983940124512, + "learning_rate": 2.137796331094717e-05, + "loss": 0.0021, + "step": 1643 + }, + { + "epoch": 1.2754072924747866, + "grad_norm": 9.78063678741455, + "learning_rate": 2.1349571623877144e-05, + "loss": 0.444, + "step": 1644 + }, + { + "epoch": 1.2761830876648566, + "grad_norm": 7.0321149826049805, + "learning_rate": 2.1321172989033727e-05, + "loss": 0.277, + "step": 1645 + }, + { + "epoch": 1.2769588828549263, + "grad_norm": 9.730364799499512, + "learning_rate": 2.1292767500391927e-05, + "loss": 0.6617, + "step": 1646 + }, + { + "epoch": 1.277734678044996, + "grad_norm": 5.500321865081787, + "learning_rate": 2.1264355251949462e-05, + "loss": 0.2703, + "step": 1647 + }, + { + "epoch": 1.278510473235066, + "grad_norm": 6.235525131225586, + "learning_rate": 2.123593633772638e-05, + "loss": 0.4062, + "step": 1648 + }, + { + "epoch": 1.2792862684251358, + "grad_norm": 3.4607746601104736, + "learning_rate": 2.1207510851764816e-05, + "loss": 0.0847, + "step": 1649 + }, + { + "epoch": 1.2800620636152056, + "grad_norm": 12.611079216003418, + "learning_rate": 2.1179078888128636e-05, + "loss": 0.8878, + "step": 1650 + }, + { + "epoch": 1.2808378588052753, + "grad_norm": 0.5125525593757629, + "learning_rate": 2.1150640540903144e-05, + "loss": 0.0281, + "step": 1651 + }, + { + "epoch": 1.281613653995345, + "grad_norm": 0.08136854320764542, + "learning_rate": 2.112219590419478e-05, + "loss": 0.0008, + "step": 1652 + }, + { + "epoch": 1.282389449185415, + "grad_norm": 7.408838748931885, + "learning_rate": 2.109374507213079e-05, + "loss": 0.2374, + "step": 1653 + }, + { + "epoch": 1.2831652443754848, + "grad_norm": 4.708815574645996, + "learning_rate": 2.10652881388589e-05, + "loss": 0.1523, + "step": 1654 + }, + { + "epoch": 1.2839410395655548, + "grad_norm": 12.45781421661377, + "learning_rate": 2.103682519854706e-05, + "loss": 0.8189, + "step": 1655 + }, + { + "epoch": 1.2847168347556246, + "grad_norm": 9.500694274902344, + "learning_rate": 2.1008356345383085e-05, + "loss": 0.4487, + "step": 1656 + }, + { + "epoch": 1.2854926299456944, + "grad_norm": 7.158897876739502, + "learning_rate": 2.0979881673574345e-05, + "loss": 0.3432, + "step": 1657 + }, + { + "epoch": 1.2862684251357641, + "grad_norm": 12.860493659973145, + "learning_rate": 2.095140127734748e-05, + "loss": 0.7214, + "step": 1658 + }, + { + "epoch": 1.2870442203258339, + "grad_norm": 9.061240196228027, + "learning_rate": 2.0922915250948066e-05, + "loss": 0.6608, + "step": 1659 + }, + { + "epoch": 1.2878200155159039, + "grad_norm": 9.543268203735352, + "learning_rate": 2.089442368864031e-05, + "loss": 0.7502, + "step": 1660 + }, + { + "epoch": 1.2878200155159039, + "eval_Qnli-dev_cosine_accuracy": 0.689453125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.7893164157867432, + "eval_Qnli-dev_cosine_ap": 0.7152658140564143, + "eval_Qnli-dev_cosine_f1": 0.6903225806451613, + "eval_Qnli-dev_cosine_f1_threshold": 0.7037264108657837, + "eval_Qnli-dev_cosine_precision": 0.5572916666666666, + "eval_Qnli-dev_cosine_recall": 0.9067796610169492, + "eval_Qnli-dev_dot_accuracy": 0.66796875, + "eval_Qnli-dev_dot_accuracy_threshold": 372.81182861328125, + "eval_Qnli-dev_dot_ap": 0.6297259199074229, + "eval_Qnli-dev_dot_f1": 0.67430441898527, + "eval_Qnli-dev_dot_f1_threshold": 318.84478759765625, + "eval_Qnli-dev_dot_precision": 0.5493333333333333, + "eval_Qnli-dev_dot_recall": 0.8728813559322034, + "eval_Qnli-dev_euclidean_accuracy": 0.6953125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 14.50367546081543, + "eval_Qnli-dev_euclidean_ap": 0.720908585372269, + "eval_Qnli-dev_euclidean_f1": 0.6949429037520392, + "eval_Qnli-dev_euclidean_f1_threshold": 16.626895904541016, + "eval_Qnli-dev_euclidean_precision": 0.5649867374005305, + "eval_Qnli-dev_euclidean_recall": 0.902542372881356, + "eval_Qnli-dev_manhattan_accuracy": 0.697265625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 254.69720458984375, + "eval_Qnli-dev_manhattan_ap": 0.7207542467826241, + "eval_Qnli-dev_manhattan_f1": 0.6843853820598007, + "eval_Qnli-dev_manhattan_f1_threshold": 298.4431457519531, + "eval_Qnli-dev_manhattan_precision": 0.5628415300546448, + "eval_Qnli-dev_manhattan_recall": 0.8728813559322034, + "eval_Qnli-dev_max_accuracy": 0.697265625, + "eval_Qnli-dev_max_accuracy_threshold": 372.81182861328125, + "eval_Qnli-dev_max_ap": 0.720908585372269, + "eval_Qnli-dev_max_f1": 0.6949429037520392, + "eval_Qnli-dev_max_f1_threshold": 318.84478759765625, + "eval_Qnli-dev_max_precision": 0.5649867374005305, + "eval_Qnli-dev_max_recall": 0.9067796610169492, + "eval_allNLI-dev_cosine_accuracy": 0.728515625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8691446781158447, + "eval_allNLI-dev_cosine_ap": 0.6146328757995156, + "eval_allNLI-dev_cosine_f1": 0.6170212765957447, + "eval_allNLI-dev_cosine_f1_threshold": 0.7771510481834412, + "eval_allNLI-dev_cosine_precision": 0.4882154882154882, + "eval_allNLI-dev_cosine_recall": 0.838150289017341, + "eval_allNLI-dev_dot_accuracy": 0.6875, + "eval_allNLI-dev_dot_accuracy_threshold": 388.3375244140625, + "eval_allNLI-dev_dot_ap": 0.5250854180136727, + "eval_allNLI-dev_dot_f1": 0.5843230403800475, + "eval_allNLI-dev_dot_f1_threshold": 356.353759765625, + "eval_allNLI-dev_dot_precision": 0.4959677419354839, + "eval_allNLI-dev_dot_recall": 0.7109826589595376, + "eval_allNLI-dev_euclidean_accuracy": 0.736328125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.17059326171875, + "eval_allNLI-dev_euclidean_ap": 0.6162793436822088, + "eval_allNLI-dev_euclidean_f1": 0.6247379454926625, + "eval_allNLI-dev_euclidean_f1_threshold": 14.343178749084473, + "eval_allNLI-dev_euclidean_precision": 0.4901315789473684, + "eval_allNLI-dev_euclidean_recall": 0.861271676300578, + "eval_allNLI-dev_manhattan_accuracy": 0.734375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 204.16429138183594, + "eval_allNLI-dev_manhattan_ap": 0.6186489686425896, + "eval_allNLI-dev_manhattan_f1": 0.6276595744680851, + "eval_allNLI-dev_manhattan_f1_threshold": 223.9955291748047, + "eval_allNLI-dev_manhattan_precision": 0.5812807881773399, + "eval_allNLI-dev_manhattan_recall": 0.6820809248554913, + "eval_allNLI-dev_max_accuracy": 0.736328125, + "eval_allNLI-dev_max_accuracy_threshold": 388.3375244140625, + "eval_allNLI-dev_max_ap": 0.6186489686425896, + "eval_allNLI-dev_max_f1": 0.6276595744680851, + "eval_allNLI-dev_max_f1_threshold": 356.353759765625, + "eval_allNLI-dev_max_precision": 0.5812807881773399, + "eval_allNLI-dev_max_recall": 0.861271676300578, + "eval_sequential_score": 0.720908585372269, + "eval_sts-test_pearson_cosine": 0.8293849330398687, + "eval_sts-test_pearson_dot": 0.7732109066136641, + "eval_sts-test_pearson_euclidean": 0.8681586553447108, + "eval_sts-test_pearson_manhattan": 0.8665268616013031, + "eval_sts-test_pearson_max": 0.8681586553447108, + "eval_sts-test_spearman_cosine": 0.873892742422482, + "eval_sts-test_spearman_dot": 0.7674105170375128, + "eval_sts-test_spearman_euclidean": 0.8707063631088918, + "eval_sts-test_spearman_manhattan": 0.8702443063350255, + "eval_sts-test_spearman_max": 0.873892742422482, + "eval_vitaminc-pairs_loss": 3.828678607940674, + "eval_vitaminc-pairs_runtime": 5.4857, + "eval_vitaminc-pairs_samples_per_second": 23.333, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 1660 + }, + { + "epoch": 1.2878200155159039, + "eval_negation-triplets_loss": 1.6117511987686157, + "eval_negation-triplets_runtime": 1.0761, + "eval_negation-triplets_samples_per_second": 118.947, + "eval_negation-triplets_steps_per_second": 1.859, + "step": 1660 + }, + { + "epoch": 1.2878200155159039, + "eval_scitail-pairs-pos_loss": 0.12270583212375641, + "eval_scitail-pairs-pos_runtime": 1.2994, + "eval_scitail-pairs-pos_samples_per_second": 98.508, + "eval_scitail-pairs-pos_steps_per_second": 1.539, + "step": 1660 + }, + { + "epoch": 1.2878200155159039, + "eval_scitail-pairs-qa_loss": 0.0032372463028877974, + "eval_scitail-pairs-qa_runtime": 0.783, + "eval_scitail-pairs-qa_samples_per_second": 163.479, + "eval_scitail-pairs-qa_steps_per_second": 2.554, + "step": 1660 + }, + { + "epoch": 1.2878200155159039, + "eval_xsum-pairs_loss": 0.10785119235515594, + "eval_xsum-pairs_runtime": 3.9258, + "eval_xsum-pairs_samples_per_second": 32.605, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 1660 + }, + { + "epoch": 1.2878200155159039, + "eval_sciq_pairs_loss": 0.05435860529541969, + "eval_sciq_pairs_runtime": 6.1249, + "eval_sciq_pairs_samples_per_second": 20.898, + "eval_sciq_pairs_steps_per_second": 0.327, + "step": 1660 + }, + { + "epoch": 1.2878200155159039, + "eval_qasc_pairs_loss": 0.2630281150341034, + "eval_qasc_pairs_runtime": 0.9235, + "eval_qasc_pairs_samples_per_second": 138.603, + "eval_qasc_pairs_steps_per_second": 2.166, + "step": 1660 + }, + { + "epoch": 1.2878200155159039, + "eval_openbookqa_pairs_loss": 1.1082996129989624, + "eval_openbookqa_pairs_runtime": 0.8776, + "eval_openbookqa_pairs_samples_per_second": 145.848, + "eval_openbookqa_pairs_steps_per_second": 2.279, + "step": 1660 + }, + { + "epoch": 1.2878200155159039, + "eval_msmarco_pairs_loss": 0.3556932806968689, + "eval_msmarco_pairs_runtime": 1.9756, + "eval_msmarco_pairs_samples_per_second": 64.791, + "eval_msmarco_pairs_steps_per_second": 1.012, + "step": 1660 + }, + { + "epoch": 1.2878200155159039, + "eval_nq_pairs_loss": 0.5807768106460571, + "eval_nq_pairs_runtime": 3.3392, + "eval_nq_pairs_samples_per_second": 38.333, + "eval_nq_pairs_steps_per_second": 0.599, + "step": 1660 + }, + { + "epoch": 1.2878200155159039, + "eval_trivia_pairs_loss": 0.7337329983711243, + "eval_trivia_pairs_runtime": 4.4877, + "eval_trivia_pairs_samples_per_second": 28.522, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 1660 + }, + { + "epoch": 1.2878200155159039, + "eval_gooaq_pairs_loss": 0.6279823184013367, + "eval_gooaq_pairs_runtime": 1.4852, + "eval_gooaq_pairs_samples_per_second": 86.181, + "eval_gooaq_pairs_steps_per_second": 1.347, + "step": 1660 + }, + { + "epoch": 1.2878200155159039, + "eval_paws-pos_loss": 0.031691815704107285, + "eval_paws-pos_runtime": 1.0387, + "eval_paws-pos_samples_per_second": 123.225, + "eval_paws-pos_steps_per_second": 1.925, + "step": 1660 + }, + { + "epoch": 1.2878200155159039, + "eval_global_dataset_loss": 0.4095148742198944, + "eval_global_dataset_runtime": 19.422, + "eval_global_dataset_samples_per_second": 21.419, + "eval_global_dataset_steps_per_second": 0.36, + "step": 1660 + }, + { + "epoch": 1.2885958107059736, + "grad_norm": 7.454000473022461, + "learning_rate": 2.086592668470674e-05, + "loss": 0.3143, + "step": 1661 + }, + { + "epoch": 1.2893716058960434, + "grad_norm": 0.10004957020282745, + "learning_rate": 2.0837424333447883e-05, + "loss": 0.0012, + "step": 1662 + }, + { + "epoch": 1.2901474010861134, + "grad_norm": 8.724870681762695, + "learning_rate": 2.0808916729181984e-05, + "loss": 0.4845, + "step": 1663 + }, + { + "epoch": 1.2909231962761831, + "grad_norm": 6.756128311157227, + "learning_rate": 2.0780403966244645e-05, + "loss": 0.2046, + "step": 1664 + }, + { + "epoch": 1.2916989914662529, + "grad_norm": 13.086095809936523, + "learning_rate": 2.0751886138988544e-05, + "loss": 1.1954, + "step": 1665 + }, + { + "epoch": 1.2924747866563226, + "grad_norm": 7.50347375869751, + "learning_rate": 2.072336334178313e-05, + "loss": 0.4912, + "step": 1666 + }, + { + "epoch": 1.2932505818463926, + "grad_norm": 9.048041343688965, + "learning_rate": 2.0694835669014292e-05, + "loss": 0.7404, + "step": 1667 + }, + { + "epoch": 1.2940263770364624, + "grad_norm": 3.8070995807647705, + "learning_rate": 2.0666303215084044e-05, + "loss": 0.1908, + "step": 1668 + }, + { + "epoch": 1.2948021722265322, + "grad_norm": 4.572576999664307, + "learning_rate": 2.063776607441024e-05, + "loss": 0.33, + "step": 1669 + }, + { + "epoch": 1.2955779674166021, + "grad_norm": 3.6509835720062256, + "learning_rate": 2.0609224341426223e-05, + "loss": 0.0908, + "step": 1670 + }, + { + "epoch": 1.296353762606672, + "grad_norm": 5.7009358406066895, + "learning_rate": 2.0580678110580542e-05, + "loss": 0.1861, + "step": 1671 + }, + { + "epoch": 1.2971295577967417, + "grad_norm": 4.028733253479004, + "learning_rate": 2.0552127476336638e-05, + "loss": 0.2022, + "step": 1672 + }, + { + "epoch": 1.2979053529868114, + "grad_norm": 5.880012035369873, + "learning_rate": 2.0523572533172503e-05, + "loss": 0.2413, + "step": 1673 + }, + { + "epoch": 1.2986811481768812, + "grad_norm": 6.58522367477417, + "learning_rate": 2.049501337558041e-05, + "loss": 0.3496, + "step": 1674 + }, + { + "epoch": 1.2994569433669512, + "grad_norm": 8.29002571105957, + "learning_rate": 2.0466450098066565e-05, + "loss": 0.6232, + "step": 1675 + }, + { + "epoch": 1.300232738557021, + "grad_norm": 8.405040740966797, + "learning_rate": 2.0437882795150814e-05, + "loss": 0.7277, + "step": 1676 + }, + { + "epoch": 1.3010085337470907, + "grad_norm": 6.34568977355957, + "learning_rate": 2.0409311561366313e-05, + "loss": 0.3247, + "step": 1677 + }, + { + "epoch": 1.3017843289371607, + "grad_norm": 8.699254989624023, + "learning_rate": 2.0380736491259236e-05, + "loss": 0.6035, + "step": 1678 + }, + { + "epoch": 1.3025601241272304, + "grad_norm": 4.6270575523376465, + "learning_rate": 2.0352157679388454e-05, + "loss": 0.1986, + "step": 1679 + }, + { + "epoch": 1.3033359193173002, + "grad_norm": 12.221156120300293, + "learning_rate": 2.032357522032521e-05, + "loss": 1.3264, + "step": 1680 + }, + { + "epoch": 1.3033359193173002, + "eval_Qnli-dev_cosine_accuracy": 0.6796875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8367424607276917, + "eval_Qnli-dev_cosine_ap": 0.7113915778622408, + "eval_Qnli-dev_cosine_f1": 0.684297520661157, + "eval_Qnli-dev_cosine_f1_threshold": 0.7372596263885498, + "eval_Qnli-dev_cosine_precision": 0.5609756097560976, + "eval_Qnli-dev_cosine_recall": 0.8771186440677966, + "eval_Qnli-dev_dot_accuracy": 0.673828125, + "eval_Qnli-dev_dot_accuracy_threshold": 407.44366455078125, + "eval_Qnli-dev_dot_ap": 0.6277581156449731, + "eval_Qnli-dev_dot_f1": 0.6707317073170731, + "eval_Qnli-dev_dot_f1_threshold": 330.45330810546875, + "eval_Qnli-dev_dot_precision": 0.5238095238095238, + "eval_Qnli-dev_dot_recall": 0.9322033898305084, + "eval_Qnli-dev_euclidean_accuracy": 0.67578125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.505901336669922, + "eval_Qnli-dev_euclidean_ap": 0.7130026325034099, + "eval_Qnli-dev_euclidean_f1": 0.6882067851373183, + "eval_Qnli-dev_euclidean_f1_threshold": 16.273117065429688, + "eval_Qnli-dev_euclidean_precision": 0.556135770234987, + "eval_Qnli-dev_euclidean_recall": 0.902542372881356, + "eval_Qnli-dev_manhattan_accuracy": 0.66796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 213.2294921875, + "eval_Qnli-dev_manhattan_ap": 0.7115075855960601, + "eval_Qnli-dev_manhattan_f1": 0.6782334384858044, + "eval_Qnli-dev_manhattan_f1_threshold": 301.8433837890625, + "eval_Qnli-dev_manhattan_precision": 0.5402010050251256, + "eval_Qnli-dev_manhattan_recall": 0.9110169491525424, + "eval_Qnli-dev_max_accuracy": 0.6796875, + "eval_Qnli-dev_max_accuracy_threshold": 407.44366455078125, + "eval_Qnli-dev_max_ap": 0.7130026325034099, + "eval_Qnli-dev_max_f1": 0.6882067851373183, + "eval_Qnli-dev_max_f1_threshold": 330.45330810546875, + "eval_Qnli-dev_max_precision": 0.5609756097560976, + "eval_Qnli-dev_max_recall": 0.9322033898305084, + "eval_allNLI-dev_cosine_accuracy": 0.7265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8741326332092285, + "eval_allNLI-dev_cosine_ap": 0.6136296258353366, + "eval_allNLI-dev_cosine_f1": 0.6279069767441859, + "eval_allNLI-dev_cosine_f1_threshold": 0.8192355036735535, + "eval_allNLI-dev_cosine_precision": 0.5252918287937743, + "eval_allNLI-dev_cosine_recall": 0.7803468208092486, + "eval_allNLI-dev_dot_accuracy": 0.69140625, + "eval_allNLI-dev_dot_accuracy_threshold": 409.74066162109375, + "eval_allNLI-dev_dot_ap": 0.5097151781640198, + "eval_allNLI-dev_dot_f1": 0.5714285714285714, + "eval_allNLI-dev_dot_f1_threshold": 358.73797607421875, + "eval_allNLI-dev_dot_precision": 0.4416403785488959, + "eval_allNLI-dev_dot_recall": 0.8092485549132948, + "eval_allNLI-dev_euclidean_accuracy": 0.7265625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.764975547790527, + "eval_allNLI-dev_euclidean_ap": 0.6151729398080977, + "eval_allNLI-dev_euclidean_f1": 0.6338028169014085, + "eval_allNLI-dev_euclidean_f1_threshold": 12.948188781738281, + "eval_allNLI-dev_euclidean_precision": 0.5335968379446641, + "eval_allNLI-dev_euclidean_recall": 0.7803468208092486, + "eval_allNLI-dev_manhattan_accuracy": 0.734375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 202.09075927734375, + "eval_allNLI-dev_manhattan_ap": 0.6172933425524525, + "eval_allNLI-dev_manhattan_f1": 0.6258823529411764, + "eval_allNLI-dev_manhattan_f1_threshold": 234.14852905273438, + "eval_allNLI-dev_manhattan_precision": 0.5277777777777778, + "eval_allNLI-dev_manhattan_recall": 0.7687861271676301, + "eval_allNLI-dev_max_accuracy": 0.734375, + "eval_allNLI-dev_max_accuracy_threshold": 409.74066162109375, + "eval_allNLI-dev_max_ap": 0.6172933425524525, + "eval_allNLI-dev_max_f1": 0.6338028169014085, + "eval_allNLI-dev_max_f1_threshold": 358.73797607421875, + "eval_allNLI-dev_max_precision": 0.5335968379446641, + "eval_allNLI-dev_max_recall": 0.8092485549132948, + "eval_sequential_score": 0.7130026325034099, + "eval_sts-test_pearson_cosine": 0.8229595830408541, + "eval_sts-test_pearson_dot": 0.7796359189029936, + "eval_sts-test_pearson_euclidean": 0.860242021602312, + "eval_sts-test_pearson_manhattan": 0.8593042870595752, + "eval_sts-test_pearson_max": 0.860242021602312, + "eval_sts-test_spearman_cosine": 0.8693666445966175, + "eval_sts-test_spearman_dot": 0.775279399835531, + "eval_sts-test_spearman_euclidean": 0.863815563879844, + "eval_sts-test_spearman_manhattan": 0.8644430660670742, + "eval_sts-test_spearman_max": 0.8693666445966175, + "eval_vitaminc-pairs_loss": 3.9550254344940186, + "eval_vitaminc-pairs_runtime": 5.4548, + "eval_vitaminc-pairs_samples_per_second": 23.465, + "eval_vitaminc-pairs_steps_per_second": 0.367, + "step": 1680 + }, + { + "epoch": 1.3033359193173002, + "eval_negation-triplets_loss": 1.5367538928985596, + "eval_negation-triplets_runtime": 1.0738, + "eval_negation-triplets_samples_per_second": 119.199, + "eval_negation-triplets_steps_per_second": 1.862, + "step": 1680 + }, + { + "epoch": 1.3033359193173002, + "eval_scitail-pairs-pos_loss": 0.09544986486434937, + "eval_scitail-pairs-pos_runtime": 1.2885, + "eval_scitail-pairs-pos_samples_per_second": 99.336, + "eval_scitail-pairs-pos_steps_per_second": 1.552, + "step": 1680 + }, + { + "epoch": 1.3033359193173002, + "eval_scitail-pairs-qa_loss": 0.005798015743494034, + "eval_scitail-pairs-qa_runtime": 0.7728, + "eval_scitail-pairs-qa_samples_per_second": 165.624, + "eval_scitail-pairs-qa_steps_per_second": 2.588, + "step": 1680 + }, + { + "epoch": 1.3033359193173002, + "eval_xsum-pairs_loss": 0.08904580771923065, + "eval_xsum-pairs_runtime": 3.9238, + "eval_xsum-pairs_samples_per_second": 32.622, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 1680 + }, + { + "epoch": 1.3033359193173002, + "eval_sciq_pairs_loss": 0.04328731819987297, + "eval_sciq_pairs_runtime": 6.0733, + "eval_sciq_pairs_samples_per_second": 21.076, + "eval_sciq_pairs_steps_per_second": 0.329, + "step": 1680 + }, + { + "epoch": 1.3033359193173002, + "eval_qasc_pairs_loss": 0.25460538268089294, + "eval_qasc_pairs_runtime": 0.9218, + "eval_qasc_pairs_samples_per_second": 138.859, + "eval_qasc_pairs_steps_per_second": 2.17, + "step": 1680 + }, + { + "epoch": 1.3033359193173002, + "eval_openbookqa_pairs_loss": 1.0830907821655273, + "eval_openbookqa_pairs_runtime": 0.8856, + "eval_openbookqa_pairs_samples_per_second": 144.536, + "eval_openbookqa_pairs_steps_per_second": 2.258, + "step": 1680 + }, + { + "epoch": 1.3033359193173002, + "eval_msmarco_pairs_loss": 0.36894744634628296, + "eval_msmarco_pairs_runtime": 1.9701, + "eval_msmarco_pairs_samples_per_second": 64.971, + "eval_msmarco_pairs_steps_per_second": 1.015, + "step": 1680 + }, + { + "epoch": 1.3033359193173002, + "eval_nq_pairs_loss": 0.5505943894386292, + "eval_nq_pairs_runtime": 3.3522, + "eval_nq_pairs_samples_per_second": 38.184, + "eval_nq_pairs_steps_per_second": 0.597, + "step": 1680 + }, + { + "epoch": 1.3033359193173002, + "eval_trivia_pairs_loss": 0.718930184841156, + "eval_trivia_pairs_runtime": 4.4852, + "eval_trivia_pairs_samples_per_second": 28.539, + "eval_trivia_pairs_steps_per_second": 0.446, + "step": 1680 + }, + { + "epoch": 1.3033359193173002, + "eval_gooaq_pairs_loss": 0.569939136505127, + "eval_gooaq_pairs_runtime": 1.4622, + "eval_gooaq_pairs_samples_per_second": 87.537, + "eval_gooaq_pairs_steps_per_second": 1.368, + "step": 1680 + }, + { + "epoch": 1.3033359193173002, + "eval_paws-pos_loss": 0.03240665793418884, + "eval_paws-pos_runtime": 1.0309, + "eval_paws-pos_samples_per_second": 124.16, + "eval_paws-pos_steps_per_second": 1.94, + "step": 1680 + }, + { + "epoch": 1.3033359193173002, + "eval_global_dataset_loss": 0.3939647078514099, + "eval_global_dataset_runtime": 19.3694, + "eval_global_dataset_samples_per_second": 21.477, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1680 + }, + { + "epoch": 1.30411171450737, + "grad_norm": 3.9158437252044678, + "learning_rate": 2.0294989208652822e-05, + "loss": 0.1393, + "step": 1681 + }, + { + "epoch": 1.30488750969744, + "grad_norm": 8.6829252243042, + "learning_rate": 2.0266399738966366e-05, + "loss": 0.6284, + "step": 1682 + }, + { + "epoch": 1.3056633048875097, + "grad_norm": 5.217443943023682, + "learning_rate": 2.0237806905872346e-05, + "loss": 0.1543, + "step": 1683 + }, + { + "epoch": 1.3064391000775795, + "grad_norm": 14.338728904724121, + "learning_rate": 2.0209210803988424e-05, + "loss": 2.1294, + "step": 1684 + }, + { + "epoch": 1.3072148952676494, + "grad_norm": 11.265268325805664, + "learning_rate": 2.0180611527943057e-05, + "loss": 0.7529, + "step": 1685 + }, + { + "epoch": 1.3079906904577192, + "grad_norm": 0.3426232635974884, + "learning_rate": 2.015200917237521e-05, + "loss": 0.004, + "step": 1686 + }, + { + "epoch": 1.308766485647789, + "grad_norm": 8.815573692321777, + "learning_rate": 2.0123403831934042e-05, + "loss": 0.5542, + "step": 1687 + }, + { + "epoch": 1.3095422808378587, + "grad_norm": 0.903327465057373, + "learning_rate": 2.0094795601278588e-05, + "loss": 0.024, + "step": 1688 + }, + { + "epoch": 1.3103180760279285, + "grad_norm": 8.407374382019043, + "learning_rate": 2.006618457507745e-05, + "loss": 0.4979, + "step": 1689 + }, + { + "epoch": 1.3110938712179985, + "grad_norm": 10.85273265838623, + "learning_rate": 2.0037570848008466e-05, + "loss": 0.8255, + "step": 1690 + }, + { + "epoch": 1.3118696664080682, + "grad_norm": 3.58357834815979, + "learning_rate": 2.0008954514758428e-05, + "loss": 0.1706, + "step": 1691 + }, + { + "epoch": 1.3126454615981382, + "grad_norm": 0.7591521143913269, + "learning_rate": 1.998033567002276e-05, + "loss": 0.0582, + "step": 1692 + }, + { + "epoch": 1.313421256788208, + "grad_norm": 3.6885528564453125, + "learning_rate": 1.995171440850517e-05, + "loss": 0.0958, + "step": 1693 + }, + { + "epoch": 1.3141970519782777, + "grad_norm": 3.8304409980773926, + "learning_rate": 1.9923090824917383e-05, + "loss": 0.1162, + "step": 1694 + }, + { + "epoch": 1.3149728471683475, + "grad_norm": 6.742178440093994, + "learning_rate": 1.9894465013978803e-05, + "loss": 0.2958, + "step": 1695 + }, + { + "epoch": 1.3157486423584173, + "grad_norm": 5.418529510498047, + "learning_rate": 1.9865837070416204e-05, + "loss": 0.1517, + "step": 1696 + }, + { + "epoch": 1.3165244375484872, + "grad_norm": 7.939877510070801, + "learning_rate": 1.983720708896341e-05, + "loss": 0.4918, + "step": 1697 + }, + { + "epoch": 1.317300232738557, + "grad_norm": 11.16855525970459, + "learning_rate": 1.980857516436101e-05, + "loss": 0.6698, + "step": 1698 + }, + { + "epoch": 1.3180760279286268, + "grad_norm": 13.760820388793945, + "learning_rate": 1.9779941391355997e-05, + "loss": 2.1935, + "step": 1699 + }, + { + "epoch": 1.3188518231186968, + "grad_norm": 4.531033039093018, + "learning_rate": 1.9751305864701504e-05, + "loss": 0.1472, + "step": 1700 + }, + { + "epoch": 1.3188518231186968, + "eval_Qnli-dev_cosine_accuracy": 0.6875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8134920597076416, + "eval_Qnli-dev_cosine_ap": 0.7158159349894619, + "eval_Qnli-dev_cosine_f1": 0.6830122591943958, + "eval_Qnli-dev_cosine_f1_threshold": 0.7581806778907776, + "eval_Qnli-dev_cosine_precision": 0.582089552238806, + "eval_Qnli-dev_cosine_recall": 0.826271186440678, + "eval_Qnli-dev_dot_accuracy": 0.66015625, + "eval_Qnli-dev_dot_accuracy_threshold": 410.5783996582031, + "eval_Qnli-dev_dot_ap": 0.6332114616905549, + "eval_Qnli-dev_dot_f1": 0.6729857819905213, + "eval_Qnli-dev_dot_f1_threshold": 339.686767578125, + "eval_Qnli-dev_dot_precision": 0.5365239294710328, + "eval_Qnli-dev_dot_recall": 0.902542372881356, + "eval_Qnli-dev_euclidean_accuracy": 0.689453125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.863306045532227, + "eval_Qnli-dev_euclidean_ap": 0.7191293929528234, + "eval_Qnli-dev_euclidean_f1": 0.6861788617886179, + "eval_Qnli-dev_euclidean_f1_threshold": 16.419885635375977, + "eval_Qnli-dev_euclidean_precision": 0.5567282321899736, + "eval_Qnli-dev_euclidean_recall": 0.8940677966101694, + "eval_Qnli-dev_manhattan_accuracy": 0.6796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 214.75872802734375, + "eval_Qnli-dev_manhattan_ap": 0.7194659345147012, + "eval_Qnli-dev_manhattan_f1": 0.6821963394342762, + "eval_Qnli-dev_manhattan_f1_threshold": 292.33270263671875, + "eval_Qnli-dev_manhattan_precision": 0.5616438356164384, + "eval_Qnli-dev_manhattan_recall": 0.8686440677966102, + "eval_Qnli-dev_max_accuracy": 0.689453125, + "eval_Qnli-dev_max_accuracy_threshold": 410.5783996582031, + "eval_Qnli-dev_max_ap": 0.7194659345147012, + "eval_Qnli-dev_max_f1": 0.6861788617886179, + "eval_Qnli-dev_max_f1_threshold": 339.686767578125, + "eval_Qnli-dev_max_precision": 0.582089552238806, + "eval_Qnli-dev_max_recall": 0.902542372881356, + "eval_allNLI-dev_cosine_accuracy": 0.7265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8822686076164246, + "eval_allNLI-dev_cosine_ap": 0.6040295442883673, + "eval_allNLI-dev_cosine_f1": 0.6163793103448275, + "eval_allNLI-dev_cosine_f1_threshold": 0.796970546245575, + "eval_allNLI-dev_cosine_precision": 0.49140893470790376, + "eval_allNLI-dev_cosine_recall": 0.8265895953757225, + "eval_allNLI-dev_dot_accuracy": 0.68359375, + "eval_allNLI-dev_dot_accuracy_threshold": 460.2188720703125, + "eval_allNLI-dev_dot_ap": 0.49873651397533886, + "eval_allNLI-dev_dot_f1": 0.5719844357976653, + "eval_allNLI-dev_dot_f1_threshold": 351.12066650390625, + "eval_allNLI-dev_dot_precision": 0.4310850439882698, + "eval_allNLI-dev_dot_recall": 0.8497109826589595, + "eval_allNLI-dev_euclidean_accuracy": 0.736328125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.943103790283203, + "eval_allNLI-dev_euclidean_ap": 0.6117330781662392, + "eval_allNLI-dev_euclidean_f1": 0.6234309623430961, + "eval_allNLI-dev_euclidean_f1_threshold": 14.179935455322266, + "eval_allNLI-dev_euclidean_precision": 0.4885245901639344, + "eval_allNLI-dev_euclidean_recall": 0.861271676300578, + "eval_allNLI-dev_manhattan_accuracy": 0.73828125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 210.18309020996094, + "eval_allNLI-dev_manhattan_ap": 0.6146429057704276, + "eval_allNLI-dev_manhattan_f1": 0.6278586278586278, + "eval_allNLI-dev_manhattan_f1_threshold": 263.7506103515625, + "eval_allNLI-dev_manhattan_precision": 0.4902597402597403, + "eval_allNLI-dev_manhattan_recall": 0.8728323699421965, + "eval_allNLI-dev_max_accuracy": 0.73828125, + "eval_allNLI-dev_max_accuracy_threshold": 460.2188720703125, + "eval_allNLI-dev_max_ap": 0.6146429057704276, + "eval_allNLI-dev_max_f1": 0.6278586278586278, + "eval_allNLI-dev_max_f1_threshold": 351.12066650390625, + "eval_allNLI-dev_max_precision": 0.49140893470790376, + "eval_allNLI-dev_max_recall": 0.8728323699421965, + "eval_sequential_score": 0.7194659345147012, + "eval_sts-test_pearson_cosine": 0.8280010817911989, + "eval_sts-test_pearson_dot": 0.7700924836653391, + "eval_sts-test_pearson_euclidean": 0.8642404154211518, + "eval_sts-test_pearson_manhattan": 0.8639739966388303, + "eval_sts-test_pearson_max": 0.8642404154211518, + "eval_sts-test_spearman_cosine": 0.8719389918526563, + "eval_sts-test_spearman_dot": 0.7575245158835388, + "eval_sts-test_spearman_euclidean": 0.8660633302370113, + "eval_sts-test_spearman_manhattan": 0.8662945152533592, + "eval_sts-test_spearman_max": 0.8719389918526563, + "eval_vitaminc-pairs_loss": 3.7672722339630127, + "eval_vitaminc-pairs_runtime": 5.4502, + "eval_vitaminc-pairs_samples_per_second": 23.485, + "eval_vitaminc-pairs_steps_per_second": 0.367, + "step": 1700 + }, + { + "epoch": 1.3188518231186968, + "eval_negation-triplets_loss": 1.614029884338379, + "eval_negation-triplets_runtime": 1.1098, + "eval_negation-triplets_samples_per_second": 115.337, + "eval_negation-triplets_steps_per_second": 1.802, + "step": 1700 + }, + { + "epoch": 1.3188518231186968, + "eval_scitail-pairs-pos_loss": 0.11092796176671982, + "eval_scitail-pairs-pos_runtime": 1.296, + "eval_scitail-pairs-pos_samples_per_second": 98.762, + "eval_scitail-pairs-pos_steps_per_second": 1.543, + "step": 1700 + }, + { + "epoch": 1.3188518231186968, + "eval_scitail-pairs-qa_loss": 0.006590584293007851, + "eval_scitail-pairs-qa_runtime": 0.7848, + "eval_scitail-pairs-qa_samples_per_second": 163.098, + "eval_scitail-pairs-qa_steps_per_second": 2.548, + "step": 1700 + }, + { + "epoch": 1.3188518231186968, + "eval_xsum-pairs_loss": 0.0976954996585846, + "eval_xsum-pairs_runtime": 3.9275, + "eval_xsum-pairs_samples_per_second": 32.591, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 1700 + }, + { + "epoch": 1.3188518231186968, + "eval_sciq_pairs_loss": 0.044491760432720184, + "eval_sciq_pairs_runtime": 5.9846, + "eval_sciq_pairs_samples_per_second": 21.388, + "eval_sciq_pairs_steps_per_second": 0.334, + "step": 1700 + }, + { + "epoch": 1.3188518231186968, + "eval_qasc_pairs_loss": 0.2388419657945633, + "eval_qasc_pairs_runtime": 0.9124, + "eval_qasc_pairs_samples_per_second": 140.295, + "eval_qasc_pairs_steps_per_second": 2.192, + "step": 1700 + }, + { + "epoch": 1.3188518231186968, + "eval_openbookqa_pairs_loss": 1.0515943765640259, + "eval_openbookqa_pairs_runtime": 0.8659, + "eval_openbookqa_pairs_samples_per_second": 147.829, + "eval_openbookqa_pairs_steps_per_second": 2.31, + "step": 1700 + }, + { + "epoch": 1.3188518231186968, + "eval_msmarco_pairs_loss": 0.3378428518772125, + "eval_msmarco_pairs_runtime": 1.972, + "eval_msmarco_pairs_samples_per_second": 64.907, + "eval_msmarco_pairs_steps_per_second": 1.014, + "step": 1700 + }, + { + "epoch": 1.3188518231186968, + "eval_nq_pairs_loss": 0.5268489718437195, + "eval_nq_pairs_runtime": 3.3381, + "eval_nq_pairs_samples_per_second": 38.346, + "eval_nq_pairs_steps_per_second": 0.599, + "step": 1700 + }, + { + "epoch": 1.3188518231186968, + "eval_trivia_pairs_loss": 0.7603480815887451, + "eval_trivia_pairs_runtime": 4.4723, + "eval_trivia_pairs_samples_per_second": 28.62, + "eval_trivia_pairs_steps_per_second": 0.447, + "step": 1700 + }, + { + "epoch": 1.3188518231186968, + "eval_gooaq_pairs_loss": 0.6031200885772705, + "eval_gooaq_pairs_runtime": 1.4701, + "eval_gooaq_pairs_samples_per_second": 87.068, + "eval_gooaq_pairs_steps_per_second": 1.36, + "step": 1700 + }, + { + "epoch": 1.3188518231186968, + "eval_paws-pos_loss": 0.0322817787528038, + "eval_paws-pos_runtime": 1.056, + "eval_paws-pos_samples_per_second": 121.208, + "eval_paws-pos_steps_per_second": 1.894, + "step": 1700 + }, + { + "epoch": 1.3188518231186968, + "eval_global_dataset_loss": 0.40005502104759216, + "eval_global_dataset_runtime": 19.4166, + "eval_global_dataset_samples_per_second": 21.425, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1700 + }, + { + "epoch": 1.3196276183087665, + "grad_norm": 5.785887241363525, + "learning_rate": 1.9722668679156454e-05, + "loss": 0.1604, + "step": 1701 + }, + { + "epoch": 1.3204034134988363, + "grad_norm": 5.983560085296631, + "learning_rate": 1.969402992948525e-05, + "loss": 0.2131, + "step": 1702 + }, + { + "epoch": 1.321179208688906, + "grad_norm": 9.850682258605957, + "learning_rate": 1.9665389710457497e-05, + "loss": 0.5133, + "step": 1703 + }, + { + "epoch": 1.321955003878976, + "grad_norm": 3.7730085849761963, + "learning_rate": 1.963674811684764e-05, + "loss": 0.069, + "step": 1704 + }, + { + "epoch": 1.3227307990690458, + "grad_norm": 0.4810899794101715, + "learning_rate": 1.9608105243434677e-05, + "loss": 0.0325, + "step": 1705 + }, + { + "epoch": 1.3235065942591155, + "grad_norm": 5.205442428588867, + "learning_rate": 1.9579461185001856e-05, + "loss": 0.2207, + "step": 1706 + }, + { + "epoch": 1.3242823894491855, + "grad_norm": 0.08708298206329346, + "learning_rate": 1.955081603633632e-05, + "loss": 0.0013, + "step": 1707 + }, + { + "epoch": 1.3250581846392553, + "grad_norm": 6.353906631469727, + "learning_rate": 1.9522169892228853e-05, + "loss": 0.2004, + "step": 1708 + }, + { + "epoch": 1.325833979829325, + "grad_norm": 8.46937370300293, + "learning_rate": 1.94935228474735e-05, + "loss": 0.3886, + "step": 1709 + }, + { + "epoch": 1.3266097750193948, + "grad_norm": 2.9853732585906982, + "learning_rate": 1.9464874996867302e-05, + "loss": 0.0816, + "step": 1710 + }, + { + "epoch": 1.3273855702094646, + "grad_norm": 6.22665548324585, + "learning_rate": 1.9436226435209973e-05, + "loss": 0.2267, + "step": 1711 + }, + { + "epoch": 1.3281613653995346, + "grad_norm": 6.36606502532959, + "learning_rate": 1.940757725730357e-05, + "loss": 0.1539, + "step": 1712 + }, + { + "epoch": 1.3289371605896043, + "grad_norm": 0.5327939391136169, + "learning_rate": 1.9378927557952185e-05, + "loss": 0.0141, + "step": 1713 + }, + { + "epoch": 1.3297129557796743, + "grad_norm": 11.666084289550781, + "learning_rate": 1.935027743196165e-05, + "loss": 0.7857, + "step": 1714 + }, + { + "epoch": 1.330488750969744, + "grad_norm": 7.120075225830078, + "learning_rate": 1.9321626974139205e-05, + "loss": 0.2806, + "step": 1715 + }, + { + "epoch": 1.3312645461598138, + "grad_norm": 6.328937530517578, + "learning_rate": 1.929297627929317e-05, + "loss": 0.3993, + "step": 1716 + }, + { + "epoch": 1.3320403413498836, + "grad_norm": 13.161340713500977, + "learning_rate": 1.9264325442232683e-05, + "loss": 0.8072, + "step": 1717 + }, + { + "epoch": 1.3328161365399533, + "grad_norm": 6.1413044929504395, + "learning_rate": 1.9235674557767315e-05, + "loss": 0.1761, + "step": 1718 + }, + { + "epoch": 1.3335919317300233, + "grad_norm": 6.826531410217285, + "learning_rate": 1.9207023720706827e-05, + "loss": 0.3838, + "step": 1719 + }, + { + "epoch": 1.334367726920093, + "grad_norm": 10.711207389831543, + "learning_rate": 1.9178373025860795e-05, + "loss": 1.0871, + "step": 1720 + }, + { + "epoch": 1.334367726920093, + "eval_Qnli-dev_cosine_accuracy": 0.693359375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8189771771430969, + "eval_Qnli-dev_cosine_ap": 0.7137928507719671, + "eval_Qnli-dev_cosine_f1": 0.6832504145936982, + "eval_Qnli-dev_cosine_f1_threshold": 0.7291297912597656, + "eval_Qnli-dev_cosine_precision": 0.5613079019073569, + "eval_Qnli-dev_cosine_recall": 0.8728813559322034, + "eval_Qnli-dev_dot_accuracy": 0.66015625, + "eval_Qnli-dev_dot_accuracy_threshold": 401.0465393066406, + "eval_Qnli-dev_dot_ap": 0.6321544208110721, + "eval_Qnli-dev_dot_f1": 0.6731391585760518, + "eval_Qnli-dev_dot_f1_threshold": 343.5312805175781, + "eval_Qnli-dev_dot_precision": 0.5445026178010471, + "eval_Qnli-dev_dot_recall": 0.8813559322033898, + "eval_Qnli-dev_euclidean_accuracy": 0.6796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 14.031986236572266, + "eval_Qnli-dev_euclidean_ap": 0.715995523683812, + "eval_Qnli-dev_euclidean_f1": 0.6821705426356589, + "eval_Qnli-dev_euclidean_f1_threshold": 14.44270133972168, + "eval_Qnli-dev_euclidean_precision": 0.6285714285714286, + "eval_Qnli-dev_euclidean_recall": 0.7457627118644068, + "eval_Qnli-dev_manhattan_accuracy": 0.677734375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 248.22784423828125, + "eval_Qnli-dev_manhattan_ap": 0.7157555197465306, + "eval_Qnli-dev_manhattan_f1": 0.6845637583892616, + "eval_Qnli-dev_manhattan_f1_threshold": 296.9533996582031, + "eval_Qnli-dev_manhattan_precision": 0.5666666666666667, + "eval_Qnli-dev_manhattan_recall": 0.864406779661017, + "eval_Qnli-dev_max_accuracy": 0.693359375, + "eval_Qnli-dev_max_accuracy_threshold": 401.0465393066406, + "eval_Qnli-dev_max_ap": 0.715995523683812, + "eval_Qnli-dev_max_f1": 0.6845637583892616, + "eval_Qnli-dev_max_f1_threshold": 343.5312805175781, + "eval_Qnli-dev_max_precision": 0.6285714285714286, + "eval_Qnli-dev_max_recall": 0.8813559322033898, + "eval_allNLI-dev_cosine_accuracy": 0.7265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.879446268081665, + "eval_allNLI-dev_cosine_ap": 0.6121722135620142, + "eval_allNLI-dev_cosine_f1": 0.6246851385390428, + "eval_allNLI-dev_cosine_f1_threshold": 0.8359977602958679, + "eval_allNLI-dev_cosine_precision": 0.5535714285714286, + "eval_allNLI-dev_cosine_recall": 0.7167630057803468, + "eval_allNLI-dev_dot_accuracy": 0.6875, + "eval_allNLI-dev_dot_accuracy_threshold": 397.56231689453125, + "eval_allNLI-dev_dot_ap": 0.5178513402847004, + "eval_allNLI-dev_dot_f1": 0.5721153846153847, + "eval_allNLI-dev_dot_f1_threshold": 376.5932922363281, + "eval_allNLI-dev_dot_precision": 0.4897119341563786, + "eval_allNLI-dev_dot_recall": 0.6878612716763006, + "eval_allNLI-dev_euclidean_accuracy": 0.73046875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.692667007446289, + "eval_allNLI-dev_euclidean_ap": 0.616481074984663, + "eval_allNLI-dev_euclidean_f1": 0.6246851385390428, + "eval_allNLI-dev_euclidean_f1_threshold": 12.38914966583252, + "eval_allNLI-dev_euclidean_precision": 0.5535714285714286, + "eval_allNLI-dev_euclidean_recall": 0.7167630057803468, + "eval_allNLI-dev_manhattan_accuracy": 0.7421875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 210.55300903320312, + "eval_allNLI-dev_manhattan_ap": 0.6136453635781026, + "eval_allNLI-dev_manhattan_f1": 0.6320987654320988, + "eval_allNLI-dev_manhattan_f1_threshold": 227.63290405273438, + "eval_allNLI-dev_manhattan_precision": 0.5517241379310345, + "eval_allNLI-dev_manhattan_recall": 0.7398843930635838, + "eval_allNLI-dev_max_accuracy": 0.7421875, + "eval_allNLI-dev_max_accuracy_threshold": 397.56231689453125, + "eval_allNLI-dev_max_ap": 0.616481074984663, + "eval_allNLI-dev_max_f1": 0.6320987654320988, + "eval_allNLI-dev_max_f1_threshold": 376.5932922363281, + "eval_allNLI-dev_max_precision": 0.5535714285714286, + "eval_allNLI-dev_max_recall": 0.7398843930635838, + "eval_sequential_score": 0.715995523683812, + "eval_sts-test_pearson_cosine": 0.8194114259272549, + "eval_sts-test_pearson_dot": 0.7696392331066577, + "eval_sts-test_pearson_euclidean": 0.8578536223945645, + "eval_sts-test_pearson_manhattan": 0.8573023875333354, + "eval_sts-test_pearson_max": 0.8578536223945645, + "eval_sts-test_spearman_cosine": 0.8671053186072486, + "eval_sts-test_spearman_dot": 0.7631203926129647, + "eval_sts-test_spearman_euclidean": 0.8619767345492714, + "eval_sts-test_spearman_manhattan": 0.8639808750394238, + "eval_sts-test_spearman_max": 0.8671053186072486, + "eval_vitaminc-pairs_loss": 3.953892946243286, + "eval_vitaminc-pairs_runtime": 5.5188, + "eval_vitaminc-pairs_samples_per_second": 23.193, + "eval_vitaminc-pairs_steps_per_second": 0.362, + "step": 1720 + }, + { + "epoch": 1.334367726920093, + "eval_negation-triplets_loss": 1.6196330785751343, + "eval_negation-triplets_runtime": 1.124, + "eval_negation-triplets_samples_per_second": 113.883, + "eval_negation-triplets_steps_per_second": 1.779, + "step": 1720 + }, + { + "epoch": 1.334367726920093, + "eval_scitail-pairs-pos_loss": 0.13077253103256226, + "eval_scitail-pairs-pos_runtime": 1.3862, + "eval_scitail-pairs-pos_samples_per_second": 92.339, + "eval_scitail-pairs-pos_steps_per_second": 1.443, + "step": 1720 + }, + { + "epoch": 1.334367726920093, + "eval_scitail-pairs-qa_loss": 0.007284028455615044, + "eval_scitail-pairs-qa_runtime": 0.7862, + "eval_scitail-pairs-qa_samples_per_second": 162.811, + "eval_scitail-pairs-qa_steps_per_second": 2.544, + "step": 1720 + }, + { + "epoch": 1.334367726920093, + "eval_xsum-pairs_loss": 0.09171821922063828, + "eval_xsum-pairs_runtime": 3.9361, + "eval_xsum-pairs_samples_per_second": 32.519, + "eval_xsum-pairs_steps_per_second": 0.508, + "step": 1720 + }, + { + "epoch": 1.334367726920093, + "eval_sciq_pairs_loss": 0.04136815667152405, + "eval_sciq_pairs_runtime": 6.039, + "eval_sciq_pairs_samples_per_second": 21.196, + "eval_sciq_pairs_steps_per_second": 0.331, + "step": 1720 + }, + { + "epoch": 1.334367726920093, + "eval_qasc_pairs_loss": 0.23705051839351654, + "eval_qasc_pairs_runtime": 0.928, + "eval_qasc_pairs_samples_per_second": 137.937, + "eval_qasc_pairs_steps_per_second": 2.155, + "step": 1720 + }, + { + "epoch": 1.334367726920093, + "eval_openbookqa_pairs_loss": 1.0086489915847778, + "eval_openbookqa_pairs_runtime": 0.8798, + "eval_openbookqa_pairs_samples_per_second": 145.488, + "eval_openbookqa_pairs_steps_per_second": 2.273, + "step": 1720 + }, + { + "epoch": 1.334367726920093, + "eval_msmarco_pairs_loss": 0.3558310568332672, + "eval_msmarco_pairs_runtime": 1.9948, + "eval_msmarco_pairs_samples_per_second": 64.166, + "eval_msmarco_pairs_steps_per_second": 1.003, + "step": 1720 + }, + { + "epoch": 1.334367726920093, + "eval_nq_pairs_loss": 0.5158048272132874, + "eval_nq_pairs_runtime": 3.3551, + "eval_nq_pairs_samples_per_second": 38.151, + "eval_nq_pairs_steps_per_second": 0.596, + "step": 1720 + }, + { + "epoch": 1.334367726920093, + "eval_trivia_pairs_loss": 0.7480645775794983, + "eval_trivia_pairs_runtime": 4.5024, + "eval_trivia_pairs_samples_per_second": 28.429, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 1720 + }, + { + "epoch": 1.334367726920093, + "eval_gooaq_pairs_loss": 0.6158685088157654, + "eval_gooaq_pairs_runtime": 1.4806, + "eval_gooaq_pairs_samples_per_second": 86.449, + "eval_gooaq_pairs_steps_per_second": 1.351, + "step": 1720 + }, + { + "epoch": 1.334367726920093, + "eval_paws-pos_loss": 0.03172389417886734, + "eval_paws-pos_runtime": 1.0424, + "eval_paws-pos_samples_per_second": 122.799, + "eval_paws-pos_steps_per_second": 1.919, + "step": 1720 + }, + { + "epoch": 1.334367726920093, + "eval_global_dataset_loss": 0.40301206707954407, + "eval_global_dataset_runtime": 19.4978, + "eval_global_dataset_samples_per_second": 21.336, + "eval_global_dataset_steps_per_second": 0.359, + "step": 1720 + }, + { + "epoch": 1.3351435221101629, + "grad_norm": 5.908486843109131, + "learning_rate": 1.9149722568038347e-05, + "loss": 0.1416, + "step": 1721 + }, + { + "epoch": 1.3359193173002328, + "grad_norm": 15.63554573059082, + "learning_rate": 1.9121072442047815e-05, + "loss": 2.4614, + "step": 1722 + }, + { + "epoch": 1.3366951124903026, + "grad_norm": 12.723451614379883, + "learning_rate": 1.9092422742696432e-05, + "loss": 0.889, + "step": 1723 + }, + { + "epoch": 1.3374709076803724, + "grad_norm": 11.989603042602539, + "learning_rate": 1.9063773564790027e-05, + "loss": 0.8296, + "step": 1724 + }, + { + "epoch": 1.3382467028704421, + "grad_norm": 4.495663166046143, + "learning_rate": 1.90351250031327e-05, + "loss": 0.1145, + "step": 1725 + }, + { + "epoch": 1.339022498060512, + "grad_norm": 7.5452165603637695, + "learning_rate": 1.9006477152526502e-05, + "loss": 0.4375, + "step": 1726 + }, + { + "epoch": 1.3397982932505819, + "grad_norm": 7.630861282348633, + "learning_rate": 1.897783010777115e-05, + "loss": 0.4656, + "step": 1727 + }, + { + "epoch": 1.3405740884406516, + "grad_norm": 0.6148019433021545, + "learning_rate": 1.8949183963663677e-05, + "loss": 0.0124, + "step": 1728 + }, + { + "epoch": 1.3413498836307216, + "grad_norm": 11.428054809570312, + "learning_rate": 1.8920538814998144e-05, + "loss": 0.9157, + "step": 1729 + }, + { + "epoch": 1.3421256788207914, + "grad_norm": 2.4426681995391846, + "learning_rate": 1.889189475656532e-05, + "loss": 0.0589, + "step": 1730 + }, + { + "epoch": 1.3429014740108611, + "grad_norm": 2.937155246734619, + "learning_rate": 1.8863251883152362e-05, + "loss": 0.0641, + "step": 1731 + }, + { + "epoch": 1.343677269200931, + "grad_norm": 7.9810686111450195, + "learning_rate": 1.8834610289542504e-05, + "loss": 0.464, + "step": 1732 + }, + { + "epoch": 1.3444530643910007, + "grad_norm": 1.203864336013794, + "learning_rate": 1.8805970070514748e-05, + "loss": 0.0258, + "step": 1733 + }, + { + "epoch": 1.3452288595810706, + "grad_norm": 11.484176635742188, + "learning_rate": 1.8777331320843546e-05, + "loss": 0.7976, + "step": 1734 + }, + { + "epoch": 1.3460046547711404, + "grad_norm": 9.385132789611816, + "learning_rate": 1.874869413529849e-05, + "loss": 0.5184, + "step": 1735 + }, + { + "epoch": 1.3467804499612102, + "grad_norm": 4.999711036682129, + "learning_rate": 1.8720058608644e-05, + "loss": 0.2261, + "step": 1736 + }, + { + "epoch": 1.3475562451512801, + "grad_norm": 9.523402214050293, + "learning_rate": 1.8691424835638986e-05, + "loss": 0.5473, + "step": 1737 + }, + { + "epoch": 1.34833204034135, + "grad_norm": 7.102741718292236, + "learning_rate": 1.8662792911036588e-05, + "loss": 0.3477, + "step": 1738 + }, + { + "epoch": 1.3491078355314197, + "grad_norm": 0.7534499168395996, + "learning_rate": 1.8634162929583797e-05, + "loss": 0.0635, + "step": 1739 + }, + { + "epoch": 1.3498836307214894, + "grad_norm": 7.467993259429932, + "learning_rate": 1.8605534986021197e-05, + "loss": 0.2952, + "step": 1740 + }, + { + "epoch": 1.3498836307214894, + "eval_Qnli-dev_cosine_accuracy": 0.68359375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8260684013366699, + "eval_Qnli-dev_cosine_ap": 0.710167328474358, + "eval_Qnli-dev_cosine_f1": 0.6879194630872484, + "eval_Qnli-dev_cosine_f1_threshold": 0.7387106418609619, + "eval_Qnli-dev_cosine_precision": 0.5694444444444444, + "eval_Qnli-dev_cosine_recall": 0.8686440677966102, + "eval_Qnli-dev_dot_accuracy": 0.650390625, + "eval_Qnli-dev_dot_accuracy_threshold": 417.19366455078125, + "eval_Qnli-dev_dot_ap": 0.6250178634400758, + "eval_Qnli-dev_dot_f1": 0.6707882534775889, + "eval_Qnli-dev_dot_f1_threshold": 332.7039489746094, + "eval_Qnli-dev_dot_precision": 0.5279805352798054, + "eval_Qnli-dev_dot_recall": 0.9194915254237288, + "eval_Qnli-dev_euclidean_accuracy": 0.68359375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.803336143493652, + "eval_Qnli-dev_euclidean_ap": 0.7125352835367065, + "eval_Qnli-dev_euclidean_f1": 0.6873857404021938, + "eval_Qnli-dev_euclidean_f1_threshold": 15.179275512695312, + "eval_Qnli-dev_euclidean_precision": 0.6045016077170418, + "eval_Qnli-dev_euclidean_recall": 0.7966101694915254, + "eval_Qnli-dev_manhattan_accuracy": 0.68359375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 245.06686401367188, + "eval_Qnli-dev_manhattan_ap": 0.713502922180456, + "eval_Qnli-dev_manhattan_f1": 0.6836734693877551, + "eval_Qnli-dev_manhattan_f1_threshold": 290.7037048339844, + "eval_Qnli-dev_manhattan_precision": 0.5710227272727273, + "eval_Qnli-dev_manhattan_recall": 0.8516949152542372, + "eval_Qnli-dev_max_accuracy": 0.68359375, + "eval_Qnli-dev_max_accuracy_threshold": 417.19366455078125, + "eval_Qnli-dev_max_ap": 0.713502922180456, + "eval_Qnli-dev_max_f1": 0.6879194630872484, + "eval_Qnli-dev_max_f1_threshold": 332.7039489746094, + "eval_Qnli-dev_max_precision": 0.6045016077170418, + "eval_Qnli-dev_max_recall": 0.9194915254237288, + "eval_allNLI-dev_cosine_accuracy": 0.73046875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8699325323104858, + "eval_allNLI-dev_cosine_ap": 0.6107502599697213, + "eval_allNLI-dev_cosine_f1": 0.6214099216710183, + "eval_allNLI-dev_cosine_f1_threshold": 0.8432725667953491, + "eval_allNLI-dev_cosine_precision": 0.5666666666666667, + "eval_allNLI-dev_cosine_recall": 0.6878612716763006, + "eval_allNLI-dev_dot_accuracy": 0.68359375, + "eval_allNLI-dev_dot_accuracy_threshold": 466.9180603027344, + "eval_allNLI-dev_dot_ap": 0.5033486606785589, + "eval_allNLI-dev_dot_f1": 0.5707196029776674, + "eval_allNLI-dev_dot_f1_threshold": 373.4122314453125, + "eval_allNLI-dev_dot_precision": 0.5, + "eval_allNLI-dev_dot_recall": 0.6647398843930635, + "eval_allNLI-dev_euclidean_accuracy": 0.732421875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.99596881866455, + "eval_allNLI-dev_euclidean_ap": 0.6170798482380073, + "eval_allNLI-dev_euclidean_f1": 0.6342710997442456, + "eval_allNLI-dev_euclidean_f1_threshold": 12.049517631530762, + "eval_allNLI-dev_euclidean_precision": 0.5688073394495413, + "eval_allNLI-dev_euclidean_recall": 0.7167630057803468, + "eval_allNLI-dev_manhattan_accuracy": 0.73828125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 201.16879272460938, + "eval_allNLI-dev_manhattan_ap": 0.6184987292312732, + "eval_allNLI-dev_manhattan_f1": 0.6313131313131313, + "eval_allNLI-dev_manhattan_f1_threshold": 220.9921112060547, + "eval_allNLI-dev_manhattan_precision": 0.5605381165919282, + "eval_allNLI-dev_manhattan_recall": 0.7225433526011561, + "eval_allNLI-dev_max_accuracy": 0.73828125, + "eval_allNLI-dev_max_accuracy_threshold": 466.9180603027344, + "eval_allNLI-dev_max_ap": 0.6184987292312732, + "eval_allNLI-dev_max_f1": 0.6342710997442456, + "eval_allNLI-dev_max_f1_threshold": 373.4122314453125, + "eval_allNLI-dev_max_precision": 0.5688073394495413, + "eval_allNLI-dev_max_recall": 0.7225433526011561, + "eval_sequential_score": 0.713502922180456, + "eval_sts-test_pearson_cosine": 0.8165337528270065, + "eval_sts-test_pearson_dot": 0.7679722338090162, + "eval_sts-test_pearson_euclidean": 0.8583705669573758, + "eval_sts-test_pearson_manhattan": 0.8584062835995685, + "eval_sts-test_pearson_max": 0.8584062835995685, + "eval_sts-test_spearman_cosine": 0.8701769109353701, + "eval_sts-test_spearman_dot": 0.768467586582839, + "eval_sts-test_spearman_euclidean": 0.8650016513117618, + "eval_sts-test_spearman_manhattan": 0.8664706562181957, + "eval_sts-test_spearman_max": 0.8701769109353701, + "eval_vitaminc-pairs_loss": 3.8053059577941895, + "eval_vitaminc-pairs_runtime": 5.5342, + "eval_vitaminc-pairs_samples_per_second": 23.129, + "eval_vitaminc-pairs_steps_per_second": 0.361, + "step": 1740 + }, + { + "epoch": 1.3498836307214894, + "eval_negation-triplets_loss": 1.559175968170166, + "eval_negation-triplets_runtime": 1.0922, + "eval_negation-triplets_samples_per_second": 117.199, + "eval_negation-triplets_steps_per_second": 1.831, + "step": 1740 + }, + { + "epoch": 1.3498836307214894, + "eval_scitail-pairs-pos_loss": 0.1014331579208374, + "eval_scitail-pairs-pos_runtime": 1.332, + "eval_scitail-pairs-pos_samples_per_second": 96.093, + "eval_scitail-pairs-pos_steps_per_second": 1.501, + "step": 1740 + }, + { + "epoch": 1.3498836307214894, + "eval_scitail-pairs-qa_loss": 0.005077951587736607, + "eval_scitail-pairs-qa_runtime": 0.7885, + "eval_scitail-pairs-qa_samples_per_second": 162.343, + "eval_scitail-pairs-qa_steps_per_second": 2.537, + "step": 1740 + }, + { + "epoch": 1.3498836307214894, + "eval_xsum-pairs_loss": 0.09790876507759094, + "eval_xsum-pairs_runtime": 3.9392, + "eval_xsum-pairs_samples_per_second": 32.494, + "eval_xsum-pairs_steps_per_second": 0.508, + "step": 1740 + }, + { + "epoch": 1.3498836307214894, + "eval_sciq_pairs_loss": 0.04244896396994591, + "eval_sciq_pairs_runtime": 6.1367, + "eval_sciq_pairs_samples_per_second": 20.858, + "eval_sciq_pairs_steps_per_second": 0.326, + "step": 1740 + }, + { + "epoch": 1.3498836307214894, + "eval_qasc_pairs_loss": 0.23306579887866974, + "eval_qasc_pairs_runtime": 0.9332, + "eval_qasc_pairs_samples_per_second": 137.165, + "eval_qasc_pairs_steps_per_second": 2.143, + "step": 1740 + }, + { + "epoch": 1.3498836307214894, + "eval_openbookqa_pairs_loss": 0.9666154980659485, + "eval_openbookqa_pairs_runtime": 0.8888, + "eval_openbookqa_pairs_samples_per_second": 144.011, + "eval_openbookqa_pairs_steps_per_second": 2.25, + "step": 1740 + }, + { + "epoch": 1.3498836307214894, + "eval_msmarco_pairs_loss": 0.3814646005630493, + "eval_msmarco_pairs_runtime": 1.9823, + "eval_msmarco_pairs_samples_per_second": 64.57, + "eval_msmarco_pairs_steps_per_second": 1.009, + "step": 1740 + }, + { + "epoch": 1.3498836307214894, + "eval_nq_pairs_loss": 0.5080822706222534, + "eval_nq_pairs_runtime": 3.3671, + "eval_nq_pairs_samples_per_second": 38.015, + "eval_nq_pairs_steps_per_second": 0.594, + "step": 1740 + }, + { + "epoch": 1.3498836307214894, + "eval_trivia_pairs_loss": 0.7755739092826843, + "eval_trivia_pairs_runtime": 4.4978, + "eval_trivia_pairs_samples_per_second": 28.458, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 1740 + }, + { + "epoch": 1.3498836307214894, + "eval_gooaq_pairs_loss": 0.6271055936813354, + "eval_gooaq_pairs_runtime": 1.4773, + "eval_gooaq_pairs_samples_per_second": 86.646, + "eval_gooaq_pairs_steps_per_second": 1.354, + "step": 1740 + }, + { + "epoch": 1.3498836307214894, + "eval_paws-pos_loss": 0.03230850771069527, + "eval_paws-pos_runtime": 1.0439, + "eval_paws-pos_samples_per_second": 122.617, + "eval_paws-pos_steps_per_second": 1.916, + "step": 1740 + }, + { + "epoch": 1.3498836307214894, + "eval_global_dataset_loss": 0.3831497132778168, + "eval_global_dataset_runtime": 19.8669, + "eval_global_dataset_samples_per_second": 20.939, + "eval_global_dataset_steps_per_second": 0.352, + "step": 1740 + }, + { + "epoch": 1.3506594259115594, + "grad_norm": 10.861262321472168, + "learning_rate": 1.8576909175082618e-05, + "loss": 0.834, + "step": 1741 + }, + { + "epoch": 1.3514352211016292, + "grad_norm": 4.407554626464844, + "learning_rate": 1.8548285591494832e-05, + "loss": 0.1005, + "step": 1742 + }, + { + "epoch": 1.352211016291699, + "grad_norm": 10.520988464355469, + "learning_rate": 1.851966432997724e-05, + "loss": 0.8344, + "step": 1743 + }, + { + "epoch": 1.352986811481769, + "grad_norm": 10.74624252319336, + "learning_rate": 1.849104548524157e-05, + "loss": 0.7943, + "step": 1744 + }, + { + "epoch": 1.3537626066718387, + "grad_norm": 4.92795991897583, + "learning_rate": 1.8462429151991535e-05, + "loss": 0.1538, + "step": 1745 + }, + { + "epoch": 1.3545384018619084, + "grad_norm": 5.671476364135742, + "learning_rate": 1.8433815424922556e-05, + "loss": 0.1648, + "step": 1746 + }, + { + "epoch": 1.3553141970519782, + "grad_norm": 4.4333391189575195, + "learning_rate": 1.8405204398721413e-05, + "loss": 0.1345, + "step": 1747 + }, + { + "epoch": 1.3560899922420482, + "grad_norm": 9.296590805053711, + "learning_rate": 1.837659616806596e-05, + "loss": 0.4956, + "step": 1748 + }, + { + "epoch": 1.356865787432118, + "grad_norm": 8.126757621765137, + "learning_rate": 1.834799082762479e-05, + "loss": 0.5533, + "step": 1749 + }, + { + "epoch": 1.3576415826221877, + "grad_norm": 11.96084213256836, + "learning_rate": 1.831938847205694e-05, + "loss": 1.6977, + "step": 1750 + }, + { + "epoch": 1.3584173778122577, + "grad_norm": 10.518410682678223, + "learning_rate": 1.8290789196011577e-05, + "loss": 0.6022, + "step": 1751 + }, + { + "epoch": 1.3591931730023274, + "grad_norm": 6.022642135620117, + "learning_rate": 1.8262193094127655e-05, + "loss": 0.1943, + "step": 1752 + }, + { + "epoch": 1.3599689681923972, + "grad_norm": 5.988025665283203, + "learning_rate": 1.8233600261033638e-05, + "loss": 0.2082, + "step": 1753 + }, + { + "epoch": 1.360744763382467, + "grad_norm": 6.095760822296143, + "learning_rate": 1.820501079134718e-05, + "loss": 0.1851, + "step": 1754 + }, + { + "epoch": 1.3615205585725367, + "grad_norm": 5.737420082092285, + "learning_rate": 1.8176424779674787e-05, + "loss": 0.2646, + "step": 1755 + }, + { + "epoch": 1.3622963537626067, + "grad_norm": 3.998337507247925, + "learning_rate": 1.8147842320611543e-05, + "loss": 0.1533, + "step": 1756 + }, + { + "epoch": 1.3630721489526765, + "grad_norm": 4.371984481811523, + "learning_rate": 1.811926350874076e-05, + "loss": 0.2075, + "step": 1757 + }, + { + "epoch": 1.3638479441427462, + "grad_norm": 8.951359748840332, + "learning_rate": 1.8090688438633687e-05, + "loss": 0.6163, + "step": 1758 + }, + { + "epoch": 1.3646237393328162, + "grad_norm": 3.40568470954895, + "learning_rate": 1.806211720484919e-05, + "loss": 0.1796, + "step": 1759 + }, + { + "epoch": 1.365399534522886, + "grad_norm": 6.764983177185059, + "learning_rate": 1.8033549901933435e-05, + "loss": 0.3106, + "step": 1760 + }, + { + "epoch": 1.365399534522886, + "eval_Qnli-dev_cosine_accuracy": 0.69140625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8534146547317505, + "eval_Qnli-dev_cosine_ap": 0.7096166604762388, + "eval_Qnli-dev_cosine_f1": 0.6883116883116883, + "eval_Qnli-dev_cosine_f1_threshold": 0.7433306574821472, + "eval_Qnli-dev_cosine_precision": 0.5578947368421052, + "eval_Qnli-dev_cosine_recall": 0.8983050847457628, + "eval_Qnli-dev_dot_accuracy": 0.654296875, + "eval_Qnli-dev_dot_accuracy_threshold": 423.22418212890625, + "eval_Qnli-dev_dot_ap": 0.6287333336948954, + "eval_Qnli-dev_dot_f1": 0.682170542635659, + "eval_Qnli-dev_dot_f1_threshold": 340.4381103515625, + "eval_Qnli-dev_dot_precision": 0.5378973105134475, + "eval_Qnli-dev_dot_recall": 0.9322033898305084, + "eval_Qnli-dev_euclidean_accuracy": 0.6796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.604022979736328, + "eval_Qnli-dev_euclidean_ap": 0.7104364120785174, + "eval_Qnli-dev_euclidean_f1": 0.6911519198664441, + "eval_Qnli-dev_euclidean_f1_threshold": 15.598084449768066, + "eval_Qnli-dev_euclidean_precision": 0.5702479338842975, + "eval_Qnli-dev_euclidean_recall": 0.8771186440677966, + "eval_Qnli-dev_manhattan_accuracy": 0.68359375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 230.25997924804688, + "eval_Qnli-dev_manhattan_ap": 0.7135814820459, + "eval_Qnli-dev_manhattan_f1": 0.6930693069306931, + "eval_Qnli-dev_manhattan_f1_threshold": 287.4403991699219, + "eval_Qnli-dev_manhattan_precision": 0.5675675675675675, + "eval_Qnli-dev_manhattan_recall": 0.8898305084745762, + "eval_Qnli-dev_max_accuracy": 0.69140625, + "eval_Qnli-dev_max_accuracy_threshold": 423.22418212890625, + "eval_Qnli-dev_max_ap": 0.7135814820459, + "eval_Qnli-dev_max_f1": 0.6930693069306931, + "eval_Qnli-dev_max_f1_threshold": 340.4381103515625, + "eval_Qnli-dev_max_precision": 0.5702479338842975, + "eval_Qnli-dev_max_recall": 0.9322033898305084, + "eval_allNLI-dev_cosine_accuracy": 0.7265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8718122839927673, + "eval_allNLI-dev_cosine_ap": 0.6088186326733481, + "eval_allNLI-dev_cosine_f1": 0.6142857142857143, + "eval_allNLI-dev_cosine_f1_threshold": 0.8326781392097473, + "eval_allNLI-dev_cosine_precision": 0.5222672064777328, + "eval_allNLI-dev_cosine_recall": 0.7456647398843931, + "eval_allNLI-dev_dot_accuracy": 0.68359375, + "eval_allNLI-dev_dot_accuracy_threshold": 402.9331970214844, + "eval_allNLI-dev_dot_ap": 0.5121654816949537, + "eval_allNLI-dev_dot_f1": 0.5714285714285715, + "eval_allNLI-dev_dot_f1_threshold": 338.00372314453125, + "eval_allNLI-dev_dot_precision": 0.4134366925064599, + "eval_allNLI-dev_dot_recall": 0.9248554913294798, + "eval_allNLI-dev_euclidean_accuracy": 0.736328125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.73448371887207, + "eval_allNLI-dev_euclidean_ap": 0.6123610160376709, + "eval_allNLI-dev_euclidean_f1": 0.6218905472636818, + "eval_allNLI-dev_euclidean_f1_threshold": 12.136075973510742, + "eval_allNLI-dev_euclidean_precision": 0.5458515283842795, + "eval_allNLI-dev_euclidean_recall": 0.7225433526011561, + "eval_allNLI-dev_manhattan_accuracy": 0.732421875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 198.0540771484375, + "eval_allNLI-dev_manhattan_ap": 0.6130194515927789, + "eval_allNLI-dev_manhattan_f1": 0.63, + "eval_allNLI-dev_manhattan_f1_threshold": 216.90907287597656, + "eval_allNLI-dev_manhattan_precision": 0.5550660792951542, + "eval_allNLI-dev_manhattan_recall": 0.7283236994219653, + "eval_allNLI-dev_max_accuracy": 0.736328125, + "eval_allNLI-dev_max_accuracy_threshold": 402.9331970214844, + "eval_allNLI-dev_max_ap": 0.6130194515927789, + "eval_allNLI-dev_max_f1": 0.63, + "eval_allNLI-dev_max_f1_threshold": 338.00372314453125, + "eval_allNLI-dev_max_precision": 0.5550660792951542, + "eval_allNLI-dev_max_recall": 0.9248554913294798, + "eval_sequential_score": 0.7135814820459, + "eval_sts-test_pearson_cosine": 0.8244905275628311, + "eval_sts-test_pearson_dot": 0.7741357779777056, + "eval_sts-test_pearson_euclidean": 0.8653290322107638, + "eval_sts-test_pearson_manhattan": 0.8645541364609031, + "eval_sts-test_pearson_max": 0.8653290322107638, + "eval_sts-test_spearman_cosine": 0.8753732484030938, + "eval_sts-test_spearman_dot": 0.7694458044055936, + "eval_sts-test_spearman_euclidean": 0.8708215976069745, + "eval_sts-test_spearman_manhattan": 0.8707891081912044, + "eval_sts-test_spearman_max": 0.8753732484030938, + "eval_vitaminc-pairs_loss": 3.7672548294067383, + "eval_vitaminc-pairs_runtime": 5.5421, + "eval_vitaminc-pairs_samples_per_second": 23.096, + "eval_vitaminc-pairs_steps_per_second": 0.361, + "step": 1760 + }, + { + "epoch": 1.365399534522886, + "eval_negation-triplets_loss": 1.599892020225525, + "eval_negation-triplets_runtime": 1.0968, + "eval_negation-triplets_samples_per_second": 116.702, + "eval_negation-triplets_steps_per_second": 1.823, + "step": 1760 + }, + { + "epoch": 1.365399534522886, + "eval_scitail-pairs-pos_loss": 0.13184364140033722, + "eval_scitail-pairs-pos_runtime": 1.332, + "eval_scitail-pairs-pos_samples_per_second": 96.098, + "eval_scitail-pairs-pos_steps_per_second": 1.502, + "step": 1760 + }, + { + "epoch": 1.365399534522886, + "eval_scitail-pairs-qa_loss": 0.00707290368154645, + "eval_scitail-pairs-qa_runtime": 0.7888, + "eval_scitail-pairs-qa_samples_per_second": 162.272, + "eval_scitail-pairs-qa_steps_per_second": 2.536, + "step": 1760 + }, + { + "epoch": 1.365399534522886, + "eval_xsum-pairs_loss": 0.09621061384677887, + "eval_xsum-pairs_runtime": 3.9594, + "eval_xsum-pairs_samples_per_second": 32.328, + "eval_xsum-pairs_steps_per_second": 0.505, + "step": 1760 + }, + { + "epoch": 1.365399534522886, + "eval_sciq_pairs_loss": 0.04128202423453331, + "eval_sciq_pairs_runtime": 6.2654, + "eval_sciq_pairs_samples_per_second": 20.43, + "eval_sciq_pairs_steps_per_second": 0.319, + "step": 1760 + }, + { + "epoch": 1.365399534522886, + "eval_qasc_pairs_loss": 0.2518593668937683, + "eval_qasc_pairs_runtime": 0.9392, + "eval_qasc_pairs_samples_per_second": 136.292, + "eval_qasc_pairs_steps_per_second": 2.13, + "step": 1760 + }, + { + "epoch": 1.365399534522886, + "eval_openbookqa_pairs_loss": 1.0094834566116333, + "eval_openbookqa_pairs_runtime": 0.8928, + "eval_openbookqa_pairs_samples_per_second": 143.363, + "eval_openbookqa_pairs_steps_per_second": 2.24, + "step": 1760 + }, + { + "epoch": 1.365399534522886, + "eval_msmarco_pairs_loss": 0.36656421422958374, + "eval_msmarco_pairs_runtime": 1.9824, + "eval_msmarco_pairs_samples_per_second": 64.569, + "eval_msmarco_pairs_steps_per_second": 1.009, + "step": 1760 + }, + { + "epoch": 1.365399534522886, + "eval_nq_pairs_loss": 0.5261555314064026, + "eval_nq_pairs_runtime": 3.35, + "eval_nq_pairs_samples_per_second": 38.209, + "eval_nq_pairs_steps_per_second": 0.597, + "step": 1760 + }, + { + "epoch": 1.365399534522886, + "eval_trivia_pairs_loss": 0.6970615983009338, + "eval_trivia_pairs_runtime": 4.4983, + "eval_trivia_pairs_samples_per_second": 28.455, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 1760 + }, + { + "epoch": 1.365399534522886, + "eval_gooaq_pairs_loss": 0.5899819731712341, + "eval_gooaq_pairs_runtime": 1.584, + "eval_gooaq_pairs_samples_per_second": 80.81, + "eval_gooaq_pairs_steps_per_second": 1.263, + "step": 1760 + }, + { + "epoch": 1.365399534522886, + "eval_paws-pos_loss": 0.031426802277565, + "eval_paws-pos_runtime": 1.0981, + "eval_paws-pos_samples_per_second": 116.563, + "eval_paws-pos_steps_per_second": 1.821, + "step": 1760 + }, + { + "epoch": 1.365399534522886, + "eval_global_dataset_loss": 0.3777310252189636, + "eval_global_dataset_runtime": 19.4825, + "eval_global_dataset_samples_per_second": 21.353, + "eval_global_dataset_steps_per_second": 0.359, + "step": 1760 + }, + { + "epoch": 1.3661753297129557, + "grad_norm": 3.4789788722991943, + "learning_rate": 1.800498662441959e-05, + "loss": 0.0962, + "step": 1761 + }, + { + "epoch": 1.3669511249030255, + "grad_norm": 6.056489944458008, + "learning_rate": 1.7976427466827498e-05, + "loss": 0.2075, + "step": 1762 + }, + { + "epoch": 1.3677269200930955, + "grad_norm": 4.120059490203857, + "learning_rate": 1.7947872523663366e-05, + "loss": 0.0891, + "step": 1763 + }, + { + "epoch": 1.3685027152831652, + "grad_norm": 4.467137813568115, + "learning_rate": 1.7919321889419456e-05, + "loss": 0.1119, + "step": 1764 + }, + { + "epoch": 1.369278510473235, + "grad_norm": 0.636376678943634, + "learning_rate": 1.789077565857378e-05, + "loss": 0.0664, + "step": 1765 + }, + { + "epoch": 1.370054305663305, + "grad_norm": 0.47923678159713745, + "learning_rate": 1.786223392558976e-05, + "loss": 0.0108, + "step": 1766 + }, + { + "epoch": 1.3708301008533748, + "grad_norm": 8.116889953613281, + "learning_rate": 1.7833696784915953e-05, + "loss": 0.5007, + "step": 1767 + }, + { + "epoch": 1.3716058960434445, + "grad_norm": 4.673745155334473, + "learning_rate": 1.780516433098571e-05, + "loss": 0.2386, + "step": 1768 + }, + { + "epoch": 1.3723816912335143, + "grad_norm": 12.978338241577148, + "learning_rate": 1.777663665821687e-05, + "loss": 0.873, + "step": 1769 + }, + { + "epoch": 1.373157486423584, + "grad_norm": 4.008668899536133, + "learning_rate": 1.7748113861011457e-05, + "loss": 0.1004, + "step": 1770 + }, + { + "epoch": 1.373933281613654, + "grad_norm": 8.647369384765625, + "learning_rate": 1.7719596033755355e-05, + "loss": 0.4765, + "step": 1771 + }, + { + "epoch": 1.3747090768037238, + "grad_norm": 7.652889728546143, + "learning_rate": 1.7691083270818017e-05, + "loss": 0.3339, + "step": 1772 + }, + { + "epoch": 1.3754848719937938, + "grad_norm": 6.335665225982666, + "learning_rate": 1.766257566655211e-05, + "loss": 0.3352, + "step": 1773 + }, + { + "epoch": 1.3762606671838635, + "grad_norm": 7.106398582458496, + "learning_rate": 1.7634073315293263e-05, + "loss": 0.2759, + "step": 1774 + }, + { + "epoch": 1.3770364623739333, + "grad_norm": 4.669564247131348, + "learning_rate": 1.760557631135969e-05, + "loss": 0.2888, + "step": 1775 + }, + { + "epoch": 1.377812257564003, + "grad_norm": 8.145456314086914, + "learning_rate": 1.7577084749051934e-05, + "loss": 0.4233, + "step": 1776 + }, + { + "epoch": 1.3785880527540728, + "grad_norm": 7.680240631103516, + "learning_rate": 1.7548598722652516e-05, + "loss": 0.3015, + "step": 1777 + }, + { + "epoch": 1.3793638479441428, + "grad_norm": 3.4261608123779297, + "learning_rate": 1.7520118326425652e-05, + "loss": 0.0663, + "step": 1778 + }, + { + "epoch": 1.3801396431342126, + "grad_norm": 8.44232177734375, + "learning_rate": 1.7491643654616916e-05, + "loss": 0.5453, + "step": 1779 + }, + { + "epoch": 1.3809154383242823, + "grad_norm": 0.5656716227531433, + "learning_rate": 1.746317480145294e-05, + "loss": 0.0111, + "step": 1780 + }, + { + "epoch": 1.3809154383242823, + "eval_Qnli-dev_cosine_accuracy": 0.69140625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8344383239746094, + "eval_Qnli-dev_cosine_ap": 0.713570906268918, + "eval_Qnli-dev_cosine_f1": 0.6895424836601307, + "eval_Qnli-dev_cosine_f1_threshold": 0.7196956276893616, + "eval_Qnli-dev_cosine_precision": 0.5611702127659575, + "eval_Qnli-dev_cosine_recall": 0.8940677966101694, + "eval_Qnli-dev_dot_accuracy": 0.658203125, + "eval_Qnli-dev_dot_accuracy_threshold": 402.57080078125, + "eval_Qnli-dev_dot_ap": 0.6211282093852062, + "eval_Qnli-dev_dot_f1": 0.6797583081570997, + "eval_Qnli-dev_dot_f1_threshold": 311.3315124511719, + "eval_Qnli-dev_dot_precision": 0.528169014084507, + "eval_Qnli-dev_dot_recall": 0.9533898305084746, + "eval_Qnli-dev_euclidean_accuracy": 0.68359375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 14.798521041870117, + "eval_Qnli-dev_euclidean_ap": 0.7181824318635213, + "eval_Qnli-dev_euclidean_f1": 0.6873977086743045, + "eval_Qnli-dev_euclidean_f1_threshold": 16.626020431518555, + "eval_Qnli-dev_euclidean_precision": 0.56, + "eval_Qnli-dev_euclidean_recall": 0.8898305084745762, + "eval_Qnli-dev_manhattan_accuracy": 0.681640625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 237.8972625732422, + "eval_Qnli-dev_manhattan_ap": 0.7189290163755491, + "eval_Qnli-dev_manhattan_f1": 0.6895424836601307, + "eval_Qnli-dev_manhattan_f1_threshold": 301.86907958984375, + "eval_Qnli-dev_manhattan_precision": 0.5611702127659575, + "eval_Qnli-dev_manhattan_recall": 0.8940677966101694, + "eval_Qnli-dev_max_accuracy": 0.69140625, + "eval_Qnli-dev_max_accuracy_threshold": 402.57080078125, + "eval_Qnli-dev_max_ap": 0.7189290163755491, + "eval_Qnli-dev_max_f1": 0.6895424836601307, + "eval_Qnli-dev_max_f1_threshold": 311.3315124511719, + "eval_Qnli-dev_max_precision": 0.5611702127659575, + "eval_Qnli-dev_max_recall": 0.9533898305084746, + "eval_allNLI-dev_cosine_accuracy": 0.724609375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8609904050827026, + "eval_allNLI-dev_cosine_ap": 0.6056731062004843, + "eval_allNLI-dev_cosine_f1": 0.6165413533834587, + "eval_allNLI-dev_cosine_f1_threshold": 0.8228631019592285, + "eval_allNLI-dev_cosine_precision": 0.5442477876106194, + "eval_allNLI-dev_cosine_recall": 0.7109826589595376, + "eval_allNLI-dev_dot_accuracy": 0.6875, + "eval_allNLI-dev_dot_accuracy_threshold": 439.8603210449219, + "eval_allNLI-dev_dot_ap": 0.5093886392294766, + "eval_allNLI-dev_dot_f1": 0.5714285714285714, + "eval_allNLI-dev_dot_f1_threshold": 347.76812744140625, + "eval_allNLI-dev_dot_precision": 0.44884488448844884, + "eval_allNLI-dev_dot_recall": 0.7861271676300579, + "eval_allNLI-dev_euclidean_accuracy": 0.72265625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.70199203491211, + "eval_allNLI-dev_euclidean_ap": 0.6097019653773605, + "eval_allNLI-dev_euclidean_f1": 0.6187050359712231, + "eval_allNLI-dev_euclidean_f1_threshold": 13.009403228759766, + "eval_allNLI-dev_euclidean_precision": 0.5286885245901639, + "eval_allNLI-dev_euclidean_recall": 0.7456647398843931, + "eval_allNLI-dev_manhattan_accuracy": 0.728515625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 200.24722290039062, + "eval_allNLI-dev_manhattan_ap": 0.6086689939812514, + "eval_allNLI-dev_manhattan_f1": 0.6240409207161125, + "eval_allNLI-dev_manhattan_f1_threshold": 224.26364135742188, + "eval_allNLI-dev_manhattan_precision": 0.5596330275229358, + "eval_allNLI-dev_manhattan_recall": 0.7052023121387283, + "eval_allNLI-dev_max_accuracy": 0.728515625, + "eval_allNLI-dev_max_accuracy_threshold": 439.8603210449219, + "eval_allNLI-dev_max_ap": 0.6097019653773605, + "eval_allNLI-dev_max_f1": 0.6240409207161125, + "eval_allNLI-dev_max_f1_threshold": 347.76812744140625, + "eval_allNLI-dev_max_precision": 0.5596330275229358, + "eval_allNLI-dev_max_recall": 0.7861271676300579, + "eval_sequential_score": 0.7189290163755491, + "eval_sts-test_pearson_cosine": 0.8269618790420523, + "eval_sts-test_pearson_dot": 0.7772969207489703, + "eval_sts-test_pearson_euclidean": 0.8674292760715094, + "eval_sts-test_pearson_manhattan": 0.866279462582368, + "eval_sts-test_pearson_max": 0.8674292760715094, + "eval_sts-test_spearman_cosine": 0.8757597650477315, + "eval_sts-test_spearman_dot": 0.7726652101500908, + "eval_sts-test_spearman_euclidean": 0.8712502251885813, + "eval_sts-test_spearman_manhattan": 0.8719343824727331, + "eval_sts-test_spearman_max": 0.8757597650477315, + "eval_vitaminc-pairs_loss": 3.9482483863830566, + "eval_vitaminc-pairs_runtime": 5.5205, + "eval_vitaminc-pairs_samples_per_second": 23.186, + "eval_vitaminc-pairs_steps_per_second": 0.362, + "step": 1780 + }, + { + "epoch": 1.3809154383242823, + "eval_negation-triplets_loss": 1.6817325353622437, + "eval_negation-triplets_runtime": 1.0907, + "eval_negation-triplets_samples_per_second": 117.352, + "eval_negation-triplets_steps_per_second": 1.834, + "step": 1780 + }, + { + "epoch": 1.3809154383242823, + "eval_scitail-pairs-pos_loss": 0.14056001603603363, + "eval_scitail-pairs-pos_runtime": 1.3303, + "eval_scitail-pairs-pos_samples_per_second": 96.219, + "eval_scitail-pairs-pos_steps_per_second": 1.503, + "step": 1780 + }, + { + "epoch": 1.3809154383242823, + "eval_scitail-pairs-qa_loss": 0.005900491029024124, + "eval_scitail-pairs-qa_runtime": 0.7861, + "eval_scitail-pairs-qa_samples_per_second": 162.837, + "eval_scitail-pairs-qa_steps_per_second": 2.544, + "step": 1780 + }, + { + "epoch": 1.3809154383242823, + "eval_xsum-pairs_loss": 0.09856506437063217, + "eval_xsum-pairs_runtime": 3.9478, + "eval_xsum-pairs_samples_per_second": 32.423, + "eval_xsum-pairs_steps_per_second": 0.507, + "step": 1780 + }, + { + "epoch": 1.3809154383242823, + "eval_sciq_pairs_loss": 0.04453403502702713, + "eval_sciq_pairs_runtime": 6.1881, + "eval_sciq_pairs_samples_per_second": 20.685, + "eval_sciq_pairs_steps_per_second": 0.323, + "step": 1780 + }, + { + "epoch": 1.3809154383242823, + "eval_qasc_pairs_loss": 0.24179349839687347, + "eval_qasc_pairs_runtime": 0.9373, + "eval_qasc_pairs_samples_per_second": 136.557, + "eval_qasc_pairs_steps_per_second": 2.134, + "step": 1780 + }, + { + "epoch": 1.3809154383242823, + "eval_openbookqa_pairs_loss": 1.0625941753387451, + "eval_openbookqa_pairs_runtime": 0.8898, + "eval_openbookqa_pairs_samples_per_second": 143.855, + "eval_openbookqa_pairs_steps_per_second": 2.248, + "step": 1780 + }, + { + "epoch": 1.3809154383242823, + "eval_msmarco_pairs_loss": 0.36683905124664307, + "eval_msmarco_pairs_runtime": 1.9994, + "eval_msmarco_pairs_samples_per_second": 64.02, + "eval_msmarco_pairs_steps_per_second": 1.0, + "step": 1780 + }, + { + "epoch": 1.3809154383242823, + "eval_nq_pairs_loss": 0.5549561381340027, + "eval_nq_pairs_runtime": 3.3522, + "eval_nq_pairs_samples_per_second": 38.184, + "eval_nq_pairs_steps_per_second": 0.597, + "step": 1780 + }, + { + "epoch": 1.3809154383242823, + "eval_trivia_pairs_loss": 0.7266644239425659, + "eval_trivia_pairs_runtime": 4.5071, + "eval_trivia_pairs_samples_per_second": 28.4, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 1780 + }, + { + "epoch": 1.3809154383242823, + "eval_gooaq_pairs_loss": 0.5850292444229126, + "eval_gooaq_pairs_runtime": 1.5268, + "eval_gooaq_pairs_samples_per_second": 83.833, + "eval_gooaq_pairs_steps_per_second": 1.31, + "step": 1780 + }, + { + "epoch": 1.3809154383242823, + "eval_paws-pos_loss": 0.030058603733778, + "eval_paws-pos_runtime": 1.0474, + "eval_paws-pos_samples_per_second": 122.212, + "eval_paws-pos_steps_per_second": 1.91, + "step": 1780 + }, + { + "epoch": 1.3809154383242823, + "eval_global_dataset_loss": 0.393206924200058, + "eval_global_dataset_runtime": 19.5009, + "eval_global_dataset_samples_per_second": 21.332, + "eval_global_dataset_steps_per_second": 0.359, + "step": 1780 + }, + { + "epoch": 1.3816912335143523, + "grad_norm": 10.680549621582031, + "learning_rate": 1.74347118611411e-05, + "loss": 0.9002, + "step": 1781 + }, + { + "epoch": 1.382467028704422, + "grad_norm": 10.885297775268555, + "learning_rate": 1.7406254927869212e-05, + "loss": 0.6257, + "step": 1782 + }, + { + "epoch": 1.3832428238944918, + "grad_norm": 10.978306770324707, + "learning_rate": 1.7377804095805216e-05, + "loss": 0.7902, + "step": 1783 + }, + { + "epoch": 1.3840186190845616, + "grad_norm": 8.765235900878906, + "learning_rate": 1.734935945909685e-05, + "loss": 0.4995, + "step": 1784 + }, + { + "epoch": 1.3847944142746316, + "grad_norm": 8.499723434448242, + "learning_rate": 1.7320921111871368e-05, + "loss": 0.4442, + "step": 1785 + }, + { + "epoch": 1.3855702094647013, + "grad_norm": 6.003756523132324, + "learning_rate": 1.7292489148235185e-05, + "loss": 0.2353, + "step": 1786 + }, + { + "epoch": 1.386346004654771, + "grad_norm": 6.424333095550537, + "learning_rate": 1.726406366227362e-05, + "loss": 0.2383, + "step": 1787 + }, + { + "epoch": 1.387121799844841, + "grad_norm": 5.591697692871094, + "learning_rate": 1.7235644748050536e-05, + "loss": 0.1427, + "step": 1788 + }, + { + "epoch": 1.3878975950349108, + "grad_norm": 8.849912643432617, + "learning_rate": 1.7207232499608067e-05, + "loss": 0.5918, + "step": 1789 + }, + { + "epoch": 1.3886733902249806, + "grad_norm": 7.9435906410217285, + "learning_rate": 1.7178827010966278e-05, + "loss": 0.6451, + "step": 1790 + }, + { + "epoch": 1.3894491854150504, + "grad_norm": 8.700857162475586, + "learning_rate": 1.715042837612286e-05, + "loss": 0.5438, + "step": 1791 + }, + { + "epoch": 1.3902249806051201, + "grad_norm": 9.01171875, + "learning_rate": 1.712203668905283e-05, + "loss": 0.3999, + "step": 1792 + }, + { + "epoch": 1.39100077579519, + "grad_norm": 6.732156276702881, + "learning_rate": 1.7093652043708215e-05, + "loss": 0.3888, + "step": 1793 + }, + { + "epoch": 1.3917765709852599, + "grad_norm": 5.008310794830322, + "learning_rate": 1.706527453401775e-05, + "loss": 0.2781, + "step": 1794 + }, + { + "epoch": 1.3925523661753298, + "grad_norm": 5.369287490844727, + "learning_rate": 1.7036904253886536e-05, + "loss": 0.1583, + "step": 1795 + }, + { + "epoch": 1.3933281613653996, + "grad_norm": 4.665939807891846, + "learning_rate": 1.700854129719576e-05, + "loss": 0.2095, + "step": 1796 + }, + { + "epoch": 1.3941039565554694, + "grad_norm": 1.1131080389022827, + "learning_rate": 1.6980185757802373e-05, + "loss": 0.0132, + "step": 1797 + }, + { + "epoch": 1.3948797517455391, + "grad_norm": 4.151784420013428, + "learning_rate": 1.6951837729538792e-05, + "loss": 0.0737, + "step": 1798 + }, + { + "epoch": 1.395655546935609, + "grad_norm": 9.108835220336914, + "learning_rate": 1.6923497306212565e-05, + "loss": 0.4911, + "step": 1799 + }, + { + "epoch": 1.3964313421256789, + "grad_norm": 3.4478728771209717, + "learning_rate": 1.6895164581606085e-05, + "loss": 0.0831, + "step": 1800 + }, + { + "epoch": 1.3964313421256789, + "eval_Qnli-dev_cosine_accuracy": 0.689453125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8336964845657349, + "eval_Qnli-dev_cosine_ap": 0.7150954851838882, + "eval_Qnli-dev_cosine_f1": 0.6956521739130435, + "eval_Qnli-dev_cosine_f1_threshold": 0.7298338413238525, + "eval_Qnli-dev_cosine_precision": 0.574585635359116, + "eval_Qnli-dev_cosine_recall": 0.8813559322033898, + "eval_Qnli-dev_dot_accuracy": 0.65625, + "eval_Qnli-dev_dot_accuracy_threshold": 401.4566955566406, + "eval_Qnli-dev_dot_ap": 0.6258851257827565, + "eval_Qnli-dev_dot_f1": 0.6767830045523521, + "eval_Qnli-dev_dot_f1_threshold": 314.2878723144531, + "eval_Qnli-dev_dot_precision": 0.5271867612293144, + "eval_Qnli-dev_dot_recall": 0.9449152542372882, + "eval_Qnli-dev_euclidean_accuracy": 0.681640625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 14.45042610168457, + "eval_Qnli-dev_euclidean_ap": 0.7170386966688913, + "eval_Qnli-dev_euclidean_f1": 0.6902086677367576, + "eval_Qnli-dev_euclidean_f1_threshold": 16.819618225097656, + "eval_Qnli-dev_euclidean_precision": 0.5555555555555556, + "eval_Qnli-dev_euclidean_recall": 0.9110169491525424, + "eval_Qnli-dev_manhattan_accuracy": 0.677734375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 255.0120391845703, + "eval_Qnli-dev_manhattan_ap": 0.7200489482781656, + "eval_Qnli-dev_manhattan_f1": 0.6934189406099519, + "eval_Qnli-dev_manhattan_f1_threshold": 303.8094482421875, + "eval_Qnli-dev_manhattan_precision": 0.5581395348837209, + "eval_Qnli-dev_manhattan_recall": 0.9152542372881356, + "eval_Qnli-dev_max_accuracy": 0.689453125, + "eval_Qnli-dev_max_accuracy_threshold": 401.4566955566406, + "eval_Qnli-dev_max_ap": 0.7200489482781656, + "eval_Qnli-dev_max_f1": 0.6956521739130435, + "eval_Qnli-dev_max_f1_threshold": 314.2878723144531, + "eval_Qnli-dev_max_precision": 0.574585635359116, + "eval_Qnli-dev_max_recall": 0.9449152542372882, + "eval_allNLI-dev_cosine_accuracy": 0.7265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8573634624481201, + "eval_allNLI-dev_cosine_ap": 0.6076363324003046, + "eval_allNLI-dev_cosine_f1": 0.626865671641791, + "eval_allNLI-dev_cosine_f1_threshold": 0.8096942901611328, + "eval_allNLI-dev_cosine_precision": 0.5502183406113537, + "eval_allNLI-dev_cosine_recall": 0.7283236994219653, + "eval_allNLI-dev_dot_accuracy": 0.6875, + "eval_allNLI-dev_dot_accuracy_threshold": 377.677490234375, + "eval_allNLI-dev_dot_ap": 0.5087142011975201, + "eval_allNLI-dev_dot_f1": 0.5793103448275863, + "eval_allNLI-dev_dot_f1_threshold": 349.49420166015625, + "eval_allNLI-dev_dot_precision": 0.48091603053435117, + "eval_allNLI-dev_dot_recall": 0.7283236994219653, + "eval_allNLI-dev_euclidean_accuracy": 0.732421875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.052558898925781, + "eval_allNLI-dev_euclidean_ap": 0.6138325149634452, + "eval_allNLI-dev_euclidean_f1": 0.6265060240963856, + "eval_allNLI-dev_euclidean_f1_threshold": 13.346977233886719, + "eval_allNLI-dev_euclidean_precision": 0.5371900826446281, + "eval_allNLI-dev_euclidean_recall": 0.7514450867052023, + "eval_allNLI-dev_manhattan_accuracy": 0.734375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 224.94667053222656, + "eval_allNLI-dev_manhattan_ap": 0.6125662203517585, + "eval_allNLI-dev_manhattan_f1": 0.6304347826086957, + "eval_allNLI-dev_manhattan_f1_threshold": 224.94667053222656, + "eval_allNLI-dev_manhattan_precision": 0.5948717948717949, + "eval_allNLI-dev_manhattan_recall": 0.6705202312138728, + "eval_allNLI-dev_max_accuracy": 0.734375, + "eval_allNLI-dev_max_accuracy_threshold": 377.677490234375, + "eval_allNLI-dev_max_ap": 0.6138325149634452, + "eval_allNLI-dev_max_f1": 0.6304347826086957, + "eval_allNLI-dev_max_f1_threshold": 349.49420166015625, + "eval_allNLI-dev_max_precision": 0.5948717948717949, + "eval_allNLI-dev_max_recall": 0.7514450867052023, + "eval_sequential_score": 0.7200489482781656, + "eval_sts-test_pearson_cosine": 0.8311435641864657, + "eval_sts-test_pearson_dot": 0.7955147226773802, + "eval_sts-test_pearson_euclidean": 0.869047473852348, + "eval_sts-test_pearson_manhattan": 0.8672954236203759, + "eval_sts-test_pearson_max": 0.869047473852348, + "eval_sts-test_spearman_cosine": 0.8752627127874572, + "eval_sts-test_spearman_dot": 0.7939026820148233, + "eval_sts-test_spearman_euclidean": 0.8705060117116423, + "eval_sts-test_spearman_manhattan": 0.8705819546022237, + "eval_sts-test_spearman_max": 0.8752627127874572, + "eval_vitaminc-pairs_loss": 3.8420679569244385, + "eval_vitaminc-pairs_runtime": 5.5169, + "eval_vitaminc-pairs_samples_per_second": 23.202, + "eval_vitaminc-pairs_steps_per_second": 0.363, + "step": 1800 + }, + { + "epoch": 1.3964313421256789, + "eval_negation-triplets_loss": 1.6764999628067017, + "eval_negation-triplets_runtime": 1.092, + "eval_negation-triplets_samples_per_second": 117.221, + "eval_negation-triplets_steps_per_second": 1.832, + "step": 1800 + }, + { + "epoch": 1.3964313421256789, + "eval_scitail-pairs-pos_loss": 0.14320363104343414, + "eval_scitail-pairs-pos_runtime": 1.3334, + "eval_scitail-pairs-pos_samples_per_second": 95.992, + "eval_scitail-pairs-pos_steps_per_second": 1.5, + "step": 1800 + }, + { + "epoch": 1.3964313421256789, + "eval_scitail-pairs-qa_loss": 0.007326332852244377, + "eval_scitail-pairs-qa_runtime": 0.7891, + "eval_scitail-pairs-qa_samples_per_second": 162.202, + "eval_scitail-pairs-qa_steps_per_second": 2.534, + "step": 1800 + }, + { + "epoch": 1.3964313421256789, + "eval_xsum-pairs_loss": 0.08128401637077332, + "eval_xsum-pairs_runtime": 3.9358, + "eval_xsum-pairs_samples_per_second": 32.522, + "eval_xsum-pairs_steps_per_second": 0.508, + "step": 1800 + }, + { + "epoch": 1.3964313421256789, + "eval_sciq_pairs_loss": 0.04576320946216583, + "eval_sciq_pairs_runtime": 6.1897, + "eval_sciq_pairs_samples_per_second": 20.679, + "eval_sciq_pairs_steps_per_second": 0.323, + "step": 1800 + }, + { + "epoch": 1.3964313421256789, + "eval_qasc_pairs_loss": 0.21950416266918182, + "eval_qasc_pairs_runtime": 0.9619, + "eval_qasc_pairs_samples_per_second": 133.074, + "eval_qasc_pairs_steps_per_second": 2.079, + "step": 1800 + }, + { + "epoch": 1.3964313421256789, + "eval_openbookqa_pairs_loss": 1.048283576965332, + "eval_openbookqa_pairs_runtime": 0.8947, + "eval_openbookqa_pairs_samples_per_second": 143.067, + "eval_openbookqa_pairs_steps_per_second": 2.235, + "step": 1800 + }, + { + "epoch": 1.3964313421256789, + "eval_msmarco_pairs_loss": 0.36549484729766846, + "eval_msmarco_pairs_runtime": 1.9854, + "eval_msmarco_pairs_samples_per_second": 64.472, + "eval_msmarco_pairs_steps_per_second": 1.007, + "step": 1800 + }, + { + "epoch": 1.3964313421256789, + "eval_nq_pairs_loss": 0.5500632524490356, + "eval_nq_pairs_runtime": 3.35, + "eval_nq_pairs_samples_per_second": 38.209, + "eval_nq_pairs_steps_per_second": 0.597, + "step": 1800 + }, + { + "epoch": 1.3964313421256789, + "eval_trivia_pairs_loss": 0.7217636108398438, + "eval_trivia_pairs_runtime": 4.4922, + "eval_trivia_pairs_samples_per_second": 28.494, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 1800 + }, + { + "epoch": 1.3964313421256789, + "eval_gooaq_pairs_loss": 0.6220347881317139, + "eval_gooaq_pairs_runtime": 1.4848, + "eval_gooaq_pairs_samples_per_second": 86.204, + "eval_gooaq_pairs_steps_per_second": 1.347, + "step": 1800 + }, + { + "epoch": 1.3964313421256789, + "eval_paws-pos_loss": 0.02932680770754814, + "eval_paws-pos_runtime": 1.0563, + "eval_paws-pos_samples_per_second": 121.177, + "eval_paws-pos_steps_per_second": 1.893, + "step": 1800 + }, + { + "epoch": 1.3964313421256789, + "eval_global_dataset_loss": 0.3950449228286743, + "eval_global_dataset_runtime": 19.454, + "eval_global_dataset_samples_per_second": 21.384, + "eval_global_dataset_steps_per_second": 0.36, + "step": 1800 + }, + { + "epoch": 1.3972071373157486, + "grad_norm": 4.628012657165527, + "learning_rate": 1.6866839649476254e-05, + "loss": 0.1333, + "step": 1801 + }, + { + "epoch": 1.3979829325058184, + "grad_norm": 5.75248908996582, + "learning_rate": 1.6838522603554206e-05, + "loss": 0.2781, + "step": 1802 + }, + { + "epoch": 1.3987587276958884, + "grad_norm": 9.245781898498535, + "learning_rate": 1.6810213537544956e-05, + "loss": 0.6254, + "step": 1803 + }, + { + "epoch": 1.3995345228859581, + "grad_norm": 9.885970115661621, + "learning_rate": 1.6781912545127137e-05, + "loss": 0.5593, + "step": 1804 + }, + { + "epoch": 1.400310318076028, + "grad_norm": 10.413599967956543, + "learning_rate": 1.6753619719952652e-05, + "loss": 0.5567, + "step": 1805 + }, + { + "epoch": 1.4010861132660977, + "grad_norm": 10.420858383178711, + "learning_rate": 1.6725335155646376e-05, + "loss": 0.6962, + "step": 1806 + }, + { + "epoch": 1.4018619084561676, + "grad_norm": 5.587803840637207, + "learning_rate": 1.6697058945805857e-05, + "loss": 0.4131, + "step": 1807 + }, + { + "epoch": 1.4026377036462374, + "grad_norm": 4.07176399230957, + "learning_rate": 1.666879118400098e-05, + "loss": 0.0976, + "step": 1808 + }, + { + "epoch": 1.4034134988363072, + "grad_norm": 6.005397796630859, + "learning_rate": 1.6640531963773697e-05, + "loss": 0.1917, + "step": 1809 + }, + { + "epoch": 1.4041892940263772, + "grad_norm": 9.860136985778809, + "learning_rate": 1.6612281378637678e-05, + "loss": 0.56, + "step": 1810 + }, + { + "epoch": 1.404965089216447, + "grad_norm": 13.69825553894043, + "learning_rate": 1.6584039522078023e-05, + "loss": 1.8062, + "step": 1811 + }, + { + "epoch": 1.4057408844065167, + "grad_norm": 6.53242826461792, + "learning_rate": 1.655580648755095e-05, + "loss": 0.4568, + "step": 1812 + }, + { + "epoch": 1.4065166795965864, + "grad_norm": 9.471614837646484, + "learning_rate": 1.6527582368483477e-05, + "loss": 0.5773, + "step": 1813 + }, + { + "epoch": 1.4072924747866562, + "grad_norm": 4.8946075439453125, + "learning_rate": 1.649936725827313e-05, + "loss": 0.2272, + "step": 1814 + }, + { + "epoch": 1.4080682699767262, + "grad_norm": 5.489307880401611, + "learning_rate": 1.6471161250287613e-05, + "loss": 0.142, + "step": 1815 + }, + { + "epoch": 1.408844065166796, + "grad_norm": 9.541376113891602, + "learning_rate": 1.6442964437864517e-05, + "loss": 0.5551, + "step": 1816 + }, + { + "epoch": 1.4096198603568657, + "grad_norm": 4.954698085784912, + "learning_rate": 1.6414776914311004e-05, + "loss": 0.1317, + "step": 1817 + }, + { + "epoch": 1.4103956555469357, + "grad_norm": 3.9457144737243652, + "learning_rate": 1.638659877290349e-05, + "loss": 0.1397, + "step": 1818 + }, + { + "epoch": 1.4111714507370055, + "grad_norm": 3.3471503257751465, + "learning_rate": 1.6358430106887346e-05, + "loss": 0.1382, + "step": 1819 + }, + { + "epoch": 1.4119472459270752, + "grad_norm": 3.939849853515625, + "learning_rate": 1.6330271009476597e-05, + "loss": 0.1645, + "step": 1820 + }, + { + "epoch": 1.4119472459270752, + "eval_Qnli-dev_cosine_accuracy": 0.677734375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8577120304107666, + "eval_Qnli-dev_cosine_ap": 0.709474527506375, + "eval_Qnli-dev_cosine_f1": 0.6919275123558485, + "eval_Qnli-dev_cosine_f1_threshold": 0.7467669248580933, + "eval_Qnli-dev_cosine_precision": 0.5660377358490566, + "eval_Qnli-dev_cosine_recall": 0.8898305084745762, + "eval_Qnli-dev_dot_accuracy": 0.6640625, + "eval_Qnli-dev_dot_accuracy_threshold": 417.42193603515625, + "eval_Qnli-dev_dot_ap": 0.619294047071016, + "eval_Qnli-dev_dot_f1": 0.6758409785932722, + "eval_Qnli-dev_dot_f1_threshold": 338.77923583984375, + "eval_Qnli-dev_dot_precision": 0.5287081339712919, + "eval_Qnli-dev_dot_recall": 0.9364406779661016, + "eval_Qnli-dev_euclidean_accuracy": 0.681640625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 14.043766021728516, + "eval_Qnli-dev_euclidean_ap": 0.7102445115544279, + "eval_Qnli-dev_euclidean_f1": 0.697452229299363, + "eval_Qnli-dev_euclidean_f1_threshold": 16.373180389404297, + "eval_Qnli-dev_euclidean_precision": 0.5586734693877551, + "eval_Qnli-dev_euclidean_recall": 0.9279661016949152, + "eval_Qnli-dev_manhattan_accuracy": 0.677734375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 237.9270782470703, + "eval_Qnli-dev_manhattan_ap": 0.7152206563542711, + "eval_Qnli-dev_manhattan_f1": 0.6914728682170543, + "eval_Qnli-dev_manhattan_f1_threshold": 305.93487548828125, + "eval_Qnli-dev_manhattan_precision": 0.5452322738386308, + "eval_Qnli-dev_manhattan_recall": 0.9449152542372882, + "eval_Qnli-dev_max_accuracy": 0.681640625, + "eval_Qnli-dev_max_accuracy_threshold": 417.42193603515625, + "eval_Qnli-dev_max_ap": 0.7152206563542711, + "eval_Qnli-dev_max_f1": 0.697452229299363, + "eval_Qnli-dev_max_f1_threshold": 338.77923583984375, + "eval_Qnli-dev_max_precision": 0.5660377358490566, + "eval_Qnli-dev_max_recall": 0.9449152542372882, + "eval_allNLI-dev_cosine_accuracy": 0.734375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8637614250183105, + "eval_allNLI-dev_cosine_ap": 0.6122250225550318, + "eval_allNLI-dev_cosine_f1": 0.6247086247086246, + "eval_allNLI-dev_cosine_f1_threshold": 0.8137850761413574, + "eval_allNLI-dev_cosine_precision": 0.5234375, + "eval_allNLI-dev_cosine_recall": 0.7745664739884393, + "eval_allNLI-dev_dot_accuracy": 0.6796875, + "eval_allNLI-dev_dot_accuracy_threshold": 475.9347839355469, + "eval_allNLI-dev_dot_ap": 0.5067427135509602, + "eval_allNLI-dev_dot_f1": 0.5757009345794393, + "eval_allNLI-dev_dot_f1_threshold": 342.48333740234375, + "eval_allNLI-dev_dot_precision": 0.425414364640884, + "eval_allNLI-dev_dot_recall": 0.8901734104046243, + "eval_allNLI-dev_euclidean_accuracy": 0.732421875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.192228317260742, + "eval_allNLI-dev_euclidean_ap": 0.6165918172594097, + "eval_allNLI-dev_euclidean_f1": 0.6328502415458938, + "eval_allNLI-dev_euclidean_f1_threshold": 12.864824295043945, + "eval_allNLI-dev_euclidean_precision": 0.5435684647302904, + "eval_allNLI-dev_euclidean_recall": 0.7572254335260116, + "eval_allNLI-dev_manhattan_accuracy": 0.732421875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 203.53961181640625, + "eval_allNLI-dev_manhattan_ap": 0.6153082028936737, + "eval_allNLI-dev_manhattan_f1": 0.6341463414634145, + "eval_allNLI-dev_manhattan_f1_threshold": 252.7510986328125, + "eval_allNLI-dev_manhattan_precision": 0.5143884892086331, + "eval_allNLI-dev_manhattan_recall": 0.8265895953757225, + "eval_allNLI-dev_max_accuracy": 0.734375, + "eval_allNLI-dev_max_accuracy_threshold": 475.9347839355469, + "eval_allNLI-dev_max_ap": 0.6165918172594097, + "eval_allNLI-dev_max_f1": 0.6341463414634145, + "eval_allNLI-dev_max_f1_threshold": 342.48333740234375, + "eval_allNLI-dev_max_precision": 0.5435684647302904, + "eval_allNLI-dev_max_recall": 0.8901734104046243, + "eval_sequential_score": 0.7152206563542711, + "eval_sts-test_pearson_cosine": 0.8267221648424257, + "eval_sts-test_pearson_dot": 0.7790538507127196, + "eval_sts-test_pearson_euclidean": 0.8639785392834488, + "eval_sts-test_pearson_manhattan": 0.8621201895961921, + "eval_sts-test_pearson_max": 0.8639785392834488, + "eval_sts-test_spearman_cosine": 0.8702317312305746, + "eval_sts-test_spearman_dot": 0.7736585091479344, + "eval_sts-test_spearman_euclidean": 0.8653355852242643, + "eval_sts-test_spearman_manhattan": 0.8651125449374896, + "eval_sts-test_spearman_max": 0.8702317312305746, + "eval_vitaminc-pairs_loss": 3.797407627105713, + "eval_vitaminc-pairs_runtime": 5.567, + "eval_vitaminc-pairs_samples_per_second": 22.993, + "eval_vitaminc-pairs_steps_per_second": 0.359, + "step": 1820 + }, + { + "epoch": 1.4119472459270752, + "eval_negation-triplets_loss": 1.615311861038208, + "eval_negation-triplets_runtime": 1.1565, + "eval_negation-triplets_samples_per_second": 110.679, + "eval_negation-triplets_steps_per_second": 1.729, + "step": 1820 + }, + { + "epoch": 1.4119472459270752, + "eval_scitail-pairs-pos_loss": 0.15228097140789032, + "eval_scitail-pairs-pos_runtime": 1.4761, + "eval_scitail-pairs-pos_samples_per_second": 86.716, + "eval_scitail-pairs-pos_steps_per_second": 1.355, + "step": 1820 + }, + { + "epoch": 1.4119472459270752, + "eval_scitail-pairs-qa_loss": 0.008484632708132267, + "eval_scitail-pairs-qa_runtime": 0.8123, + "eval_scitail-pairs-qa_samples_per_second": 157.584, + "eval_scitail-pairs-qa_steps_per_second": 2.462, + "step": 1820 + }, + { + "epoch": 1.4119472459270752, + "eval_xsum-pairs_loss": 0.07424400001764297, + "eval_xsum-pairs_runtime": 3.9452, + "eval_xsum-pairs_samples_per_second": 32.445, + "eval_xsum-pairs_steps_per_second": 0.507, + "step": 1820 + }, + { + "epoch": 1.4119472459270752, + "eval_sciq_pairs_loss": 0.04548928141593933, + "eval_sciq_pairs_runtime": 6.2639, + "eval_sciq_pairs_samples_per_second": 20.434, + "eval_sciq_pairs_steps_per_second": 0.319, + "step": 1820 + }, + { + "epoch": 1.4119472459270752, + "eval_qasc_pairs_loss": 0.21296152472496033, + "eval_qasc_pairs_runtime": 0.9594, + "eval_qasc_pairs_samples_per_second": 133.42, + "eval_qasc_pairs_steps_per_second": 2.085, + "step": 1820 + }, + { + "epoch": 1.4119472459270752, + "eval_openbookqa_pairs_loss": 1.0667941570281982, + "eval_openbookqa_pairs_runtime": 0.9023, + "eval_openbookqa_pairs_samples_per_second": 141.859, + "eval_openbookqa_pairs_steps_per_second": 2.217, + "step": 1820 + }, + { + "epoch": 1.4119472459270752, + "eval_msmarco_pairs_loss": 0.3414951264858246, + "eval_msmarco_pairs_runtime": 1.9865, + "eval_msmarco_pairs_samples_per_second": 64.434, + "eval_msmarco_pairs_steps_per_second": 1.007, + "step": 1820 + }, + { + "epoch": 1.4119472459270752, + "eval_nq_pairs_loss": 0.48314133286476135, + "eval_nq_pairs_runtime": 3.362, + "eval_nq_pairs_samples_per_second": 38.072, + "eval_nq_pairs_steps_per_second": 0.595, + "step": 1820 + }, + { + "epoch": 1.4119472459270752, + "eval_trivia_pairs_loss": 0.6327275037765503, + "eval_trivia_pairs_runtime": 4.5078, + "eval_trivia_pairs_samples_per_second": 28.395, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 1820 + }, + { + "epoch": 1.4119472459270752, + "eval_gooaq_pairs_loss": 0.5942782163619995, + "eval_gooaq_pairs_runtime": 1.4944, + "eval_gooaq_pairs_samples_per_second": 85.651, + "eval_gooaq_pairs_steps_per_second": 1.338, + "step": 1820 + }, + { + "epoch": 1.4119472459270752, + "eval_paws-pos_loss": 0.029733898118138313, + "eval_paws-pos_runtime": 1.0492, + "eval_paws-pos_samples_per_second": 122.001, + "eval_paws-pos_steps_per_second": 1.906, + "step": 1820 + }, + { + "epoch": 1.4119472459270752, + "eval_global_dataset_loss": 0.38853293657302856, + "eval_global_dataset_runtime": 19.5009, + "eval_global_dataset_samples_per_second": 21.332, + "eval_global_dataset_steps_per_second": 0.359, + "step": 1820 + }, + { + "epoch": 1.412723041117145, + "grad_norm": 6.45574951171875, + "learning_rate": 1.6302121573853595e-05, + "loss": 0.4355, + "step": 1821 + }, + { + "epoch": 1.413498836307215, + "grad_norm": 7.539233684539795, + "learning_rate": 1.6273981893168726e-05, + "loss": 0.4843, + "step": 1822 + }, + { + "epoch": 1.4142746314972847, + "grad_norm": 0.7213716506958008, + "learning_rate": 1.624585206054009e-05, + "loss": 0.015, + "step": 1823 + }, + { + "epoch": 1.4150504266873545, + "grad_norm": 6.887151718139648, + "learning_rate": 1.6217732169053197e-05, + "loss": 0.55, + "step": 1824 + }, + { + "epoch": 1.4158262218774245, + "grad_norm": 9.043567657470703, + "learning_rate": 1.618962231176066e-05, + "loss": 0.4399, + "step": 1825 + }, + { + "epoch": 1.4166020170674942, + "grad_norm": 13.23807430267334, + "learning_rate": 1.6161522581681904e-05, + "loss": 1.8768, + "step": 1826 + }, + { + "epoch": 1.417377812257564, + "grad_norm": 4.3568925857543945, + "learning_rate": 1.613343307180282e-05, + "loss": 0.1297, + "step": 1827 + }, + { + "epoch": 1.4181536074476337, + "grad_norm": 11.774724960327148, + "learning_rate": 1.6105353875075496e-05, + "loss": 0.8268, + "step": 1828 + }, + { + "epoch": 1.4189294026377037, + "grad_norm": 6.948408126831055, + "learning_rate": 1.6077285084417873e-05, + "loss": 0.2522, + "step": 1829 + }, + { + "epoch": 1.4197051978277735, + "grad_norm": 5.333669185638428, + "learning_rate": 1.604922679271347e-05, + "loss": 0.1362, + "step": 1830 + }, + { + "epoch": 1.4204809930178433, + "grad_norm": 0.7340044975280762, + "learning_rate": 1.602117909281107e-05, + "loss": 0.033, + "step": 1831 + }, + { + "epoch": 1.4212567882079132, + "grad_norm": 0.8986733555793762, + "learning_rate": 1.5993142077524395e-05, + "loss": 0.0242, + "step": 1832 + }, + { + "epoch": 1.422032583397983, + "grad_norm": 4.408289909362793, + "learning_rate": 1.5965115839631812e-05, + "loss": 0.1334, + "step": 1833 + }, + { + "epoch": 1.4228083785880528, + "grad_norm": 9.30521011352539, + "learning_rate": 1.5937100471876024e-05, + "loss": 0.4268, + "step": 1834 + }, + { + "epoch": 1.4235841737781225, + "grad_norm": 3.0090644359588623, + "learning_rate": 1.590909606696376e-05, + "loss": 0.1184, + "step": 1835 + }, + { + "epoch": 1.4243599689681923, + "grad_norm": 6.331893444061279, + "learning_rate": 1.588110271756548e-05, + "loss": 0.2224, + "step": 1836 + }, + { + "epoch": 1.4251357641582623, + "grad_norm": 7.605957984924316, + "learning_rate": 1.5853120516315063e-05, + "loss": 0.3772, + "step": 1837 + }, + { + "epoch": 1.425911559348332, + "grad_norm": 10.143529891967773, + "learning_rate": 1.5825149555809477e-05, + "loss": 0.4948, + "step": 1838 + }, + { + "epoch": 1.4266873545384018, + "grad_norm": 10.375722885131836, + "learning_rate": 1.579718992860851e-05, + "loss": 0.637, + "step": 1839 + }, + { + "epoch": 1.4274631497284718, + "grad_norm": 10.450224876403809, + "learning_rate": 1.5769241727234452e-05, + "loss": 0.5175, + "step": 1840 + }, + { + "epoch": 1.4274631497284718, + "eval_Qnli-dev_cosine_accuracy": 0.673828125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8249573707580566, + "eval_Qnli-dev_cosine_ap": 0.7006195939027415, + "eval_Qnli-dev_cosine_f1": 0.6883720930232559, + "eval_Qnli-dev_cosine_f1_threshold": 0.6899151802062988, + "eval_Qnli-dev_cosine_precision": 0.5427872860635696, + "eval_Qnli-dev_cosine_recall": 0.940677966101695, + "eval_Qnli-dev_dot_accuracy": 0.6640625, + "eval_Qnli-dev_dot_accuracy_threshold": 399.835205078125, + "eval_Qnli-dev_dot_ap": 0.6008559480569156, + "eval_Qnli-dev_dot_f1": 0.6756756756756758, + "eval_Qnli-dev_dot_f1_threshold": 313.7658996582031, + "eval_Qnli-dev_dot_precision": 0.5232558139534884, + "eval_Qnli-dev_dot_recall": 0.9533898305084746, + "eval_Qnli-dev_euclidean_accuracy": 0.677734375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.255086898803711, + "eval_Qnli-dev_euclidean_ap": 0.7041753433991003, + "eval_Qnli-dev_euclidean_f1": 0.6919275123558485, + "eval_Qnli-dev_euclidean_f1_threshold": 16.40647315979004, + "eval_Qnli-dev_euclidean_precision": 0.5660377358490566, + "eval_Qnli-dev_euclidean_recall": 0.8898305084745762, + "eval_Qnli-dev_manhattan_accuracy": 0.677734375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 245.8935546875, + "eval_Qnli-dev_manhattan_ap": 0.7094875226219203, + "eval_Qnli-dev_manhattan_f1": 0.6853582554517135, + "eval_Qnli-dev_manhattan_f1_threshold": 314.30340576171875, + "eval_Qnli-dev_manhattan_precision": 0.541871921182266, + "eval_Qnli-dev_manhattan_recall": 0.9322033898305084, + "eval_Qnli-dev_max_accuracy": 0.677734375, + "eval_Qnli-dev_max_accuracy_threshold": 399.835205078125, + "eval_Qnli-dev_max_ap": 0.7094875226219203, + "eval_Qnli-dev_max_f1": 0.6919275123558485, + "eval_Qnli-dev_max_f1_threshold": 314.30340576171875, + "eval_Qnli-dev_max_precision": 0.5660377358490566, + "eval_Qnli-dev_max_recall": 0.9533898305084746, + "eval_allNLI-dev_cosine_accuracy": 0.732421875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8643351793289185, + "eval_allNLI-dev_cosine_ap": 0.6115839387857255, + "eval_allNLI-dev_cosine_f1": 0.6201550387596899, + "eval_allNLI-dev_cosine_f1_threshold": 0.8219221830368042, + "eval_allNLI-dev_cosine_precision": 0.5607476635514018, + "eval_allNLI-dev_cosine_recall": 0.6936416184971098, + "eval_allNLI-dev_dot_accuracy": 0.6953125, + "eval_allNLI-dev_dot_accuracy_threshold": 378.34271240234375, + "eval_allNLI-dev_dot_ap": 0.5127166733299272, + "eval_allNLI-dev_dot_f1": 0.5909980430528375, + "eval_allNLI-dev_dot_f1_threshold": 328.07965087890625, + "eval_allNLI-dev_dot_precision": 0.4467455621301775, + "eval_allNLI-dev_dot_recall": 0.8728323699421965, + "eval_allNLI-dev_euclidean_accuracy": 0.732421875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.17978286743164, + "eval_allNLI-dev_euclidean_ap": 0.613114025203357, + "eval_allNLI-dev_euclidean_f1": 0.6256157635467982, + "eval_allNLI-dev_euclidean_f1_threshold": 13.1724853515625, + "eval_allNLI-dev_euclidean_precision": 0.5450643776824035, + "eval_allNLI-dev_euclidean_recall": 0.7341040462427746, + "eval_allNLI-dev_manhattan_accuracy": 0.734375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 210.77999877929688, + "eval_allNLI-dev_manhattan_ap": 0.6100950498761212, + "eval_allNLI-dev_manhattan_f1": 0.6345733041575492, + "eval_allNLI-dev_manhattan_f1_threshold": 260.439453125, + "eval_allNLI-dev_manhattan_precision": 0.5105633802816901, + "eval_allNLI-dev_manhattan_recall": 0.838150289017341, + "eval_allNLI-dev_max_accuracy": 0.734375, + "eval_allNLI-dev_max_accuracy_threshold": 378.34271240234375, + "eval_allNLI-dev_max_ap": 0.613114025203357, + "eval_allNLI-dev_max_f1": 0.6345733041575492, + "eval_allNLI-dev_max_f1_threshold": 328.07965087890625, + "eval_allNLI-dev_max_precision": 0.5607476635514018, + "eval_allNLI-dev_max_recall": 0.8728323699421965, + "eval_sequential_score": 0.7094875226219203, + "eval_sts-test_pearson_cosine": 0.8373205365959482, + "eval_sts-test_pearson_dot": 0.8042043616182744, + "eval_sts-test_pearson_euclidean": 0.8719501312262712, + "eval_sts-test_pearson_manhattan": 0.8690821192761715, + "eval_sts-test_pearson_max": 0.8719501312262712, + "eval_sts-test_spearman_cosine": 0.8765620656619563, + "eval_sts-test_spearman_dot": 0.7961438699365659, + "eval_sts-test_spearman_euclidean": 0.8714019319646984, + "eval_sts-test_spearman_manhattan": 0.8709197371232058, + "eval_sts-test_spearman_max": 0.8765620656619563, + "eval_vitaminc-pairs_loss": 3.7996833324432373, + "eval_vitaminc-pairs_runtime": 5.5217, + "eval_vitaminc-pairs_samples_per_second": 23.181, + "eval_vitaminc-pairs_steps_per_second": 0.362, + "step": 1840 + }, + { + "epoch": 1.4274631497284718, + "eval_negation-triplets_loss": 1.6409591436386108, + "eval_negation-triplets_runtime": 1.1086, + "eval_negation-triplets_samples_per_second": 115.458, + "eval_negation-triplets_steps_per_second": 1.804, + "step": 1840 + }, + { + "epoch": 1.4274631497284718, + "eval_scitail-pairs-pos_loss": 0.13959158957004547, + "eval_scitail-pairs-pos_runtime": 1.4028, + "eval_scitail-pairs-pos_samples_per_second": 91.249, + "eval_scitail-pairs-pos_steps_per_second": 1.426, + "step": 1840 + }, + { + "epoch": 1.4274631497284718, + "eval_scitail-pairs-qa_loss": 0.00729799410328269, + "eval_scitail-pairs-qa_runtime": 0.8122, + "eval_scitail-pairs-qa_samples_per_second": 157.604, + "eval_scitail-pairs-qa_steps_per_second": 2.463, + "step": 1840 + }, + { + "epoch": 1.4274631497284718, + "eval_xsum-pairs_loss": 0.07058010995388031, + "eval_xsum-pairs_runtime": 3.9587, + "eval_xsum-pairs_samples_per_second": 32.334, + "eval_xsum-pairs_steps_per_second": 0.505, + "step": 1840 + }, + { + "epoch": 1.4274631497284718, + "eval_sciq_pairs_loss": 0.046764593571424484, + "eval_sciq_pairs_runtime": 6.283, + "eval_sciq_pairs_samples_per_second": 20.372, + "eval_sciq_pairs_steps_per_second": 0.318, + "step": 1840 + }, + { + "epoch": 1.4274631497284718, + "eval_qasc_pairs_loss": 0.19537001848220825, + "eval_qasc_pairs_runtime": 0.9419, + "eval_qasc_pairs_samples_per_second": 135.892, + "eval_qasc_pairs_steps_per_second": 2.123, + "step": 1840 + }, + { + "epoch": 1.4274631497284718, + "eval_openbookqa_pairs_loss": 1.0662832260131836, + "eval_openbookqa_pairs_runtime": 0.8926, + "eval_openbookqa_pairs_samples_per_second": 143.395, + "eval_openbookqa_pairs_steps_per_second": 2.241, + "step": 1840 + }, + { + "epoch": 1.4274631497284718, + "eval_msmarco_pairs_loss": 0.33644944429397583, + "eval_msmarco_pairs_runtime": 1.9855, + "eval_msmarco_pairs_samples_per_second": 64.467, + "eval_msmarco_pairs_steps_per_second": 1.007, + "step": 1840 + }, + { + "epoch": 1.4274631497284718, + "eval_nq_pairs_loss": 0.5248653292655945, + "eval_nq_pairs_runtime": 3.3635, + "eval_nq_pairs_samples_per_second": 38.055, + "eval_nq_pairs_steps_per_second": 0.595, + "step": 1840 + }, + { + "epoch": 1.4274631497284718, + "eval_trivia_pairs_loss": 0.7039917707443237, + "eval_trivia_pairs_runtime": 4.5165, + "eval_trivia_pairs_samples_per_second": 28.341, + "eval_trivia_pairs_steps_per_second": 0.443, + "step": 1840 + }, + { + "epoch": 1.4274631497284718, + "eval_gooaq_pairs_loss": 0.6478250622749329, + "eval_gooaq_pairs_runtime": 1.4882, + "eval_gooaq_pairs_samples_per_second": 86.009, + "eval_gooaq_pairs_steps_per_second": 1.344, + "step": 1840 + }, + { + "epoch": 1.4274631497284718, + "eval_paws-pos_loss": 0.031161140650510788, + "eval_paws-pos_runtime": 1.0493, + "eval_paws-pos_samples_per_second": 121.991, + "eval_paws-pos_steps_per_second": 1.906, + "step": 1840 + }, + { + "epoch": 1.4274631497284718, + "eval_global_dataset_loss": 0.40719538927078247, + "eval_global_dataset_runtime": 19.4524, + "eval_global_dataset_samples_per_second": 21.386, + "eval_global_dataset_steps_per_second": 0.36, + "step": 1840 + }, + { + "epoch": 1.4282389449185415, + "grad_norm": 4.9047112464904785, + "learning_rate": 1.574130504417175e-05, + "loss": 0.1617, + "step": 1841 + }, + { + "epoch": 1.4290147401086113, + "grad_norm": 10.677031517028809, + "learning_rate": 1.5713379971866783e-05, + "loss": 0.6259, + "step": 1842 + }, + { + "epoch": 1.429790535298681, + "grad_norm": 7.786196708679199, + "learning_rate": 1.5685466602727477e-05, + "loss": 0.4134, + "step": 1843 + }, + { + "epoch": 1.430566330488751, + "grad_norm": 5.014035701751709, + "learning_rate": 1.5657565029123034e-05, + "loss": 0.1234, + "step": 1844 + }, + { + "epoch": 1.4313421256788208, + "grad_norm": 7.045890808105469, + "learning_rate": 1.562967534338363e-05, + "loss": 0.2915, + "step": 1845 + }, + { + "epoch": 1.4321179208688906, + "grad_norm": 6.4519853591918945, + "learning_rate": 1.5601797637800087e-05, + "loss": 0.3538, + "step": 1846 + }, + { + "epoch": 1.4328937160589605, + "grad_norm": 4.449458122253418, + "learning_rate": 1.5573932004623615e-05, + "loss": 0.1225, + "step": 1847 + }, + { + "epoch": 1.4336695112490303, + "grad_norm": 2.6542670726776123, + "learning_rate": 1.5546078536065444e-05, + "loss": 0.0631, + "step": 1848 + }, + { + "epoch": 1.4344453064391, + "grad_norm": 7.313796043395996, + "learning_rate": 1.5518237324296563e-05, + "loss": 0.3229, + "step": 1849 + }, + { + "epoch": 1.4352211016291698, + "grad_norm": 2.7030460834503174, + "learning_rate": 1.5490408461447397e-05, + "loss": 0.1176, + "step": 1850 + }, + { + "epoch": 1.4359968968192396, + "grad_norm": 5.962095737457275, + "learning_rate": 1.5462592039607513e-05, + "loss": 0.1955, + "step": 1851 + }, + { + "epoch": 1.4367726920093096, + "grad_norm": 1.4826499223709106, + "learning_rate": 1.543478815082529e-05, + "loss": 0.0498, + "step": 1852 + }, + { + "epoch": 1.4375484871993793, + "grad_norm": 4.61406946182251, + "learning_rate": 1.5406996887107667e-05, + "loss": 0.1463, + "step": 1853 + }, + { + "epoch": 1.4383242823894493, + "grad_norm": 0.4812264144420624, + "learning_rate": 1.537921834041978e-05, + "loss": 0.0195, + "step": 1854 + }, + { + "epoch": 1.439100077579519, + "grad_norm": 7.116182327270508, + "learning_rate": 1.5351452602684686e-05, + "loss": 0.2555, + "step": 1855 + }, + { + "epoch": 1.4398758727695888, + "grad_norm": 4.053106784820557, + "learning_rate": 1.532369976578306e-05, + "loss": 0.1071, + "step": 1856 + }, + { + "epoch": 1.4406516679596586, + "grad_norm": 7.386887073516846, + "learning_rate": 1.5295959921552878e-05, + "loss": 0.2955, + "step": 1857 + }, + { + "epoch": 1.4414274631497284, + "grad_norm": 8.269121170043945, + "learning_rate": 1.526823316178914e-05, + "loss": 0.5042, + "step": 1858 + }, + { + "epoch": 1.4422032583397983, + "grad_norm": 9.993685722351074, + "learning_rate": 1.5240519578243524e-05, + "loss": 0.5207, + "step": 1859 + }, + { + "epoch": 1.442979053529868, + "grad_norm": 0.3865280747413635, + "learning_rate": 1.5212819262624127e-05, + "loss": 0.0077, + "step": 1860 + }, + { + "epoch": 1.442979053529868, + "eval_Qnli-dev_cosine_accuracy": 0.677734375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8205337524414062, + "eval_Qnli-dev_cosine_ap": 0.7159660840716326, + "eval_Qnli-dev_cosine_f1": 0.691588785046729, + "eval_Qnli-dev_cosine_f1_threshold": 0.6951051354408264, + "eval_Qnli-dev_cosine_precision": 0.5467980295566502, + "eval_Qnli-dev_cosine_recall": 0.940677966101695, + "eval_Qnli-dev_dot_accuracy": 0.677734375, + "eval_Qnli-dev_dot_accuracy_threshold": 385.08001708984375, + "eval_Qnli-dev_dot_ap": 0.6406159546395409, + "eval_Qnli-dev_dot_f1": 0.6948356807511737, + "eval_Qnli-dev_dot_f1_threshold": 320.15106201171875, + "eval_Qnli-dev_dot_precision": 0.5508684863523573, + "eval_Qnli-dev_dot_recall": 0.940677966101695, + "eval_Qnli-dev_euclidean_accuracy": 0.68359375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.938385009765625, + "eval_Qnli-dev_euclidean_ap": 0.7143570324612342, + "eval_Qnli-dev_euclidean_f1": 0.6943999999999998, + "eval_Qnli-dev_euclidean_f1_threshold": 16.545787811279297, + "eval_Qnli-dev_euclidean_precision": 0.5578406169665809, + "eval_Qnli-dev_euclidean_recall": 0.9194915254237288, + "eval_Qnli-dev_manhattan_accuracy": 0.681640625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 241.22779846191406, + "eval_Qnli-dev_manhattan_ap": 0.7180768288231284, + "eval_Qnli-dev_manhattan_f1": 0.6963434022257552, + "eval_Qnli-dev_manhattan_f1_threshold": 303.5975646972656, + "eval_Qnli-dev_manhattan_precision": 0.5572519083969466, + "eval_Qnli-dev_manhattan_recall": 0.9279661016949152, + "eval_Qnli-dev_max_accuracy": 0.68359375, + "eval_Qnli-dev_max_accuracy_threshold": 385.08001708984375, + "eval_Qnli-dev_max_ap": 0.7180768288231284, + "eval_Qnli-dev_max_f1": 0.6963434022257552, + "eval_Qnli-dev_max_f1_threshold": 320.15106201171875, + "eval_Qnli-dev_max_precision": 0.5578406169665809, + "eval_Qnli-dev_max_recall": 0.940677966101695, + "eval_allNLI-dev_cosine_accuracy": 0.734375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8666304349899292, + "eval_allNLI-dev_cosine_ap": 0.6131430269056544, + "eval_allNLI-dev_cosine_f1": 0.6307692307692309, + "eval_allNLI-dev_cosine_f1_threshold": 0.8194683790206909, + "eval_allNLI-dev_cosine_precision": 0.5668202764976958, + "eval_allNLI-dev_cosine_recall": 0.7109826589595376, + "eval_allNLI-dev_dot_accuracy": 0.697265625, + "eval_allNLI-dev_dot_accuracy_threshold": 385.79217529296875, + "eval_allNLI-dev_dot_ap": 0.5162681159321405, + "eval_allNLI-dev_dot_f1": 0.5846774193548387, + "eval_allNLI-dev_dot_f1_threshold": 332.63671875, + "eval_allNLI-dev_dot_precision": 0.44891640866873067, + "eval_allNLI-dev_dot_recall": 0.838150289017341, + "eval_allNLI-dev_euclidean_accuracy": 0.7265625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.305049896240234, + "eval_allNLI-dev_euclidean_ap": 0.6180477592839047, + "eval_allNLI-dev_euclidean_f1": 0.6291079812206573, + "eval_allNLI-dev_euclidean_f1_threshold": 13.459049224853516, + "eval_allNLI-dev_euclidean_precision": 0.5296442687747036, + "eval_allNLI-dev_euclidean_recall": 0.7745664739884393, + "eval_allNLI-dev_manhattan_accuracy": 0.740234375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 214.03382873535156, + "eval_allNLI-dev_manhattan_ap": 0.6169813334005825, + "eval_allNLI-dev_manhattan_f1": 0.640198511166253, + "eval_allNLI-dev_manhattan_f1_threshold": 236.46234130859375, + "eval_allNLI-dev_manhattan_precision": 0.5608695652173913, + "eval_allNLI-dev_manhattan_recall": 0.7456647398843931, + "eval_allNLI-dev_max_accuracy": 0.740234375, + "eval_allNLI-dev_max_accuracy_threshold": 385.79217529296875, + "eval_allNLI-dev_max_ap": 0.6180477592839047, + "eval_allNLI-dev_max_f1": 0.640198511166253, + "eval_allNLI-dev_max_f1_threshold": 332.63671875, + "eval_allNLI-dev_max_precision": 0.5668202764976958, + "eval_allNLI-dev_max_recall": 0.838150289017341, + "eval_sequential_score": 0.7180768288231284, + "eval_sts-test_pearson_cosine": 0.8350268507817415, + "eval_sts-test_pearson_dot": 0.8110712177390301, + "eval_sts-test_pearson_euclidean": 0.8679612247743447, + "eval_sts-test_pearson_manhattan": 0.8657022819992763, + "eval_sts-test_pearson_max": 0.8679612247743447, + "eval_sts-test_spearman_cosine": 0.8739434903529, + "eval_sts-test_spearman_dot": 0.8109926069738005, + "eval_sts-test_spearman_euclidean": 0.867002793467401, + "eval_sts-test_spearman_manhattan": 0.8669980498336934, + "eval_sts-test_spearman_max": 0.8739434903529, + "eval_vitaminc-pairs_loss": 3.7778611183166504, + "eval_vitaminc-pairs_runtime": 5.5343, + "eval_vitaminc-pairs_samples_per_second": 23.128, + "eval_vitaminc-pairs_steps_per_second": 0.361, + "step": 1860 + }, + { + "epoch": 1.442979053529868, + "eval_negation-triplets_loss": 1.6409580707550049, + "eval_negation-triplets_runtime": 1.1182, + "eval_negation-triplets_samples_per_second": 114.472, + "eval_negation-triplets_steps_per_second": 1.789, + "step": 1860 + }, + { + "epoch": 1.442979053529868, + "eval_scitail-pairs-pos_loss": 0.15163648128509521, + "eval_scitail-pairs-pos_runtime": 1.4161, + "eval_scitail-pairs-pos_samples_per_second": 90.388, + "eval_scitail-pairs-pos_steps_per_second": 1.412, + "step": 1860 + }, + { + "epoch": 1.442979053529868, + "eval_scitail-pairs-qa_loss": 0.006420614197850227, + "eval_scitail-pairs-qa_runtime": 0.8067, + "eval_scitail-pairs-qa_samples_per_second": 158.666, + "eval_scitail-pairs-qa_steps_per_second": 2.479, + "step": 1860 + }, + { + "epoch": 1.442979053529868, + "eval_xsum-pairs_loss": 0.08534740656614304, + "eval_xsum-pairs_runtime": 3.9561, + "eval_xsum-pairs_samples_per_second": 32.355, + "eval_xsum-pairs_steps_per_second": 0.506, + "step": 1860 + }, + { + "epoch": 1.442979053529868, + "eval_sciq_pairs_loss": 0.0424392893910408, + "eval_sciq_pairs_runtime": 6.2242, + "eval_sciq_pairs_samples_per_second": 20.565, + "eval_sciq_pairs_steps_per_second": 0.321, + "step": 1860 + }, + { + "epoch": 1.442979053529868, + "eval_qasc_pairs_loss": 0.2004096806049347, + "eval_qasc_pairs_runtime": 0.9439, + "eval_qasc_pairs_samples_per_second": 135.608, + "eval_qasc_pairs_steps_per_second": 2.119, + "step": 1860 + }, + { + "epoch": 1.442979053529868, + "eval_openbookqa_pairs_loss": 1.0335988998413086, + "eval_openbookqa_pairs_runtime": 0.8986, + "eval_openbookqa_pairs_samples_per_second": 142.446, + "eval_openbookqa_pairs_steps_per_second": 2.226, + "step": 1860 + }, + { + "epoch": 1.442979053529868, + "eval_msmarco_pairs_loss": 0.3472045958042145, + "eval_msmarco_pairs_runtime": 1.99, + "eval_msmarco_pairs_samples_per_second": 64.322, + "eval_msmarco_pairs_steps_per_second": 1.005, + "step": 1860 + }, + { + "epoch": 1.442979053529868, + "eval_nq_pairs_loss": 0.5380181074142456, + "eval_nq_pairs_runtime": 3.3673, + "eval_nq_pairs_samples_per_second": 38.013, + "eval_nq_pairs_steps_per_second": 0.594, + "step": 1860 + }, + { + "epoch": 1.442979053529868, + "eval_trivia_pairs_loss": 0.7067691087722778, + "eval_trivia_pairs_runtime": 4.5409, + "eval_trivia_pairs_samples_per_second": 28.188, + "eval_trivia_pairs_steps_per_second": 0.44, + "step": 1860 + }, + { + "epoch": 1.442979053529868, + "eval_gooaq_pairs_loss": 0.6585148572921753, + "eval_gooaq_pairs_runtime": 1.4791, + "eval_gooaq_pairs_samples_per_second": 86.537, + "eval_gooaq_pairs_steps_per_second": 1.352, + "step": 1860 + }, + { + "epoch": 1.442979053529868, + "eval_paws-pos_loss": 0.03135106712579727, + "eval_paws-pos_runtime": 1.053, + "eval_paws-pos_samples_per_second": 121.553, + "eval_paws-pos_steps_per_second": 1.899, + "step": 1860 + }, + { + "epoch": 1.442979053529868, + "eval_global_dataset_loss": 0.407550573348999, + "eval_global_dataset_runtime": 19.5341, + "eval_global_dataset_samples_per_second": 21.296, + "eval_global_dataset_steps_per_second": 0.358, + "step": 1860 + }, + { + "epoch": 1.4437548487199379, + "grad_norm": 0.39913633465766907, + "learning_rate": 1.5185132306595124e-05, + "loss": 0.008, + "step": 1861 + }, + { + "epoch": 1.4445306439100078, + "grad_norm": 8.190966606140137, + "learning_rate": 1.5157458801776494e-05, + "loss": 0.3834, + "step": 1862 + }, + { + "epoch": 1.4453064391000776, + "grad_norm": 10.092744827270508, + "learning_rate": 1.5129798839743687e-05, + "loss": 0.5169, + "step": 1863 + }, + { + "epoch": 1.4460822342901474, + "grad_norm": 6.415419101715088, + "learning_rate": 1.5102152512027362e-05, + "loss": 0.4644, + "step": 1864 + }, + { + "epoch": 1.4468580294802171, + "grad_norm": 4.3949198722839355, + "learning_rate": 1.5074519910113047e-05, + "loss": 0.1231, + "step": 1865 + }, + { + "epoch": 1.4476338246702871, + "grad_norm": 8.93814468383789, + "learning_rate": 1.5046901125440846e-05, + "loss": 0.3934, + "step": 1866 + }, + { + "epoch": 1.4484096198603569, + "grad_norm": 1.5640870332717896, + "learning_rate": 1.5019296249405146e-05, + "loss": 0.0248, + "step": 1867 + }, + { + "epoch": 1.4491854150504266, + "grad_norm": 9.639640808105469, + "learning_rate": 1.4991705373354298e-05, + "loss": 0.4832, + "step": 1868 + }, + { + "epoch": 1.4499612102404966, + "grad_norm": 5.728815078735352, + "learning_rate": 1.4964128588590345e-05, + "loss": 0.1717, + "step": 1869 + }, + { + "epoch": 1.4507370054305664, + "grad_norm": 6.835094928741455, + "learning_rate": 1.4936565986368688e-05, + "loss": 0.2287, + "step": 1870 + }, + { + "epoch": 1.4515128006206361, + "grad_norm": 10.740663528442383, + "learning_rate": 1.4909017657897789e-05, + "loss": 0.9166, + "step": 1871 + }, + { + "epoch": 1.452288595810706, + "grad_norm": 0.451395183801651, + "learning_rate": 1.4881483694338889e-05, + "loss": 0.0074, + "step": 1872 + }, + { + "epoch": 1.4530643910007757, + "grad_norm": 5.0044708251953125, + "learning_rate": 1.4853964186805687e-05, + "loss": 0.2021, + "step": 1873 + }, + { + "epoch": 1.4538401861908457, + "grad_norm": 3.4364075660705566, + "learning_rate": 1.4826459226364036e-05, + "loss": 0.0618, + "step": 1874 + }, + { + "epoch": 1.4546159813809154, + "grad_norm": 9.861668586730957, + "learning_rate": 1.479896890403168e-05, + "loss": 0.6004, + "step": 1875 + }, + { + "epoch": 1.4553917765709854, + "grad_norm": 5.263766288757324, + "learning_rate": 1.4771493310777894e-05, + "loss": 0.1177, + "step": 1876 + }, + { + "epoch": 1.4561675717610552, + "grad_norm": 7.743444442749023, + "learning_rate": 1.4744032537523221e-05, + "loss": 0.355, + "step": 1877 + }, + { + "epoch": 1.456943366951125, + "grad_norm": 9.26172161102295, + "learning_rate": 1.471658667513917e-05, + "loss": 0.4237, + "step": 1878 + }, + { + "epoch": 1.4577191621411947, + "grad_norm": 15.931780815124512, + "learning_rate": 1.468915581444789e-05, + "loss": 2.1223, + "step": 1879 + }, + { + "epoch": 1.4584949573312644, + "grad_norm": 8.42387866973877, + "learning_rate": 1.4661740046221914e-05, + "loss": 0.4203, + "step": 1880 + }, + { + "epoch": 1.4584949573312644, + "eval_Qnli-dev_cosine_accuracy": 0.677734375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8284502029418945, + "eval_Qnli-dev_cosine_ap": 0.7174859376133815, + "eval_Qnli-dev_cosine_f1": 0.6943999999999998, + "eval_Qnli-dev_cosine_f1_threshold": 0.7081899046897888, + "eval_Qnli-dev_cosine_precision": 0.5578406169665809, + "eval_Qnli-dev_cosine_recall": 0.9194915254237288, + "eval_Qnli-dev_dot_accuracy": 0.673828125, + "eval_Qnli-dev_dot_accuracy_threshold": 392.69921875, + "eval_Qnli-dev_dot_ap": 0.6428399242367437, + "eval_Qnli-dev_dot_f1": 0.6947040498442367, + "eval_Qnli-dev_dot_f1_threshold": 321.2248229980469, + "eval_Qnli-dev_dot_precision": 0.5492610837438424, + "eval_Qnli-dev_dot_recall": 0.9449152542372882, + "eval_Qnli-dev_euclidean_accuracy": 0.6796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.526029586791992, + "eval_Qnli-dev_euclidean_ap": 0.7182124253900506, + "eval_Qnli-dev_euclidean_f1": 0.7015873015873015, + "eval_Qnli-dev_euclidean_f1_threshold": 16.896026611328125, + "eval_Qnli-dev_euclidean_precision": 0.5609137055837563, + "eval_Qnli-dev_euclidean_recall": 0.9364406779661016, + "eval_Qnli-dev_manhattan_accuracy": 0.6796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 240.1810302734375, + "eval_Qnli-dev_manhattan_ap": 0.7181450716323455, + "eval_Qnli-dev_manhattan_f1": 0.6985172981878089, + "eval_Qnli-dev_manhattan_f1_threshold": 296.4964294433594, + "eval_Qnli-dev_manhattan_precision": 0.5714285714285714, + "eval_Qnli-dev_manhattan_recall": 0.8983050847457628, + "eval_Qnli-dev_max_accuracy": 0.6796875, + "eval_Qnli-dev_max_accuracy_threshold": 392.69921875, + "eval_Qnli-dev_max_ap": 0.7182124253900506, + "eval_Qnli-dev_max_f1": 0.7015873015873015, + "eval_Qnli-dev_max_f1_threshold": 321.2248229980469, + "eval_Qnli-dev_max_precision": 0.5714285714285714, + "eval_Qnli-dev_max_recall": 0.9449152542372882, + "eval_allNLI-dev_cosine_accuracy": 0.736328125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8664657473564148, + "eval_allNLI-dev_cosine_ap": 0.6128089042164357, + "eval_allNLI-dev_cosine_f1": 0.6331658291457286, + "eval_allNLI-dev_cosine_f1_threshold": 0.8111977577209473, + "eval_allNLI-dev_cosine_precision": 0.56, + "eval_allNLI-dev_cosine_recall": 0.7283236994219653, + "eval_allNLI-dev_dot_accuracy": 0.69921875, + "eval_allNLI-dev_dot_accuracy_threshold": 384.8896484375, + "eval_allNLI-dev_dot_ap": 0.5265866211033887, + "eval_allNLI-dev_dot_f1": 0.5858585858585859, + "eval_allNLI-dev_dot_f1_threshold": 331.3429870605469, + "eval_allNLI-dev_dot_precision": 0.4503105590062112, + "eval_allNLI-dev_dot_recall": 0.838150289017341, + "eval_allNLI-dev_euclidean_accuracy": 0.73046875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.321836471557617, + "eval_allNLI-dev_euclidean_ap": 0.6148476277677892, + "eval_allNLI-dev_euclidean_f1": 0.6352357320099256, + "eval_allNLI-dev_euclidean_f1_threshold": 13.124080657958984, + "eval_allNLI-dev_euclidean_precision": 0.5565217391304348, + "eval_allNLI-dev_euclidean_recall": 0.7398843930635838, + "eval_allNLI-dev_manhattan_accuracy": 0.7421875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 220.26966857910156, + "eval_allNLI-dev_manhattan_ap": 0.6137038710372589, + "eval_allNLI-dev_manhattan_f1": 0.6381909547738693, + "eval_allNLI-dev_manhattan_f1_threshold": 236.4145965576172, + "eval_allNLI-dev_manhattan_precision": 0.5644444444444444, + "eval_allNLI-dev_manhattan_recall": 0.7341040462427746, + "eval_allNLI-dev_max_accuracy": 0.7421875, + "eval_allNLI-dev_max_accuracy_threshold": 384.8896484375, + "eval_allNLI-dev_max_ap": 0.6148476277677892, + "eval_allNLI-dev_max_f1": 0.6381909547738693, + "eval_allNLI-dev_max_f1_threshold": 331.3429870605469, + "eval_allNLI-dev_max_precision": 0.5644444444444444, + "eval_allNLI-dev_max_recall": 0.838150289017341, + "eval_sequential_score": 0.7182124253900506, + "eval_sts-test_pearson_cosine": 0.8307127036358364, + "eval_sts-test_pearson_dot": 0.8041550311307148, + "eval_sts-test_pearson_euclidean": 0.8647120695117854, + "eval_sts-test_pearson_manhattan": 0.8622695896800229, + "eval_sts-test_pearson_max": 0.8647120695117854, + "eval_sts-test_spearman_cosine": 0.8700076169135266, + "eval_sts-test_spearman_dot": 0.7999611525332522, + "eval_sts-test_spearman_euclidean": 0.8642739062990147, + "eval_sts-test_spearman_manhattan": 0.864384039153299, + "eval_sts-test_spearman_max": 0.8700076169135266, + "eval_vitaminc-pairs_loss": 3.8564577102661133, + "eval_vitaminc-pairs_runtime": 5.5264, + "eval_vitaminc-pairs_samples_per_second": 23.162, + "eval_vitaminc-pairs_steps_per_second": 0.362, + "step": 1880 + }, + { + "epoch": 1.4584949573312644, + "eval_negation-triplets_loss": 1.6949903964996338, + "eval_negation-triplets_runtime": 1.0881, + "eval_negation-triplets_samples_per_second": 117.639, + "eval_negation-triplets_steps_per_second": 1.838, + "step": 1880 + }, + { + "epoch": 1.4584949573312644, + "eval_scitail-pairs-pos_loss": 0.1661553978919983, + "eval_scitail-pairs-pos_runtime": 1.3284, + "eval_scitail-pairs-pos_samples_per_second": 96.354, + "eval_scitail-pairs-pos_steps_per_second": 1.506, + "step": 1880 + }, + { + "epoch": 1.4584949573312644, + "eval_scitail-pairs-qa_loss": 0.00592053821310401, + "eval_scitail-pairs-qa_runtime": 0.7882, + "eval_scitail-pairs-qa_samples_per_second": 162.392, + "eval_scitail-pairs-qa_steps_per_second": 2.537, + "step": 1880 + }, + { + "epoch": 1.4584949573312644, + "eval_xsum-pairs_loss": 0.0907338336110115, + "eval_xsum-pairs_runtime": 3.9509, + "eval_xsum-pairs_samples_per_second": 32.397, + "eval_xsum-pairs_steps_per_second": 0.506, + "step": 1880 + }, + { + "epoch": 1.4584949573312644, + "eval_sciq_pairs_loss": 0.04132814705371857, + "eval_sciq_pairs_runtime": 6.0747, + "eval_sciq_pairs_samples_per_second": 21.071, + "eval_sciq_pairs_steps_per_second": 0.329, + "step": 1880 + }, + { + "epoch": 1.4584949573312644, + "eval_qasc_pairs_loss": 0.21149024367332458, + "eval_qasc_pairs_runtime": 0.9296, + "eval_qasc_pairs_samples_per_second": 137.695, + "eval_qasc_pairs_steps_per_second": 2.151, + "step": 1880 + }, + { + "epoch": 1.4584949573312644, + "eval_openbookqa_pairs_loss": 1.0699771642684937, + "eval_openbookqa_pairs_runtime": 0.8799, + "eval_openbookqa_pairs_samples_per_second": 145.474, + "eval_openbookqa_pairs_steps_per_second": 2.273, + "step": 1880 + }, + { + "epoch": 1.4584949573312644, + "eval_msmarco_pairs_loss": 0.33686983585357666, + "eval_msmarco_pairs_runtime": 1.9917, + "eval_msmarco_pairs_samples_per_second": 64.265, + "eval_msmarco_pairs_steps_per_second": 1.004, + "step": 1880 + }, + { + "epoch": 1.4584949573312644, + "eval_nq_pairs_loss": 0.5288337469100952, + "eval_nq_pairs_runtime": 3.3631, + "eval_nq_pairs_samples_per_second": 38.06, + "eval_nq_pairs_steps_per_second": 0.595, + "step": 1880 + }, + { + "epoch": 1.4584949573312644, + "eval_trivia_pairs_loss": 0.7156597375869751, + "eval_trivia_pairs_runtime": 4.497, + "eval_trivia_pairs_samples_per_second": 28.463, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 1880 + }, + { + "epoch": 1.4584949573312644, + "eval_gooaq_pairs_loss": 0.6307020783424377, + "eval_gooaq_pairs_runtime": 1.4837, + "eval_gooaq_pairs_samples_per_second": 86.27, + "eval_gooaq_pairs_steps_per_second": 1.348, + "step": 1880 + }, + { + "epoch": 1.4584949573312644, + "eval_paws-pos_loss": 0.029905738309025764, + "eval_paws-pos_runtime": 1.0731, + "eval_paws-pos_samples_per_second": 119.278, + "eval_paws-pos_steps_per_second": 1.864, + "step": 1880 + }, + { + "epoch": 1.4584949573312644, + "eval_global_dataset_loss": 0.4159528911113739, + "eval_global_dataset_runtime": 19.489, + "eval_global_dataset_samples_per_second": 21.345, + "eval_global_dataset_steps_per_second": 0.359, + "step": 1880 + }, + { + "epoch": 1.4592707525213344, + "grad_norm": 0.4212203323841095, + "learning_rate": 1.4634339461183813e-05, + "loss": 0.0075, + "step": 1881 + }, + { + "epoch": 1.4600465477114042, + "grad_norm": 12.250702857971191, + "learning_rate": 1.4606954150005917e-05, + "loss": 0.5759, + "step": 1882 + }, + { + "epoch": 1.460822342901474, + "grad_norm": 5.15714168548584, + "learning_rate": 1.4579584203310018e-05, + "loss": 0.1991, + "step": 1883 + }, + { + "epoch": 1.461598138091544, + "grad_norm": 5.92712926864624, + "learning_rate": 1.4552229711667054e-05, + "loss": 0.2055, + "step": 1884 + }, + { + "epoch": 1.4623739332816137, + "grad_norm": 6.788751125335693, + "learning_rate": 1.4524890765596833e-05, + "loss": 0.4316, + "step": 1885 + }, + { + "epoch": 1.4631497284716835, + "grad_norm": 4.003585338592529, + "learning_rate": 1.4497567455567715e-05, + "loss": 0.0629, + "step": 1886 + }, + { + "epoch": 1.4639255236617532, + "grad_norm": 9.057820320129395, + "learning_rate": 1.4470259871996323e-05, + "loss": 0.5783, + "step": 1887 + }, + { + "epoch": 1.4647013188518232, + "grad_norm": 15.321743965148926, + "learning_rate": 1.444296810524723e-05, + "loss": 2.2605, + "step": 1888 + }, + { + "epoch": 1.465477114041893, + "grad_norm": 3.588271379470825, + "learning_rate": 1.4415692245632671e-05, + "loss": 0.074, + "step": 1889 + }, + { + "epoch": 1.4662529092319627, + "grad_norm": 7.726933479309082, + "learning_rate": 1.438843238341224e-05, + "loss": 0.414, + "step": 1890 + }, + { + "epoch": 1.4670287044220327, + "grad_norm": 5.311885356903076, + "learning_rate": 1.4361188608792616e-05, + "loss": 0.1246, + "step": 1891 + }, + { + "epoch": 1.4678044996121025, + "grad_norm": 6.5706787109375, + "learning_rate": 1.4333961011927207e-05, + "loss": 0.238, + "step": 1892 + }, + { + "epoch": 1.4685802948021722, + "grad_norm": 5.357235908508301, + "learning_rate": 1.430674968291591e-05, + "loss": 0.1744, + "step": 1893 + }, + { + "epoch": 1.469356089992242, + "grad_norm": 4.78502893447876, + "learning_rate": 1.4279554711804781e-05, + "loss": 0.1964, + "step": 1894 + }, + { + "epoch": 1.4701318851823117, + "grad_norm": 6.954161643981934, + "learning_rate": 1.4252376188585752e-05, + "loss": 0.2378, + "step": 1895 + }, + { + "epoch": 1.4709076803723817, + "grad_norm": 10.114001274108887, + "learning_rate": 1.4225214203196305e-05, + "loss": 0.6163, + "step": 1896 + }, + { + "epoch": 1.4716834755624515, + "grad_norm": 0.46074315905570984, + "learning_rate": 1.4198068845519233e-05, + "loss": 0.0134, + "step": 1897 + }, + { + "epoch": 1.4724592707525213, + "grad_norm": 3.6576077938079834, + "learning_rate": 1.4170940205382279e-05, + "loss": 0.0808, + "step": 1898 + }, + { + "epoch": 1.4732350659425912, + "grad_norm": 1.297433853149414, + "learning_rate": 1.414382837255787e-05, + "loss": 0.0299, + "step": 1899 + }, + { + "epoch": 1.474010861132661, + "grad_norm": 9.051212310791016, + "learning_rate": 1.4116733436762819e-05, + "loss": 0.397, + "step": 1900 + }, + { + "epoch": 1.474010861132661, + "eval_Qnli-dev_cosine_accuracy": 0.67578125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8056411743164062, + "eval_Qnli-dev_cosine_ap": 0.7122794349428349, + "eval_Qnli-dev_cosine_f1": 0.6990291262135921, + "eval_Qnli-dev_cosine_f1_threshold": 0.7107926607131958, + "eval_Qnli-dev_cosine_precision": 0.5654450261780105, + "eval_Qnli-dev_cosine_recall": 0.9152542372881356, + "eval_Qnli-dev_dot_accuracy": 0.669921875, + "eval_Qnli-dev_dot_accuracy_threshold": 384.1185302734375, + "eval_Qnli-dev_dot_ap": 0.6394898301384377, + "eval_Qnli-dev_dot_f1": 0.6840390879478827, + "eval_Qnli-dev_dot_f1_threshold": 330.908447265625, + "eval_Qnli-dev_dot_precision": 0.5555555555555556, + "eval_Qnli-dev_dot_recall": 0.8898305084745762, + "eval_Qnli-dev_euclidean_accuracy": 0.681640625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.451963424682617, + "eval_Qnli-dev_euclidean_ap": 0.7135810337674882, + "eval_Qnli-dev_euclidean_f1": 0.6948051948051949, + "eval_Qnli-dev_euclidean_f1_threshold": 16.583803176879883, + "eval_Qnli-dev_euclidean_precision": 0.5631578947368421, + "eval_Qnli-dev_euclidean_recall": 0.9067796610169492, + "eval_Qnli-dev_manhattan_accuracy": 0.681640625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 242.968994140625, + "eval_Qnli-dev_manhattan_ap": 0.7143571351883156, + "eval_Qnli-dev_manhattan_f1": 0.6929392446633826, + "eval_Qnli-dev_manhattan_f1_threshold": 299.19439697265625, + "eval_Qnli-dev_manhattan_precision": 0.5656836461126006, + "eval_Qnli-dev_manhattan_recall": 0.8940677966101694, + "eval_Qnli-dev_max_accuracy": 0.681640625, + "eval_Qnli-dev_max_accuracy_threshold": 384.1185302734375, + "eval_Qnli-dev_max_ap": 0.7143571351883156, + "eval_Qnli-dev_max_f1": 0.6990291262135921, + "eval_Qnli-dev_max_f1_threshold": 330.908447265625, + "eval_Qnli-dev_max_precision": 0.5656836461126006, + "eval_Qnli-dev_max_recall": 0.9152542372881356, + "eval_allNLI-dev_cosine_accuracy": 0.734375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8581185340881348, + "eval_allNLI-dev_cosine_ap": 0.6153250958118793, + "eval_allNLI-dev_cosine_f1": 0.6318289786223278, + "eval_allNLI-dev_cosine_f1_threshold": 0.8003972768783569, + "eval_allNLI-dev_cosine_precision": 0.5362903225806451, + "eval_allNLI-dev_cosine_recall": 0.7687861271676301, + "eval_allNLI-dev_dot_accuracy": 0.703125, + "eval_allNLI-dev_dot_accuracy_threshold": 394.1622009277344, + "eval_allNLI-dev_dot_ap": 0.5363284145847672, + "eval_allNLI-dev_dot_f1": 0.591346153846154, + "eval_allNLI-dev_dot_f1_threshold": 360.6121826171875, + "eval_allNLI-dev_dot_precision": 0.5061728395061729, + "eval_allNLI-dev_dot_recall": 0.7109826589595376, + "eval_allNLI-dev_euclidean_accuracy": 0.736328125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.446758270263672, + "eval_allNLI-dev_euclidean_ap": 0.6164205817610829, + "eval_allNLI-dev_euclidean_f1": 0.6332574031890661, + "eval_allNLI-dev_euclidean_f1_threshold": 13.972461700439453, + "eval_allNLI-dev_euclidean_precision": 0.5225563909774437, + "eval_allNLI-dev_euclidean_recall": 0.8034682080924855, + "eval_allNLI-dev_manhattan_accuracy": 0.744140625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 215.04000854492188, + "eval_allNLI-dev_manhattan_ap": 0.6157420290701557, + "eval_allNLI-dev_manhattan_f1": 0.6356968215158924, + "eval_allNLI-dev_manhattan_f1_threshold": 242.4803466796875, + "eval_allNLI-dev_manhattan_precision": 0.5508474576271186, + "eval_allNLI-dev_manhattan_recall": 0.7514450867052023, + "eval_allNLI-dev_max_accuracy": 0.744140625, + "eval_allNLI-dev_max_accuracy_threshold": 394.1622009277344, + "eval_allNLI-dev_max_ap": 0.6164205817610829, + "eval_allNLI-dev_max_f1": 0.6356968215158924, + "eval_allNLI-dev_max_f1_threshold": 360.6121826171875, + "eval_allNLI-dev_max_precision": 0.5508474576271186, + "eval_allNLI-dev_max_recall": 0.8034682080924855, + "eval_sequential_score": 0.7143571351883156, + "eval_sts-test_pearson_cosine": 0.8289814573948676, + "eval_sts-test_pearson_dot": 0.7983434367333816, + "eval_sts-test_pearson_euclidean": 0.8653429538092183, + "eval_sts-test_pearson_manhattan": 0.8623611724573763, + "eval_sts-test_pearson_max": 0.8653429538092183, + "eval_sts-test_spearman_cosine": 0.8704549952736563, + "eval_sts-test_spearman_dot": 0.7907918875817227, + "eval_sts-test_spearman_euclidean": 0.8656306750418792, + "eval_sts-test_spearman_manhattan": 0.8650079612396178, + "eval_sts-test_spearman_max": 0.8704549952736563, + "eval_vitaminc-pairs_loss": 3.9490621089935303, + "eval_vitaminc-pairs_runtime": 5.5295, + "eval_vitaminc-pairs_samples_per_second": 23.148, + "eval_vitaminc-pairs_steps_per_second": 0.362, + "step": 1900 + }, + { + "epoch": 1.474010861132661, + "eval_negation-triplets_loss": 1.732640027999878, + "eval_negation-triplets_runtime": 1.093, + "eval_negation-triplets_samples_per_second": 117.11, + "eval_negation-triplets_steps_per_second": 1.83, + "step": 1900 + }, + { + "epoch": 1.474010861132661, + "eval_scitail-pairs-pos_loss": 0.15987463295459747, + "eval_scitail-pairs-pos_runtime": 1.3568, + "eval_scitail-pairs-pos_samples_per_second": 94.341, + "eval_scitail-pairs-pos_steps_per_second": 1.474, + "step": 1900 + }, + { + "epoch": 1.474010861132661, + "eval_scitail-pairs-qa_loss": 0.004988286644220352, + "eval_scitail-pairs-qa_runtime": 0.7871, + "eval_scitail-pairs-qa_samples_per_second": 162.62, + "eval_scitail-pairs-qa_steps_per_second": 2.541, + "step": 1900 + }, + { + "epoch": 1.474010861132661, + "eval_xsum-pairs_loss": 0.09815925359725952, + "eval_xsum-pairs_runtime": 3.9669, + "eval_xsum-pairs_samples_per_second": 32.267, + "eval_xsum-pairs_steps_per_second": 0.504, + "step": 1900 + }, + { + "epoch": 1.474010861132661, + "eval_sciq_pairs_loss": 0.04286904260516167, + "eval_sciq_pairs_runtime": 6.2072, + "eval_sciq_pairs_samples_per_second": 20.621, + "eval_sciq_pairs_steps_per_second": 0.322, + "step": 1900 + }, + { + "epoch": 1.474010861132661, + "eval_qasc_pairs_loss": 0.19964127242565155, + "eval_qasc_pairs_runtime": 0.9429, + "eval_qasc_pairs_samples_per_second": 135.745, + "eval_qasc_pairs_steps_per_second": 2.121, + "step": 1900 + }, + { + "epoch": 1.474010861132661, + "eval_openbookqa_pairs_loss": 1.0724620819091797, + "eval_openbookqa_pairs_runtime": 0.8934, + "eval_openbookqa_pairs_samples_per_second": 143.272, + "eval_openbookqa_pairs_steps_per_second": 2.239, + "step": 1900 + }, + { + "epoch": 1.474010861132661, + "eval_msmarco_pairs_loss": 0.29800882935523987, + "eval_msmarco_pairs_runtime": 1.9886, + "eval_msmarco_pairs_samples_per_second": 64.367, + "eval_msmarco_pairs_steps_per_second": 1.006, + "step": 1900 + }, + { + "epoch": 1.474010861132661, + "eval_nq_pairs_loss": 0.5535090565681458, + "eval_nq_pairs_runtime": 3.3653, + "eval_nq_pairs_samples_per_second": 38.035, + "eval_nq_pairs_steps_per_second": 0.594, + "step": 1900 + }, + { + "epoch": 1.474010861132661, + "eval_trivia_pairs_loss": 0.7323253154754639, + "eval_trivia_pairs_runtime": 4.5027, + "eval_trivia_pairs_samples_per_second": 28.427, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 1900 + }, + { + "epoch": 1.474010861132661, + "eval_gooaq_pairs_loss": 0.6194378733634949, + "eval_gooaq_pairs_runtime": 1.4841, + "eval_gooaq_pairs_samples_per_second": 86.248, + "eval_gooaq_pairs_steps_per_second": 1.348, + "step": 1900 + }, + { + "epoch": 1.474010861132661, + "eval_paws-pos_loss": 0.03011227585375309, + "eval_paws-pos_runtime": 1.0465, + "eval_paws-pos_samples_per_second": 122.309, + "eval_paws-pos_steps_per_second": 1.911, + "step": 1900 + }, + { + "epoch": 1.474010861132661, + "eval_global_dataset_loss": 0.4248928725719452, + "eval_global_dataset_runtime": 19.4957, + "eval_global_dataset_samples_per_second": 21.338, + "eval_global_dataset_steps_per_second": 0.359, + "step": 1900 + }, + { + "epoch": 1.4747866563227308, + "grad_norm": 6.924256324768066, + "learning_rate": 1.4089655487658013e-05, + "loss": 0.2322, + "step": 1901 + }, + { + "epoch": 1.4755624515128005, + "grad_norm": 3.9136409759521484, + "learning_rate": 1.4062594614848152e-05, + "loss": 0.1483, + "step": 1902 + }, + { + "epoch": 1.4763382467028705, + "grad_norm": 5.443375110626221, + "learning_rate": 1.4035550907881405e-05, + "loss": 0.3458, + "step": 1903 + }, + { + "epoch": 1.4771140418929403, + "grad_norm": 4.32873010635376, + "learning_rate": 1.400852445624915e-05, + "loss": 0.1381, + "step": 1904 + }, + { + "epoch": 1.47788983708301, + "grad_norm": 3.122194290161133, + "learning_rate": 1.3981515349385654e-05, + "loss": 0.0579, + "step": 1905 + }, + { + "epoch": 1.47866563227308, + "grad_norm": 11.154013633728027, + "learning_rate": 1.3954523676667796e-05, + "loss": 0.6073, + "step": 1906 + }, + { + "epoch": 1.4794414274631498, + "grad_norm": 4.1729230880737305, + "learning_rate": 1.3927549527414754e-05, + "loss": 0.0892, + "step": 1907 + }, + { + "epoch": 1.4802172226532195, + "grad_norm": 10.311990737915039, + "learning_rate": 1.390059299088774e-05, + "loss": 0.7287, + "step": 1908 + }, + { + "epoch": 1.4809930178432893, + "grad_norm": 9.161555290222168, + "learning_rate": 1.3873654156289656e-05, + "loss": 0.4626, + "step": 1909 + }, + { + "epoch": 1.481768813033359, + "grad_norm": 0.5573193430900574, + "learning_rate": 1.3846733112764847e-05, + "loss": 0.0151, + "step": 1910 + }, + { + "epoch": 1.482544608223429, + "grad_norm": 4.960310459136963, + "learning_rate": 1.3819829949398768e-05, + "loss": 0.1189, + "step": 1911 + }, + { + "epoch": 1.4833204034134988, + "grad_norm": 6.604163646697998, + "learning_rate": 1.379294475521771e-05, + "loss": 0.2048, + "step": 1912 + }, + { + "epoch": 1.4840961986035688, + "grad_norm": 9.144646644592285, + "learning_rate": 1.3766077619188519e-05, + "loss": 0.2738, + "step": 1913 + }, + { + "epoch": 1.4848719937936385, + "grad_norm": 0.8234514594078064, + "learning_rate": 1.3739228630218261e-05, + "loss": 0.0334, + "step": 1914 + }, + { + "epoch": 1.4856477889837083, + "grad_norm": 5.421793460845947, + "learning_rate": 1.3712397877153965e-05, + "loss": 0.1126, + "step": 1915 + }, + { + "epoch": 1.486423584173778, + "grad_norm": 3.8895211219787598, + "learning_rate": 1.3685585448782307e-05, + "loss": 0.105, + "step": 1916 + }, + { + "epoch": 1.4871993793638478, + "grad_norm": 7.020480155944824, + "learning_rate": 1.3658791433829322e-05, + "loss": 0.3402, + "step": 1917 + }, + { + "epoch": 1.4879751745539178, + "grad_norm": 6.382047176361084, + "learning_rate": 1.363201592096013e-05, + "loss": 0.2484, + "step": 1918 + }, + { + "epoch": 1.4887509697439876, + "grad_norm": 1.5577524900436401, + "learning_rate": 1.3605258998778608e-05, + "loss": 0.0218, + "step": 1919 + }, + { + "epoch": 1.4895267649340573, + "grad_norm": 9.67573070526123, + "learning_rate": 1.3578520755827115e-05, + "loss": 0.4909, + "step": 1920 + }, + { + "epoch": 1.4895267649340573, + "eval_Qnli-dev_cosine_accuracy": 0.67578125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8574289083480835, + "eval_Qnli-dev_cosine_ap": 0.7131896216473332, + "eval_Qnli-dev_cosine_f1": 0.6988352745424292, + "eval_Qnli-dev_cosine_f1_threshold": 0.719751238822937, + "eval_Qnli-dev_cosine_precision": 0.5753424657534246, + "eval_Qnli-dev_cosine_recall": 0.8898305084745762, + "eval_Qnli-dev_dot_accuracy": 0.65625, + "eval_Qnli-dev_dot_accuracy_threshold": 367.4393310546875, + "eval_Qnli-dev_dot_ap": 0.6330652166932025, + "eval_Qnli-dev_dot_f1": 0.6910828025477707, + "eval_Qnli-dev_dot_f1_threshold": 322.78759765625, + "eval_Qnli-dev_dot_precision": 0.5535714285714286, + "eval_Qnli-dev_dot_recall": 0.9194915254237288, + "eval_Qnli-dev_euclidean_accuracy": 0.681640625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.72474479675293, + "eval_Qnli-dev_euclidean_ap": 0.7135685295218737, + "eval_Qnli-dev_euclidean_f1": 0.696629213483146, + "eval_Qnli-dev_euclidean_f1_threshold": 17.18631362915039, + "eval_Qnli-dev_euclidean_precision": 0.5607235142118863, + "eval_Qnli-dev_euclidean_recall": 0.9194915254237288, + "eval_Qnli-dev_manhattan_accuracy": 0.68359375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 254.84141540527344, + "eval_Qnli-dev_manhattan_ap": 0.7145169100818405, + "eval_Qnli-dev_manhattan_f1": 0.6953748006379585, + "eval_Qnli-dev_manhattan_f1_threshold": 315.40875244140625, + "eval_Qnli-dev_manhattan_precision": 0.5575447570332481, + "eval_Qnli-dev_manhattan_recall": 0.923728813559322, + "eval_Qnli-dev_max_accuracy": 0.68359375, + "eval_Qnli-dev_max_accuracy_threshold": 367.4393310546875, + "eval_Qnli-dev_max_ap": 0.7145169100818405, + "eval_Qnli-dev_max_f1": 0.6988352745424292, + "eval_Qnli-dev_max_f1_threshold": 322.78759765625, + "eval_Qnli-dev_max_precision": 0.5753424657534246, + "eval_Qnli-dev_max_recall": 0.923728813559322, + "eval_allNLI-dev_cosine_accuracy": 0.728515625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8445956707000732, + "eval_allNLI-dev_cosine_ap": 0.6160421942459223, + "eval_allNLI-dev_cosine_f1": 0.6342592592592593, + "eval_allNLI-dev_cosine_f1_threshold": 0.7812492251396179, + "eval_allNLI-dev_cosine_precision": 0.528957528957529, + "eval_allNLI-dev_cosine_recall": 0.791907514450867, + "eval_allNLI-dev_dot_accuracy": 0.697265625, + "eval_allNLI-dev_dot_accuracy_threshold": 380.1988525390625, + "eval_allNLI-dev_dot_ap": 0.5459196544641434, + "eval_allNLI-dev_dot_f1": 0.5916230366492147, + "eval_allNLI-dev_dot_f1_threshold": 356.35302734375, + "eval_allNLI-dev_dot_precision": 0.5406698564593302, + "eval_allNLI-dev_dot_recall": 0.653179190751445, + "eval_allNLI-dev_euclidean_accuracy": 0.732421875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.97309684753418, + "eval_allNLI-dev_euclidean_ap": 0.6167379218032171, + "eval_allNLI-dev_euclidean_f1": 0.6339285714285714, + "eval_allNLI-dev_euclidean_f1_threshold": 14.225412368774414, + "eval_allNLI-dev_euclidean_precision": 0.5163636363636364, + "eval_allNLI-dev_euclidean_recall": 0.8208092485549133, + "eval_allNLI-dev_manhattan_accuracy": 0.740234375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 219.45164489746094, + "eval_allNLI-dev_manhattan_ap": 0.6174222421545458, + "eval_allNLI-dev_manhattan_f1": 0.6313465783664459, + "eval_allNLI-dev_manhattan_f1_threshold": 263.63568115234375, + "eval_allNLI-dev_manhattan_precision": 0.5107142857142857, + "eval_allNLI-dev_manhattan_recall": 0.8265895953757225, + "eval_allNLI-dev_max_accuracy": 0.740234375, + "eval_allNLI-dev_max_accuracy_threshold": 380.1988525390625, + "eval_allNLI-dev_max_ap": 0.6174222421545458, + "eval_allNLI-dev_max_f1": 0.6342592592592593, + "eval_allNLI-dev_max_f1_threshold": 356.35302734375, + "eval_allNLI-dev_max_precision": 0.5406698564593302, + "eval_allNLI-dev_max_recall": 0.8265895953757225, + "eval_sequential_score": 0.7145169100818405, + "eval_sts-test_pearson_cosine": 0.8313783519178186, + "eval_sts-test_pearson_dot": 0.8037162645514555, + "eval_sts-test_pearson_euclidean": 0.866726281765954, + "eval_sts-test_pearson_manhattan": 0.863767777726559, + "eval_sts-test_pearson_max": 0.866726281765954, + "eval_sts-test_spearman_cosine": 0.8718183424519453, + "eval_sts-test_spearman_dot": 0.7996847687428988, + "eval_sts-test_spearman_euclidean": 0.8675947184018251, + "eval_sts-test_spearman_manhattan": 0.8660864218878893, + "eval_sts-test_spearman_max": 0.8718183424519453, + "eval_vitaminc-pairs_loss": 4.0265326499938965, + "eval_vitaminc-pairs_runtime": 5.5463, + "eval_vitaminc-pairs_samples_per_second": 23.079, + "eval_vitaminc-pairs_steps_per_second": 0.361, + "step": 1920 + }, + { + "epoch": 1.4895267649340573, + "eval_negation-triplets_loss": 1.7300397157669067, + "eval_negation-triplets_runtime": 1.0849, + "eval_negation-triplets_samples_per_second": 117.978, + "eval_negation-triplets_steps_per_second": 1.843, + "step": 1920 + }, + { + "epoch": 1.4895267649340573, + "eval_scitail-pairs-pos_loss": 0.1415715217590332, + "eval_scitail-pairs-pos_runtime": 1.3536, + "eval_scitail-pairs-pos_samples_per_second": 94.561, + "eval_scitail-pairs-pos_steps_per_second": 1.478, + "step": 1920 + }, + { + "epoch": 1.4895267649340573, + "eval_scitail-pairs-qa_loss": 0.004903589375317097, + "eval_scitail-pairs-qa_runtime": 0.7896, + "eval_scitail-pairs-qa_samples_per_second": 162.11, + "eval_scitail-pairs-qa_steps_per_second": 2.533, + "step": 1920 + }, + { + "epoch": 1.4895267649340573, + "eval_xsum-pairs_loss": 0.07376985996961594, + "eval_xsum-pairs_runtime": 3.9579, + "eval_xsum-pairs_samples_per_second": 32.341, + "eval_xsum-pairs_steps_per_second": 0.505, + "step": 1920 + }, + { + "epoch": 1.4895267649340573, + "eval_sciq_pairs_loss": 0.044173698872327805, + "eval_sciq_pairs_runtime": 6.2303, + "eval_sciq_pairs_samples_per_second": 20.545, + "eval_sciq_pairs_steps_per_second": 0.321, + "step": 1920 + }, + { + "epoch": 1.4895267649340573, + "eval_qasc_pairs_loss": 0.20457953214645386, + "eval_qasc_pairs_runtime": 0.9394, + "eval_qasc_pairs_samples_per_second": 136.253, + "eval_qasc_pairs_steps_per_second": 2.129, + "step": 1920 + }, + { + "epoch": 1.4895267649340573, + "eval_openbookqa_pairs_loss": 1.1103839874267578, + "eval_openbookqa_pairs_runtime": 0.8987, + "eval_openbookqa_pairs_samples_per_second": 142.42, + "eval_openbookqa_pairs_steps_per_second": 2.225, + "step": 1920 + }, + { + "epoch": 1.4895267649340573, + "eval_msmarco_pairs_loss": 0.32623329758644104, + "eval_msmarco_pairs_runtime": 1.9887, + "eval_msmarco_pairs_samples_per_second": 64.364, + "eval_msmarco_pairs_steps_per_second": 1.006, + "step": 1920 + }, + { + "epoch": 1.4895267649340573, + "eval_nq_pairs_loss": 0.5245996713638306, + "eval_nq_pairs_runtime": 3.3593, + "eval_nq_pairs_samples_per_second": 38.104, + "eval_nq_pairs_steps_per_second": 0.595, + "step": 1920 + }, + { + "epoch": 1.4895267649340573, + "eval_trivia_pairs_loss": 0.7103387713432312, + "eval_trivia_pairs_runtime": 4.4991, + "eval_trivia_pairs_samples_per_second": 28.45, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 1920 + }, + { + "epoch": 1.4895267649340573, + "eval_gooaq_pairs_loss": 0.5741066932678223, + "eval_gooaq_pairs_runtime": 1.4799, + "eval_gooaq_pairs_samples_per_second": 86.493, + "eval_gooaq_pairs_steps_per_second": 1.351, + "step": 1920 + }, + { + "epoch": 1.4895267649340573, + "eval_paws-pos_loss": 0.030585337430238724, + "eval_paws-pos_runtime": 1.0466, + "eval_paws-pos_samples_per_second": 122.301, + "eval_paws-pos_steps_per_second": 1.911, + "step": 1920 + }, + { + "epoch": 1.4895267649340573, + "eval_global_dataset_loss": 0.42454588413238525, + "eval_global_dataset_runtime": 19.4921, + "eval_global_dataset_samples_per_second": 21.342, + "eval_global_dataset_steps_per_second": 0.359, + "step": 1920 + }, + { + "epoch": 1.4903025601241273, + "grad_norm": 6.249855995178223, + "learning_rate": 1.3551801280586207e-05, + "loss": 0.15, + "step": 1921 + }, + { + "epoch": 1.491078355314197, + "grad_norm": 5.963282585144043, + "learning_rate": 1.3525100661474324e-05, + "loss": 0.2728, + "step": 1922 + }, + { + "epoch": 1.4918541505042668, + "grad_norm": 3.7456300258636475, + "learning_rate": 1.349841898684751e-05, + "loss": 0.0908, + "step": 1923 + }, + { + "epoch": 1.4926299456943366, + "grad_norm": 10.316059112548828, + "learning_rate": 1.3471756344999135e-05, + "loss": 0.5755, + "step": 1924 + }, + { + "epoch": 1.4934057408844066, + "grad_norm": 1.0556391477584839, + "learning_rate": 1.3445112824159568e-05, + "loss": 0.0171, + "step": 1925 + }, + { + "epoch": 1.4941815360744763, + "grad_norm": 0.48754870891571045, + "learning_rate": 1.3418488512495912e-05, + "loss": 0.0056, + "step": 1926 + }, + { + "epoch": 1.494957331264546, + "grad_norm": 0.06053588166832924, + "learning_rate": 1.3391883498111702e-05, + "loss": 0.0008, + "step": 1927 + }, + { + "epoch": 1.495733126454616, + "grad_norm": 3.4073486328125, + "learning_rate": 1.3365297869046613e-05, + "loss": 0.0591, + "step": 1928 + }, + { + "epoch": 1.4965089216446859, + "grad_norm": 10.400298118591309, + "learning_rate": 1.3338731713276184e-05, + "loss": 0.4441, + "step": 1929 + }, + { + "epoch": 1.4972847168347556, + "grad_norm": 7.917856216430664, + "learning_rate": 1.3312185118711502e-05, + "loss": 0.3384, + "step": 1930 + }, + { + "epoch": 1.4980605120248254, + "grad_norm": 6.580019474029541, + "learning_rate": 1.3285658173198923e-05, + "loss": 0.163, + "step": 1931 + }, + { + "epoch": 1.4988363072148951, + "grad_norm": 5.718687534332275, + "learning_rate": 1.3259150964519786e-05, + "loss": 0.0849, + "step": 1932 + }, + { + "epoch": 1.4996121024049651, + "grad_norm": 7.495270729064941, + "learning_rate": 1.3232663580390118e-05, + "loss": 0.3104, + "step": 1933 + }, + { + "epoch": 1.5003878975950349, + "grad_norm": 8.769048690795898, + "learning_rate": 1.3206196108460336e-05, + "loss": 0.4226, + "step": 1934 + }, + { + "epoch": 1.5011636927851049, + "grad_norm": 3.4028501510620117, + "learning_rate": 1.3179748636314985e-05, + "loss": 0.0918, + "step": 1935 + }, + { + "epoch": 1.5019394879751746, + "grad_norm": 3.947514772415161, + "learning_rate": 1.3153321251472406e-05, + "loss": 0.0959, + "step": 1936 + }, + { + "epoch": 1.5027152831652444, + "grad_norm": 3.8078134059906006, + "learning_rate": 1.3126914041384478e-05, + "loss": 0.0845, + "step": 1937 + }, + { + "epoch": 1.5034910783553141, + "grad_norm": 7.807572841644287, + "learning_rate": 1.3100527093436321e-05, + "loss": 0.4794, + "step": 1938 + }, + { + "epoch": 1.504266873545384, + "grad_norm": 4.6256794929504395, + "learning_rate": 1.307416049494599e-05, + "loss": 0.1041, + "step": 1939 + }, + { + "epoch": 1.505042668735454, + "grad_norm": 5.546387672424316, + "learning_rate": 1.3047814333164229e-05, + "loss": 0.1277, + "step": 1940 + }, + { + "epoch": 1.505042668735454, + "eval_Qnli-dev_cosine_accuracy": 0.689453125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8301265239715576, + "eval_Qnli-dev_cosine_ap": 0.7210188301409275, + "eval_Qnli-dev_cosine_f1": 0.6919104991394148, + "eval_Qnli-dev_cosine_f1_threshold": 0.7363235354423523, + "eval_Qnli-dev_cosine_precision": 0.5826086956521739, + "eval_Qnli-dev_cosine_recall": 0.8516949152542372, + "eval_Qnli-dev_dot_accuracy": 0.6640625, + "eval_Qnli-dev_dot_accuracy_threshold": 389.80908203125, + "eval_Qnli-dev_dot_ap": 0.6208014718691316, + "eval_Qnli-dev_dot_f1": 0.6882716049382716, + "eval_Qnli-dev_dot_f1_threshold": 316.023193359375, + "eval_Qnli-dev_dot_precision": 0.5412621359223301, + "eval_Qnli-dev_dot_recall": 0.9449152542372882, + "eval_Qnli-dev_euclidean_accuracy": 0.689453125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.254837036132812, + "eval_Qnli-dev_euclidean_ap": 0.7240178306674767, + "eval_Qnli-dev_euclidean_f1": 0.69281045751634, + "eval_Qnli-dev_euclidean_f1_threshold": 16.88025665283203, + "eval_Qnli-dev_euclidean_precision": 0.5638297872340425, + "eval_Qnli-dev_euclidean_recall": 0.8983050847457628, + "eval_Qnli-dev_manhattan_accuracy": 0.6875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 244.45188903808594, + "eval_Qnli-dev_manhattan_ap": 0.7246181180194788, + "eval_Qnli-dev_manhattan_f1": 0.6906710310965629, + "eval_Qnli-dev_manhattan_f1_threshold": 306.42779541015625, + "eval_Qnli-dev_manhattan_precision": 0.5626666666666666, + "eval_Qnli-dev_manhattan_recall": 0.8940677966101694, + "eval_Qnli-dev_max_accuracy": 0.689453125, + "eval_Qnli-dev_max_accuracy_threshold": 389.80908203125, + "eval_Qnli-dev_max_ap": 0.7246181180194788, + "eval_Qnli-dev_max_f1": 0.69281045751634, + "eval_Qnli-dev_max_f1_threshold": 316.023193359375, + "eval_Qnli-dev_max_precision": 0.5826086956521739, + "eval_Qnli-dev_max_recall": 0.9449152542372882, + "eval_allNLI-dev_cosine_accuracy": 0.7265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8389524817466736, + "eval_allNLI-dev_cosine_ap": 0.6120507114486965, + "eval_allNLI-dev_cosine_f1": 0.6270022883295194, + "eval_allNLI-dev_cosine_f1_threshold": 0.7797061204910278, + "eval_allNLI-dev_cosine_precision": 0.5189393939393939, + "eval_allNLI-dev_cosine_recall": 0.791907514450867, + "eval_allNLI-dev_dot_accuracy": 0.69921875, + "eval_allNLI-dev_dot_accuracy_threshold": 370.90350341796875, + "eval_allNLI-dev_dot_ap": 0.5314295889750881, + "eval_allNLI-dev_dot_f1": 0.5869074492099323, + "eval_allNLI-dev_dot_f1_threshold": 340.1354064941406, + "eval_allNLI-dev_dot_precision": 0.48148148148148145, + "eval_allNLI-dev_dot_recall": 0.7514450867052023, + "eval_allNLI-dev_euclidean_accuracy": 0.724609375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.3927001953125, + "eval_allNLI-dev_euclidean_ap": 0.612058379152184, + "eval_allNLI-dev_euclidean_f1": 0.6380090497737557, + "eval_allNLI-dev_euclidean_f1_threshold": 14.17930793762207, + "eval_allNLI-dev_euclidean_precision": 0.5241635687732342, + "eval_allNLI-dev_euclidean_recall": 0.815028901734104, + "eval_allNLI-dev_manhattan_accuracy": 0.73828125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 224.98875427246094, + "eval_allNLI-dev_manhattan_ap": 0.6118313476181912, + "eval_allNLI-dev_manhattan_f1": 0.6296296296296297, + "eval_allNLI-dev_manhattan_f1_threshold": 253.98919677734375, + "eval_allNLI-dev_manhattan_precision": 0.525096525096525, + "eval_allNLI-dev_manhattan_recall": 0.7861271676300579, + "eval_allNLI-dev_max_accuracy": 0.73828125, + "eval_allNLI-dev_max_accuracy_threshold": 370.90350341796875, + "eval_allNLI-dev_max_ap": 0.612058379152184, + "eval_allNLI-dev_max_f1": 0.6380090497737557, + "eval_allNLI-dev_max_f1_threshold": 340.1354064941406, + "eval_allNLI-dev_max_precision": 0.525096525096525, + "eval_allNLI-dev_max_recall": 0.815028901734104, + "eval_sequential_score": 0.7246181180194788, + "eval_sts-test_pearson_cosine": 0.8281555839666211, + "eval_sts-test_pearson_dot": 0.7953653224400192, + "eval_sts-test_pearson_euclidean": 0.8632082365973088, + "eval_sts-test_pearson_manhattan": 0.8599130489005837, + "eval_sts-test_pearson_max": 0.8632082365973088, + "eval_sts-test_spearman_cosine": 0.8674738899960682, + "eval_sts-test_spearman_dot": 0.7858068655702893, + "eval_sts-test_spearman_euclidean": 0.863390337394103, + "eval_sts-test_spearman_manhattan": 0.8616222150565287, + "eval_sts-test_spearman_max": 0.8674738899960682, + "eval_vitaminc-pairs_loss": 3.9212806224823, + "eval_vitaminc-pairs_runtime": 5.5228, + "eval_vitaminc-pairs_samples_per_second": 23.177, + "eval_vitaminc-pairs_steps_per_second": 0.362, + "step": 1940 + }, + { + "epoch": 1.505042668735454, + "eval_negation-triplets_loss": 1.7112139463424683, + "eval_negation-triplets_runtime": 1.0906, + "eval_negation-triplets_samples_per_second": 117.364, + "eval_negation-triplets_steps_per_second": 1.834, + "step": 1940 + }, + { + "epoch": 1.505042668735454, + "eval_scitail-pairs-pos_loss": 0.12974028289318085, + "eval_scitail-pairs-pos_runtime": 1.3485, + "eval_scitail-pairs-pos_samples_per_second": 94.918, + "eval_scitail-pairs-pos_steps_per_second": 1.483, + "step": 1940 + }, + { + "epoch": 1.505042668735454, + "eval_scitail-pairs-qa_loss": 0.004312725737690926, + "eval_scitail-pairs-qa_runtime": 0.7901, + "eval_scitail-pairs-qa_samples_per_second": 162.014, + "eval_scitail-pairs-qa_steps_per_second": 2.531, + "step": 1940 + }, + { + "epoch": 1.505042668735454, + "eval_xsum-pairs_loss": 0.0696437731385231, + "eval_xsum-pairs_runtime": 3.9628, + "eval_xsum-pairs_samples_per_second": 32.3, + "eval_xsum-pairs_steps_per_second": 0.505, + "step": 1940 + }, + { + "epoch": 1.505042668735454, + "eval_sciq_pairs_loss": 0.043703727424144745, + "eval_sciq_pairs_runtime": 6.2183, + "eval_sciq_pairs_samples_per_second": 20.584, + "eval_sciq_pairs_steps_per_second": 0.322, + "step": 1940 + }, + { + "epoch": 1.505042668735454, + "eval_qasc_pairs_loss": 0.19343048334121704, + "eval_qasc_pairs_runtime": 0.9435, + "eval_qasc_pairs_samples_per_second": 135.67, + "eval_qasc_pairs_steps_per_second": 2.12, + "step": 1940 + }, + { + "epoch": 1.505042668735454, + "eval_openbookqa_pairs_loss": 1.1336475610733032, + "eval_openbookqa_pairs_runtime": 0.894, + "eval_openbookqa_pairs_samples_per_second": 143.176, + "eval_openbookqa_pairs_steps_per_second": 2.237, + "step": 1940 + }, + { + "epoch": 1.505042668735454, + "eval_msmarco_pairs_loss": 0.34916937351226807, + "eval_msmarco_pairs_runtime": 1.9866, + "eval_msmarco_pairs_samples_per_second": 64.43, + "eval_msmarco_pairs_steps_per_second": 1.007, + "step": 1940 + }, + { + "epoch": 1.505042668735454, + "eval_nq_pairs_loss": 0.5556931495666504, + "eval_nq_pairs_runtime": 3.3562, + "eval_nq_pairs_samples_per_second": 38.139, + "eval_nq_pairs_steps_per_second": 0.596, + "step": 1940 + }, + { + "epoch": 1.505042668735454, + "eval_trivia_pairs_loss": 0.7403109073638916, + "eval_trivia_pairs_runtime": 4.4944, + "eval_trivia_pairs_samples_per_second": 28.48, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 1940 + }, + { + "epoch": 1.505042668735454, + "eval_gooaq_pairs_loss": 0.586372971534729, + "eval_gooaq_pairs_runtime": 1.4818, + "eval_gooaq_pairs_samples_per_second": 86.381, + "eval_gooaq_pairs_steps_per_second": 1.35, + "step": 1940 + }, + { + "epoch": 1.505042668735454, + "eval_paws-pos_loss": 0.03114868700504303, + "eval_paws-pos_runtime": 1.0488, + "eval_paws-pos_samples_per_second": 122.045, + "eval_paws-pos_steps_per_second": 1.907, + "step": 1940 + }, + { + "epoch": 1.505042668735454, + "eval_global_dataset_loss": 0.41872119903564453, + "eval_global_dataset_runtime": 19.4805, + "eval_global_dataset_samples_per_second": 21.355, + "eval_global_dataset_steps_per_second": 0.359, + "step": 1940 + }, + { + "epoch": 1.5058184639255237, + "grad_norm": 5.260847091674805, + "learning_rate": 1.3021488695274128e-05, + "loss": 0.137, + "step": 1941 + }, + { + "epoch": 1.5065942591155936, + "grad_norm": 5.195507526397705, + "learning_rate": 1.2995183668390872e-05, + "loss": 0.1588, + "step": 1942 + }, + { + "epoch": 1.5073700543056634, + "grad_norm": 6.960474967956543, + "learning_rate": 1.296889933956144e-05, + "loss": 0.3433, + "step": 1943 + }, + { + "epoch": 1.5081458494957332, + "grad_norm": 11.506572723388672, + "learning_rate": 1.2942635795764318e-05, + "loss": 0.6962, + "step": 1944 + }, + { + "epoch": 1.508921644685803, + "grad_norm": 5.9135236740112305, + "learning_rate": 1.2916393123909199e-05, + "loss": 0.207, + "step": 1945 + }, + { + "epoch": 1.5096974398758727, + "grad_norm": 11.03255558013916, + "learning_rate": 1.289017141083674e-05, + "loss": 0.6334, + "step": 1946 + }, + { + "epoch": 1.5104732350659424, + "grad_norm": 1.9664965867996216, + "learning_rate": 1.2863970743318215e-05, + "loss": 0.0221, + "step": 1947 + }, + { + "epoch": 1.5112490302560124, + "grad_norm": 2.6385130882263184, + "learning_rate": 1.283779120805526e-05, + "loss": 0.0574, + "step": 1948 + }, + { + "epoch": 1.5120248254460822, + "grad_norm": 7.735723495483398, + "learning_rate": 1.2811632891679585e-05, + "loss": 0.3144, + "step": 1949 + }, + { + "epoch": 1.5128006206361522, + "grad_norm": 1.096258282661438, + "learning_rate": 1.2785495880752678e-05, + "loss": 0.0218, + "step": 1950 + }, + { + "epoch": 1.513576415826222, + "grad_norm": 7.595947265625, + "learning_rate": 1.2759380261765541e-05, + "loss": 0.4694, + "step": 1951 + }, + { + "epoch": 1.5143522110162917, + "grad_norm": 4.479292392730713, + "learning_rate": 1.2733286121138367e-05, + "loss": 0.1281, + "step": 1952 + }, + { + "epoch": 1.5151280062063615, + "grad_norm": 6.908661365509033, + "learning_rate": 1.2707213545220288e-05, + "loss": 0.289, + "step": 1953 + }, + { + "epoch": 1.5159038013964312, + "grad_norm": 3.6389217376708984, + "learning_rate": 1.2681162620289067e-05, + "loss": 0.0704, + "step": 1954 + }, + { + "epoch": 1.5166795965865012, + "grad_norm": 3.911529779434204, + "learning_rate": 1.2655133432550819e-05, + "loss": 0.1425, + "step": 1955 + }, + { + "epoch": 1.517455391776571, + "grad_norm": 9.166714668273926, + "learning_rate": 1.2629126068139738e-05, + "loss": 0.3958, + "step": 1956 + }, + { + "epoch": 1.518231186966641, + "grad_norm": 5.50640344619751, + "learning_rate": 1.26031406131178e-05, + "loss": 0.2563, + "step": 1957 + }, + { + "epoch": 1.5190069821567107, + "grad_norm": 9.986611366271973, + "learning_rate": 1.2577177153474466e-05, + "loss": 0.4828, + "step": 1958 + }, + { + "epoch": 1.5197827773467805, + "grad_norm": 5.142590522766113, + "learning_rate": 1.2551235775126427e-05, + "loss": 0.1402, + "step": 1959 + }, + { + "epoch": 1.5205585725368502, + "grad_norm": 7.962689399719238, + "learning_rate": 1.2525316563917305e-05, + "loss": 0.4353, + "step": 1960 + }, + { + "epoch": 1.5205585725368502, + "eval_Qnli-dev_cosine_accuracy": 0.68359375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8357386589050293, + "eval_Qnli-dev_cosine_ap": 0.7216188137135886, + "eval_Qnli-dev_cosine_f1": 0.6900958466453675, + "eval_Qnli-dev_cosine_f1_threshold": 0.7127985954284668, + "eval_Qnli-dev_cosine_precision": 0.5538461538461539, + "eval_Qnli-dev_cosine_recall": 0.9152542372881356, + "eval_Qnli-dev_dot_accuracy": 0.65625, + "eval_Qnli-dev_dot_accuracy_threshold": 396.4915771484375, + "eval_Qnli-dev_dot_ap": 0.6168942480812701, + "eval_Qnli-dev_dot_f1": 0.683464566929134, + "eval_Qnli-dev_dot_f1_threshold": 321.5917053222656, + "eval_Qnli-dev_dot_precision": 0.543859649122807, + "eval_Qnli-dev_dot_recall": 0.9194915254237288, + "eval_Qnli-dev_euclidean_accuracy": 0.6796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 12.920280456542969, + "eval_Qnli-dev_euclidean_ap": 0.7248892724866826, + "eval_Qnli-dev_euclidean_f1": 0.6886326194398682, + "eval_Qnli-dev_euclidean_f1_threshold": 16.286312103271484, + "eval_Qnli-dev_euclidean_precision": 0.5633423180592992, + "eval_Qnli-dev_euclidean_recall": 0.885593220338983, + "eval_Qnli-dev_manhattan_accuracy": 0.69140625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 245.96142578125, + "eval_Qnli-dev_manhattan_ap": 0.7271141029393466, + "eval_Qnli-dev_manhattan_f1": 0.6854460093896713, + "eval_Qnli-dev_manhattan_f1_threshold": 310.1822509765625, + "eval_Qnli-dev_manhattan_precision": 0.543424317617866, + "eval_Qnli-dev_manhattan_recall": 0.9279661016949152, + "eval_Qnli-dev_max_accuracy": 0.69140625, + "eval_Qnli-dev_max_accuracy_threshold": 396.4915771484375, + "eval_Qnli-dev_max_ap": 0.7271141029393466, + "eval_Qnli-dev_max_f1": 0.6900958466453675, + "eval_Qnli-dev_max_f1_threshold": 321.5917053222656, + "eval_Qnli-dev_max_precision": 0.5633423180592992, + "eval_Qnli-dev_max_recall": 0.9279661016949152, + "eval_allNLI-dev_cosine_accuracy": 0.724609375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8540923595428467, + "eval_allNLI-dev_cosine_ap": 0.6104946588996083, + "eval_allNLI-dev_cosine_f1": 0.6240409207161125, + "eval_allNLI-dev_cosine_f1_threshold": 0.818300724029541, + "eval_allNLI-dev_cosine_precision": 0.5596330275229358, + "eval_allNLI-dev_cosine_recall": 0.7052023121387283, + "eval_allNLI-dev_dot_accuracy": 0.703125, + "eval_allNLI-dev_dot_accuracy_threshold": 377.623046875, + "eval_allNLI-dev_dot_ap": 0.5296308189097265, + "eval_allNLI-dev_dot_f1": 0.5939393939393939, + "eval_allNLI-dev_dot_f1_threshold": 334.8927001953125, + "eval_allNLI-dev_dot_precision": 0.45652173913043476, + "eval_allNLI-dev_dot_recall": 0.8497109826589595, + "eval_allNLI-dev_euclidean_accuracy": 0.73046875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.957559585571289, + "eval_allNLI-dev_euclidean_ap": 0.6092605260457771, + "eval_allNLI-dev_euclidean_f1": 0.6311111111111111, + "eval_allNLI-dev_euclidean_f1_threshold": 14.108078002929688, + "eval_allNLI-dev_euclidean_precision": 0.5126353790613718, + "eval_allNLI-dev_euclidean_recall": 0.8208092485549133, + "eval_allNLI-dev_manhattan_accuracy": 0.7265625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 211.57723999023438, + "eval_allNLI-dev_manhattan_ap": 0.6102406096703648, + "eval_allNLI-dev_manhattan_f1": 0.6292134831460674, + "eval_allNLI-dev_manhattan_f1_threshold": 255.39822387695312, + "eval_allNLI-dev_manhattan_precision": 0.5147058823529411, + "eval_allNLI-dev_manhattan_recall": 0.8092485549132948, + "eval_allNLI-dev_max_accuracy": 0.73046875, + "eval_allNLI-dev_max_accuracy_threshold": 377.623046875, + "eval_allNLI-dev_max_ap": 0.6104946588996083, + "eval_allNLI-dev_max_f1": 0.6311111111111111, + "eval_allNLI-dev_max_f1_threshold": 334.8927001953125, + "eval_allNLI-dev_max_precision": 0.5596330275229358, + "eval_allNLI-dev_max_recall": 0.8497109826589595, + "eval_sequential_score": 0.7271141029393466, + "eval_sts-test_pearson_cosine": 0.8331082077930215, + "eval_sts-test_pearson_dot": 0.79389364938934, + "eval_sts-test_pearson_euclidean": 0.867643329776058, + "eval_sts-test_pearson_manhattan": 0.8655193527995241, + "eval_sts-test_pearson_max": 0.867643329776058, + "eval_sts-test_spearman_cosine": 0.871108050431141, + "eval_sts-test_spearman_dot": 0.7875285360985391, + "eval_sts-test_spearman_euclidean": 0.8662098458668065, + "eval_sts-test_spearman_manhattan": 0.8661069627168677, + "eval_sts-test_spearman_max": 0.871108050431141, + "eval_vitaminc-pairs_loss": 3.9110500812530518, + "eval_vitaminc-pairs_runtime": 5.4949, + "eval_vitaminc-pairs_samples_per_second": 23.294, + "eval_vitaminc-pairs_steps_per_second": 0.364, + "step": 1960 + }, + { + "epoch": 1.5205585725368502, + "eval_negation-triplets_loss": 1.6790313720703125, + "eval_negation-triplets_runtime": 1.0865, + "eval_negation-triplets_samples_per_second": 117.807, + "eval_negation-triplets_steps_per_second": 1.841, + "step": 1960 + }, + { + "epoch": 1.5205585725368502, + "eval_scitail-pairs-pos_loss": 0.14731931686401367, + "eval_scitail-pairs-pos_runtime": 1.3066, + "eval_scitail-pairs-pos_samples_per_second": 97.961, + "eval_scitail-pairs-pos_steps_per_second": 1.531, + "step": 1960 + }, + { + "epoch": 1.5205585725368502, + "eval_scitail-pairs-qa_loss": 0.003832975635305047, + "eval_scitail-pairs-qa_runtime": 0.7767, + "eval_scitail-pairs-qa_samples_per_second": 164.807, + "eval_scitail-pairs-qa_steps_per_second": 2.575, + "step": 1960 + }, + { + "epoch": 1.5205585725368502, + "eval_xsum-pairs_loss": 0.061746131628751755, + "eval_xsum-pairs_runtime": 3.9203, + "eval_xsum-pairs_samples_per_second": 32.65, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 1960 + }, + { + "epoch": 1.5205585725368502, + "eval_sciq_pairs_loss": 0.043878186494112015, + "eval_sciq_pairs_runtime": 6.1594, + "eval_sciq_pairs_samples_per_second": 20.781, + "eval_sciq_pairs_steps_per_second": 0.325, + "step": 1960 + }, + { + "epoch": 1.5205585725368502, + "eval_qasc_pairs_loss": 0.1730397790670395, + "eval_qasc_pairs_runtime": 0.9345, + "eval_qasc_pairs_samples_per_second": 136.966, + "eval_qasc_pairs_steps_per_second": 2.14, + "step": 1960 + }, + { + "epoch": 1.5205585725368502, + "eval_openbookqa_pairs_loss": 1.1393791437149048, + "eval_openbookqa_pairs_runtime": 0.886, + "eval_openbookqa_pairs_samples_per_second": 144.461, + "eval_openbookqa_pairs_steps_per_second": 2.257, + "step": 1960 + }, + { + "epoch": 1.5205585725368502, + "eval_msmarco_pairs_loss": 0.333445280790329, + "eval_msmarco_pairs_runtime": 1.9857, + "eval_msmarco_pairs_samples_per_second": 64.462, + "eval_msmarco_pairs_steps_per_second": 1.007, + "step": 1960 + }, + { + "epoch": 1.5205585725368502, + "eval_nq_pairs_loss": 0.5256453156471252, + "eval_nq_pairs_runtime": 3.3447, + "eval_nq_pairs_samples_per_second": 38.27, + "eval_nq_pairs_steps_per_second": 0.598, + "step": 1960 + }, + { + "epoch": 1.5205585725368502, + "eval_trivia_pairs_loss": 0.6820062398910522, + "eval_trivia_pairs_runtime": 4.4937, + "eval_trivia_pairs_samples_per_second": 28.484, + "eval_trivia_pairs_steps_per_second": 0.445, + "step": 1960 + }, + { + "epoch": 1.5205585725368502, + "eval_gooaq_pairs_loss": 0.5639373064041138, + "eval_gooaq_pairs_runtime": 1.4857, + "eval_gooaq_pairs_samples_per_second": 86.156, + "eval_gooaq_pairs_steps_per_second": 1.346, + "step": 1960 + }, + { + "epoch": 1.5205585725368502, + "eval_paws-pos_loss": 0.031012365594506264, + "eval_paws-pos_runtime": 1.0517, + "eval_paws-pos_samples_per_second": 121.705, + "eval_paws-pos_steps_per_second": 1.902, + "step": 1960 + }, + { + "epoch": 1.5205585725368502, + "eval_global_dataset_loss": 0.4052011966705322, + "eval_global_dataset_runtime": 19.4448, + "eval_global_dataset_samples_per_second": 21.394, + "eval_global_dataset_steps_per_second": 0.36, + "step": 1960 + }, + { + "epoch": 1.52133436772692, + "grad_norm": 6.409043312072754, + "learning_rate": 1.2499419605617342e-05, + "loss": 0.2934, + "step": 1961 + }, + { + "epoch": 1.5221101629169898, + "grad_norm": 8.753270149230957, + "learning_rate": 1.2473544985923178e-05, + "loss": 0.6305, + "step": 1962 + }, + { + "epoch": 1.5228859581070597, + "grad_norm": 10.297379493713379, + "learning_rate": 1.244769279045751e-05, + "loss": 0.5462, + "step": 1963 + }, + { + "epoch": 1.5236617532971297, + "grad_norm": 7.841190338134766, + "learning_rate": 1.2421863104768829e-05, + "loss": 0.3565, + "step": 1964 + }, + { + "epoch": 1.5244375484871995, + "grad_norm": 5.738776683807373, + "learning_rate": 1.2396056014331146e-05, + "loss": 0.1647, + "step": 1965 + }, + { + "epoch": 1.5252133436772692, + "grad_norm": 5.1275529861450195, + "learning_rate": 1.2370271604543708e-05, + "loss": 0.3594, + "step": 1966 + }, + { + "epoch": 1.525989138867339, + "grad_norm": 6.638382911682129, + "learning_rate": 1.2344509960730686e-05, + "loss": 0.3324, + "step": 1967 + }, + { + "epoch": 1.5267649340574088, + "grad_norm": 8.398876190185547, + "learning_rate": 1.2318771168140931e-05, + "loss": 0.5156, + "step": 1968 + }, + { + "epoch": 1.5275407292474785, + "grad_norm": 10.445442199707031, + "learning_rate": 1.2293055311947687e-05, + "loss": 0.545, + "step": 1969 + }, + { + "epoch": 1.5283165244375485, + "grad_norm": 10.456537246704102, + "learning_rate": 1.2267362477248268e-05, + "loss": 0.5059, + "step": 1970 + }, + { + "epoch": 1.5290923196276183, + "grad_norm": 6.196310520172119, + "learning_rate": 1.2241692749063829e-05, + "loss": 0.257, + "step": 1971 + }, + { + "epoch": 1.5298681148176883, + "grad_norm": 0.7681391835212708, + "learning_rate": 1.2216046212339058e-05, + "loss": 0.0097, + "step": 1972 + }, + { + "epoch": 1.530643910007758, + "grad_norm": 0.4158245027065277, + "learning_rate": 1.2190422951941906e-05, + "loss": 0.0056, + "step": 1973 + }, + { + "epoch": 1.5314197051978278, + "grad_norm": 6.8286285400390625, + "learning_rate": 1.2164823052663281e-05, + "loss": 0.2418, + "step": 1974 + }, + { + "epoch": 1.5321955003878975, + "grad_norm": 7.76875114440918, + "learning_rate": 1.21392465992168e-05, + "loss": 0.353, + "step": 1975 + }, + { + "epoch": 1.5329712955779673, + "grad_norm": 5.978578567504883, + "learning_rate": 1.211369367623849e-05, + "loss": 0.2684, + "step": 1976 + }, + { + "epoch": 1.5337470907680373, + "grad_norm": 7.003124237060547, + "learning_rate": 1.2088164368286514e-05, + "loss": 0.3902, + "step": 1977 + }, + { + "epoch": 1.534522885958107, + "grad_norm": 5.4184651374816895, + "learning_rate": 1.2062658759840887e-05, + "loss": 0.0946, + "step": 1978 + }, + { + "epoch": 1.535298681148177, + "grad_norm": 5.319007873535156, + "learning_rate": 1.203717693530321e-05, + "loss": 0.1842, + "step": 1979 + }, + { + "epoch": 1.5360744763382468, + "grad_norm": 5.597511291503906, + "learning_rate": 1.201171897899636e-05, + "loss": 0.1735, + "step": 1980 + }, + { + "epoch": 1.5360744763382468, + "eval_Qnli-dev_cosine_accuracy": 0.677734375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8231859803199768, + "eval_Qnli-dev_cosine_ap": 0.717294715962436, + "eval_Qnli-dev_cosine_f1": 0.6870748299319728, + "eval_Qnli-dev_cosine_f1_threshold": 0.7208942174911499, + "eval_Qnli-dev_cosine_precision": 0.5738636363636364, + "eval_Qnli-dev_cosine_recall": 0.8559322033898306, + "eval_Qnli-dev_dot_accuracy": 0.666015625, + "eval_Qnli-dev_dot_accuracy_threshold": 358.2715148925781, + "eval_Qnli-dev_dot_ap": 0.6315403611821628, + "eval_Qnli-dev_dot_f1": 0.6884735202492211, + "eval_Qnli-dev_dot_f1_threshold": 304.59088134765625, + "eval_Qnli-dev_dot_precision": 0.5443349753694581, + "eval_Qnli-dev_dot_recall": 0.9364406779661016, + "eval_Qnli-dev_euclidean_accuracy": 0.68359375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.641965866088867, + "eval_Qnli-dev_euclidean_ap": 0.7195344050327303, + "eval_Qnli-dev_euclidean_f1": 0.6869983948635633, + "eval_Qnli-dev_euclidean_f1_threshold": 17.36569595336914, + "eval_Qnli-dev_euclidean_precision": 0.5529715762273901, + "eval_Qnli-dev_euclidean_recall": 0.9067796610169492, + "eval_Qnli-dev_manhattan_accuracy": 0.68359375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 249.661376953125, + "eval_Qnli-dev_manhattan_ap": 0.7221799942409435, + "eval_Qnli-dev_manhattan_f1": 0.6839546191247974, + "eval_Qnli-dev_manhattan_f1_threshold": 314.0516357421875, + "eval_Qnli-dev_manhattan_precision": 0.5538057742782152, + "eval_Qnli-dev_manhattan_recall": 0.8940677966101694, + "eval_Qnli-dev_max_accuracy": 0.68359375, + "eval_Qnli-dev_max_accuracy_threshold": 358.2715148925781, + "eval_Qnli-dev_max_ap": 0.7221799942409435, + "eval_Qnli-dev_max_f1": 0.6884735202492211, + "eval_Qnli-dev_max_f1_threshold": 314.0516357421875, + "eval_Qnli-dev_max_precision": 0.5738636363636364, + "eval_Qnli-dev_max_recall": 0.9364406779661016, + "eval_allNLI-dev_cosine_accuracy": 0.73046875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8565464615821838, + "eval_allNLI-dev_cosine_ap": 0.6167725023489327, + "eval_allNLI-dev_cosine_f1": 0.6282973621103117, + "eval_allNLI-dev_cosine_f1_threshold": 0.7900665998458862, + "eval_allNLI-dev_cosine_precision": 0.5368852459016393, + "eval_allNLI-dev_cosine_recall": 0.7572254335260116, + "eval_allNLI-dev_dot_accuracy": 0.712890625, + "eval_allNLI-dev_dot_accuracy_threshold": 369.78729248046875, + "eval_allNLI-dev_dot_ap": 0.5335401963981521, + "eval_allNLI-dev_dot_f1": 0.5906976744186047, + "eval_allNLI-dev_dot_f1_threshold": 344.4092712402344, + "eval_allNLI-dev_dot_precision": 0.49416342412451364, + "eval_allNLI-dev_dot_recall": 0.7341040462427746, + "eval_allNLI-dev_euclidean_accuracy": 0.728515625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 12.005903244018555, + "eval_allNLI-dev_euclidean_ap": 0.6164918200068363, + "eval_allNLI-dev_euclidean_f1": 0.640909090909091, + "eval_allNLI-dev_euclidean_f1_threshold": 14.150827407836914, + "eval_allNLI-dev_euclidean_precision": 0.5280898876404494, + "eval_allNLI-dev_euclidean_recall": 0.815028901734104, + "eval_allNLI-dev_manhattan_accuracy": 0.732421875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 222.51718139648438, + "eval_allNLI-dev_manhattan_ap": 0.613799738233737, + "eval_allNLI-dev_manhattan_f1": 0.6323185011709601, + "eval_allNLI-dev_manhattan_f1_threshold": 252.96902465820312, + "eval_allNLI-dev_manhattan_precision": 0.531496062992126, + "eval_allNLI-dev_manhattan_recall": 0.7803468208092486, + "eval_allNLI-dev_max_accuracy": 0.732421875, + "eval_allNLI-dev_max_accuracy_threshold": 369.78729248046875, + "eval_allNLI-dev_max_ap": 0.6167725023489327, + "eval_allNLI-dev_max_f1": 0.640909090909091, + "eval_allNLI-dev_max_f1_threshold": 344.4092712402344, + "eval_allNLI-dev_max_precision": 0.5368852459016393, + "eval_allNLI-dev_max_recall": 0.815028901734104, + "eval_sequential_score": 0.7221799942409435, + "eval_sts-test_pearson_cosine": 0.8338440690370794, + "eval_sts-test_pearson_dot": 0.8003966532430564, + "eval_sts-test_pearson_euclidean": 0.8674465709349157, + "eval_sts-test_pearson_manhattan": 0.8663742858203357, + "eval_sts-test_pearson_max": 0.8674465709349157, + "eval_sts-test_spearman_cosine": 0.8731866122689873, + "eval_sts-test_spearman_dot": 0.7977475761397881, + "eval_sts-test_spearman_euclidean": 0.8671518599190956, + "eval_sts-test_spearman_manhattan": 0.8681012579295143, + "eval_sts-test_spearman_max": 0.8731866122689873, + "eval_vitaminc-pairs_loss": 3.9380064010620117, + "eval_vitaminc-pairs_runtime": 5.4729, + "eval_vitaminc-pairs_samples_per_second": 23.388, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 1980 + }, + { + "epoch": 1.5360744763382468, + "eval_negation-triplets_loss": 1.7180496454238892, + "eval_negation-triplets_runtime": 1.0785, + "eval_negation-triplets_samples_per_second": 118.683, + "eval_negation-triplets_steps_per_second": 1.854, + "step": 1980 + }, + { + "epoch": 1.5360744763382468, + "eval_scitail-pairs-pos_loss": 0.14408084750175476, + "eval_scitail-pairs-pos_runtime": 1.3048, + "eval_scitail-pairs-pos_samples_per_second": 98.1, + "eval_scitail-pairs-pos_steps_per_second": 1.533, + "step": 1980 + }, + { + "epoch": 1.5360744763382468, + "eval_scitail-pairs-qa_loss": 0.003421870991587639, + "eval_scitail-pairs-qa_runtime": 0.7746, + "eval_scitail-pairs-qa_samples_per_second": 165.245, + "eval_scitail-pairs-qa_steps_per_second": 2.582, + "step": 1980 + }, + { + "epoch": 1.5360744763382468, + "eval_xsum-pairs_loss": 0.05674615874886513, + "eval_xsum-pairs_runtime": 3.9465, + "eval_xsum-pairs_samples_per_second": 32.434, + "eval_xsum-pairs_steps_per_second": 0.507, + "step": 1980 + }, + { + "epoch": 1.5360744763382468, + "eval_sciq_pairs_loss": 0.04385075718164444, + "eval_sciq_pairs_runtime": 6.1803, + "eval_sciq_pairs_samples_per_second": 20.711, + "eval_sciq_pairs_steps_per_second": 0.324, + "step": 1980 + }, + { + "epoch": 1.5360744763382468, + "eval_qasc_pairs_loss": 0.16773392260074615, + "eval_qasc_pairs_runtime": 0.9327, + "eval_qasc_pairs_samples_per_second": 137.23, + "eval_qasc_pairs_steps_per_second": 2.144, + "step": 1980 + }, + { + "epoch": 1.5360744763382468, + "eval_openbookqa_pairs_loss": 1.1240838766098022, + "eval_openbookqa_pairs_runtime": 0.9156, + "eval_openbookqa_pairs_samples_per_second": 139.8, + "eval_openbookqa_pairs_steps_per_second": 2.184, + "step": 1980 + }, + { + "epoch": 1.5360744763382468, + "eval_msmarco_pairs_loss": 0.34464114904403687, + "eval_msmarco_pairs_runtime": 1.9717, + "eval_msmarco_pairs_samples_per_second": 64.919, + "eval_msmarco_pairs_steps_per_second": 1.014, + "step": 1980 + }, + { + "epoch": 1.5360744763382468, + "eval_nq_pairs_loss": 0.5673208832740784, + "eval_nq_pairs_runtime": 3.3298, + "eval_nq_pairs_samples_per_second": 38.441, + "eval_nq_pairs_steps_per_second": 0.601, + "step": 1980 + }, + { + "epoch": 1.5360744763382468, + "eval_trivia_pairs_loss": 0.6362181305885315, + "eval_trivia_pairs_runtime": 4.5074, + "eval_trivia_pairs_samples_per_second": 28.398, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 1980 + }, + { + "epoch": 1.5360744763382468, + "eval_gooaq_pairs_loss": 0.5666394829750061, + "eval_gooaq_pairs_runtime": 1.4678, + "eval_gooaq_pairs_samples_per_second": 87.203, + "eval_gooaq_pairs_steps_per_second": 1.363, + "step": 1980 + }, + { + "epoch": 1.5360744763382468, + "eval_paws-pos_loss": 0.031658001244068146, + "eval_paws-pos_runtime": 1.0363, + "eval_paws-pos_samples_per_second": 123.512, + "eval_paws-pos_steps_per_second": 1.93, + "step": 1980 + }, + { + "epoch": 1.5360744763382468, + "eval_global_dataset_loss": 0.4061763882637024, + "eval_global_dataset_runtime": 19.4111, + "eval_global_dataset_samples_per_second": 21.431, + "eval_global_dataset_steps_per_second": 0.361, + "step": 1980 + }, + { + "epoch": 1.5368502715283165, + "grad_norm": 14.809601783752441, + "learning_rate": 1.1986284975164245e-05, + "loss": 1.8022, + "step": 1981 + }, + { + "epoch": 1.5376260667183863, + "grad_norm": 8.542431831359863, + "learning_rate": 1.1960875007971514e-05, + "loss": 0.367, + "step": 1982 + }, + { + "epoch": 1.538401861908456, + "grad_norm": 2.4377570152282715, + "learning_rate": 1.1935489161503256e-05, + "loss": 0.0372, + "step": 1983 + }, + { + "epoch": 1.5391776570985258, + "grad_norm": 6.3503241539001465, + "learning_rate": 1.1910127519764762e-05, + "loss": 0.1395, + "step": 1984 + }, + { + "epoch": 1.5399534522885958, + "grad_norm": 9.770283699035645, + "learning_rate": 1.1884790166681223e-05, + "loss": 0.5904, + "step": 1985 + }, + { + "epoch": 1.5407292474786658, + "grad_norm": 3.4733684062957764, + "learning_rate": 1.1859477186097439e-05, + "loss": 0.1003, + "step": 1986 + }, + { + "epoch": 1.5415050426687356, + "grad_norm": 4.320209980010986, + "learning_rate": 1.1834188661777578e-05, + "loss": 0.1523, + "step": 1987 + }, + { + "epoch": 1.5422808378588053, + "grad_norm": 7.825626373291016, + "learning_rate": 1.1808924677404873e-05, + "loss": 0.3867, + "step": 1988 + }, + { + "epoch": 1.543056633048875, + "grad_norm": 2.2164692878723145, + "learning_rate": 1.1783685316581337e-05, + "loss": 0.0557, + "step": 1989 + }, + { + "epoch": 1.5438324282389448, + "grad_norm": 10.42979907989502, + "learning_rate": 1.1758470662827519e-05, + "loss": 0.5611, + "step": 1990 + }, + { + "epoch": 1.5446082234290146, + "grad_norm": 10.756227493286133, + "learning_rate": 1.1733280799582211e-05, + "loss": 0.5927, + "step": 1991 + }, + { + "epoch": 1.5453840186190846, + "grad_norm": 3.7764925956726074, + "learning_rate": 1.1708115810202142e-05, + "loss": 0.1493, + "step": 1992 + }, + { + "epoch": 1.5461598138091543, + "grad_norm": 7.408784866333008, + "learning_rate": 1.168297577796176e-05, + "loss": 0.4362, + "step": 1993 + }, + { + "epoch": 1.5469356089992243, + "grad_norm": 6.916841983795166, + "learning_rate": 1.1657860786052909e-05, + "loss": 0.4276, + "step": 1994 + }, + { + "epoch": 1.547711404189294, + "grad_norm": 5.064115524291992, + "learning_rate": 1.1632770917584588e-05, + "loss": 0.1307, + "step": 1995 + }, + { + "epoch": 1.5484871993793639, + "grad_norm": 2.644122362136841, + "learning_rate": 1.1607706255582631e-05, + "loss": 0.0838, + "step": 1996 + }, + { + "epoch": 1.5492629945694336, + "grad_norm": 3.820586681365967, + "learning_rate": 1.1582666882989488e-05, + "loss": 0.091, + "step": 1997 + }, + { + "epoch": 1.5500387897595034, + "grad_norm": 5.767513751983643, + "learning_rate": 1.1557652882663915e-05, + "loss": 0.411, + "step": 1998 + }, + { + "epoch": 1.5508145849495734, + "grad_norm": 6.469566822052002, + "learning_rate": 1.1532664337380698e-05, + "loss": 0.2367, + "step": 1999 + }, + { + "epoch": 1.5515903801396431, + "grad_norm": 8.817070007324219, + "learning_rate": 1.1507701329830392e-05, + "loss": 0.3979, + "step": 2000 + }, + { + "epoch": 1.5515903801396431, + "eval_Qnli-dev_cosine_accuracy": 0.68359375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8214009404182434, + "eval_Qnli-dev_cosine_ap": 0.7195690268775171, + "eval_Qnli-dev_cosine_f1": 0.6885758998435054, + "eval_Qnli-dev_cosine_f1_threshold": 0.668143630027771, + "eval_Qnli-dev_cosine_precision": 0.5459057071960298, + "eval_Qnli-dev_cosine_recall": 0.9322033898305084, + "eval_Qnli-dev_dot_accuracy": 0.662109375, + "eval_Qnli-dev_dot_accuracy_threshold": 357.72027587890625, + "eval_Qnli-dev_dot_ap": 0.6445440797335545, + "eval_Qnli-dev_dot_f1": 0.6882716049382716, + "eval_Qnli-dev_dot_f1_threshold": 303.049072265625, + "eval_Qnli-dev_dot_precision": 0.5412621359223301, + "eval_Qnli-dev_dot_recall": 0.9449152542372882, + "eval_Qnli-dev_euclidean_accuracy": 0.6796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.49072265625, + "eval_Qnli-dev_euclidean_ap": 0.717855569425595, + "eval_Qnli-dev_euclidean_f1": 0.688888888888889, + "eval_Qnli-dev_euclidean_f1_threshold": 17.393165588378906, + "eval_Qnli-dev_euclidean_precision": 0.550761421319797, + "eval_Qnli-dev_euclidean_recall": 0.9194915254237288, + "eval_Qnli-dev_manhattan_accuracy": 0.68359375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 254.56642150878906, + "eval_Qnli-dev_manhattan_ap": 0.7200163748793393, + "eval_Qnli-dev_manhattan_f1": 0.6848, + "eval_Qnli-dev_manhattan_f1_threshold": 320.8479919433594, + "eval_Qnli-dev_manhattan_precision": 0.5501285347043702, + "eval_Qnli-dev_manhattan_recall": 0.9067796610169492, + "eval_Qnli-dev_max_accuracy": 0.68359375, + "eval_Qnli-dev_max_accuracy_threshold": 357.72027587890625, + "eval_Qnli-dev_max_ap": 0.7200163748793393, + "eval_Qnli-dev_max_f1": 0.688888888888889, + "eval_Qnli-dev_max_f1_threshold": 320.8479919433594, + "eval_Qnli-dev_max_precision": 0.550761421319797, + "eval_Qnli-dev_max_recall": 0.9449152542372882, + "eval_allNLI-dev_cosine_accuracy": 0.736328125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8489243984222412, + "eval_allNLI-dev_cosine_ap": 0.618271099125419, + "eval_allNLI-dev_cosine_f1": 0.6339066339066338, + "eval_allNLI-dev_cosine_f1_threshold": 0.7980220317840576, + "eval_allNLI-dev_cosine_precision": 0.5512820512820513, + "eval_allNLI-dev_cosine_recall": 0.7456647398843931, + "eval_allNLI-dev_dot_accuracy": 0.69921875, + "eval_allNLI-dev_dot_accuracy_threshold": 368.255615234375, + "eval_allNLI-dev_dot_ap": 0.5258246950841056, + "eval_allNLI-dev_dot_f1": 0.5957446808510639, + "eval_allNLI-dev_dot_f1_threshold": 336.3111267089844, + "eval_allNLI-dev_dot_precision": 0.4713804713804714, + "eval_allNLI-dev_dot_recall": 0.8092485549132948, + "eval_allNLI-dev_euclidean_accuracy": 0.734375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.805103302001953, + "eval_allNLI-dev_euclidean_ap": 0.62126525868756, + "eval_allNLI-dev_euclidean_f1": 0.6436781609195402, + "eval_allNLI-dev_euclidean_f1_threshold": 14.103891372680664, + "eval_allNLI-dev_euclidean_precision": 0.5343511450381679, + "eval_allNLI-dev_euclidean_recall": 0.8092485549132948, + "eval_allNLI-dev_manhattan_accuracy": 0.73828125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 226.9370574951172, + "eval_allNLI-dev_manhattan_ap": 0.616090308418179, + "eval_allNLI-dev_manhattan_f1": 0.638095238095238, + "eval_allNLI-dev_manhattan_f1_threshold": 248.9129638671875, + "eval_allNLI-dev_manhattan_precision": 0.5425101214574899, + "eval_allNLI-dev_manhattan_recall": 0.7745664739884393, + "eval_allNLI-dev_max_accuracy": 0.73828125, + "eval_allNLI-dev_max_accuracy_threshold": 368.255615234375, + "eval_allNLI-dev_max_ap": 0.62126525868756, + "eval_allNLI-dev_max_f1": 0.6436781609195402, + "eval_allNLI-dev_max_f1_threshold": 336.3111267089844, + "eval_allNLI-dev_max_precision": 0.5512820512820513, + "eval_allNLI-dev_max_recall": 0.8092485549132948, + "eval_sequential_score": 0.7200163748793393, + "eval_sts-test_pearson_cosine": 0.8320851075646946, + "eval_sts-test_pearson_dot": 0.7903761901046727, + "eval_sts-test_pearson_euclidean": 0.8663481935999435, + "eval_sts-test_pearson_manhattan": 0.8643649057521037, + "eval_sts-test_pearson_max": 0.8663481935999435, + "eval_sts-test_spearman_cosine": 0.8734033397744375, + "eval_sts-test_spearman_dot": 0.7874238181468833, + "eval_sts-test_spearman_euclidean": 0.867574490831676, + "eval_sts-test_spearman_manhattan": 0.8669370091131554, + "eval_sts-test_spearman_max": 0.8734033397744375, + "eval_vitaminc-pairs_loss": 3.953611135482788, + "eval_vitaminc-pairs_runtime": 5.4842, + "eval_vitaminc-pairs_samples_per_second": 23.34, + "eval_vitaminc-pairs_steps_per_second": 0.365, + "step": 2000 + }, + { + "epoch": 1.5515903801396431, + "eval_negation-triplets_loss": 1.687469482421875, + "eval_negation-triplets_runtime": 1.0791, + "eval_negation-triplets_samples_per_second": 118.621, + "eval_negation-triplets_steps_per_second": 1.853, + "step": 2000 + }, + { + "epoch": 1.5515903801396431, + "eval_scitail-pairs-pos_loss": 0.12901288270950317, + "eval_scitail-pairs-pos_runtime": 1.3142, + "eval_scitail-pairs-pos_samples_per_second": 97.396, + "eval_scitail-pairs-pos_steps_per_second": 1.522, + "step": 2000 + }, + { + "epoch": 1.5515903801396431, + "eval_scitail-pairs-qa_loss": 0.003454263787716627, + "eval_scitail-pairs-qa_runtime": 0.7754, + "eval_scitail-pairs-qa_samples_per_second": 165.073, + "eval_scitail-pairs-qa_steps_per_second": 2.579, + "step": 2000 + }, + { + "epoch": 1.5515903801396431, + "eval_xsum-pairs_loss": 0.09570430964231491, + "eval_xsum-pairs_runtime": 3.9202, + "eval_xsum-pairs_samples_per_second": 32.651, + "eval_xsum-pairs_steps_per_second": 0.51, + "step": 2000 + }, + { + "epoch": 1.5515903801396431, + "eval_sciq_pairs_loss": 0.041838087141513824, + "eval_sciq_pairs_runtime": 5.9906, + "eval_sciq_pairs_samples_per_second": 21.367, + "eval_sciq_pairs_steps_per_second": 0.334, + "step": 2000 + }, + { + "epoch": 1.5515903801396431, + "eval_qasc_pairs_loss": 0.1679731160402298, + "eval_qasc_pairs_runtime": 0.9165, + "eval_qasc_pairs_samples_per_second": 139.657, + "eval_qasc_pairs_steps_per_second": 2.182, + "step": 2000 + }, + { + "epoch": 1.5515903801396431, + "eval_openbookqa_pairs_loss": 1.0533864498138428, + "eval_openbookqa_pairs_runtime": 0.8821, + "eval_openbookqa_pairs_samples_per_second": 145.112, + "eval_openbookqa_pairs_steps_per_second": 2.267, + "step": 2000 + }, + { + "epoch": 1.5515903801396431, + "eval_msmarco_pairs_loss": 0.32383444905281067, + "eval_msmarco_pairs_runtime": 1.9889, + "eval_msmarco_pairs_samples_per_second": 64.358, + "eval_msmarco_pairs_steps_per_second": 1.006, + "step": 2000 + }, + { + "epoch": 1.5515903801396431, + "eval_nq_pairs_loss": 0.5685637593269348, + "eval_nq_pairs_runtime": 3.3604, + "eval_nq_pairs_samples_per_second": 38.09, + "eval_nq_pairs_steps_per_second": 0.595, + "step": 2000 + }, + { + "epoch": 1.5515903801396431, + "eval_trivia_pairs_loss": 0.6671900749206543, + "eval_trivia_pairs_runtime": 4.5038, + "eval_trivia_pairs_samples_per_second": 28.421, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 2000 + }, + { + "epoch": 1.5515903801396431, + "eval_gooaq_pairs_loss": 0.5941081047058105, + "eval_gooaq_pairs_runtime": 1.4774, + "eval_gooaq_pairs_samples_per_second": 86.636, + "eval_gooaq_pairs_steps_per_second": 1.354, + "step": 2000 + }, + { + "epoch": 1.5515903801396431, + "eval_paws-pos_loss": 0.030928507447242737, + "eval_paws-pos_runtime": 1.0437, + "eval_paws-pos_samples_per_second": 122.635, + "eval_paws-pos_steps_per_second": 1.916, + "step": 2000 + }, + { + "epoch": 1.5515903801396431, + "eval_global_dataset_loss": 0.4055573344230652, + "eval_global_dataset_runtime": 19.4412, + "eval_global_dataset_samples_per_second": 21.398, + "eval_global_dataset_steps_per_second": 0.36, + "step": 2000 + }, + { + "epoch": 1.552366175329713, + "grad_norm": 6.698705196380615, + "learning_rate": 1.148276394261906e-05, + "loss": 0.2013, + "step": 2001 + }, + { + "epoch": 1.5531419705197829, + "grad_norm": 7.605216026306152, + "learning_rate": 1.145785225826796e-05, + "loss": 0.2821, + "step": 2002 + }, + { + "epoch": 1.5539177657098526, + "grad_norm": 5.610411167144775, + "learning_rate": 1.143296635921331e-05, + "loss": 0.1185, + "step": 2003 + }, + { + "epoch": 1.5546935608999224, + "grad_norm": 17.075592041015625, + "learning_rate": 1.1408106327806e-05, + "loss": 2.1549, + "step": 2004 + }, + { + "epoch": 1.5554693560899922, + "grad_norm": 4.582951068878174, + "learning_rate": 1.1383272246311315e-05, + "loss": 0.1351, + "step": 2005 + }, + { + "epoch": 1.556245151280062, + "grad_norm": 5.0711541175842285, + "learning_rate": 1.1358464196908675e-05, + "loss": 0.1189, + "step": 2006 + }, + { + "epoch": 1.557020946470132, + "grad_norm": 8.517909049987793, + "learning_rate": 1.1333682261691353e-05, + "loss": 0.3583, + "step": 2007 + }, + { + "epoch": 1.5577967416602017, + "grad_norm": 4.800745487213135, + "learning_rate": 1.1308926522666199e-05, + "loss": 0.1171, + "step": 2008 + }, + { + "epoch": 1.5585725368502716, + "grad_norm": 2.900177001953125, + "learning_rate": 1.1284197061753389e-05, + "loss": 0.0708, + "step": 2009 + }, + { + "epoch": 1.5593483320403414, + "grad_norm": 0.14408570528030396, + "learning_rate": 1.1259493960786132e-05, + "loss": 0.0023, + "step": 2010 + }, + { + "epoch": 1.5601241272304112, + "grad_norm": 3.0433990955352783, + "learning_rate": 1.1234817301510425e-05, + "loss": 0.0636, + "step": 2011 + }, + { + "epoch": 1.560899922420481, + "grad_norm": 4.390824794769287, + "learning_rate": 1.1210167165584732e-05, + "loss": 0.0936, + "step": 2012 + }, + { + "epoch": 1.5616757176105507, + "grad_norm": 8.155580520629883, + "learning_rate": 1.1185543634579775e-05, + "loss": 0.3519, + "step": 2013 + }, + { + "epoch": 1.5624515128006207, + "grad_norm": 7.241706371307373, + "learning_rate": 1.116094678997824e-05, + "loss": 0.2882, + "step": 2014 + }, + { + "epoch": 1.5632273079906904, + "grad_norm": 0.1498359888792038, + "learning_rate": 1.113637671317448e-05, + "loss": 0.002, + "step": 2015 + }, + { + "epoch": 1.5640031031807604, + "grad_norm": 6.577805042266846, + "learning_rate": 1.111183348547429e-05, + "loss": 0.1809, + "step": 2016 + }, + { + "epoch": 1.5647788983708302, + "grad_norm": 6.228726387023926, + "learning_rate": 1.1087317188094616e-05, + "loss": 0.1562, + "step": 2017 + }, + { + "epoch": 1.5655546935609, + "grad_norm": 2.865723133087158, + "learning_rate": 1.1062827902163268e-05, + "loss": 0.0306, + "step": 2018 + }, + { + "epoch": 1.5663304887509697, + "grad_norm": 1.2260462045669556, + "learning_rate": 1.1038365708718695e-05, + "loss": 0.0163, + "step": 2019 + }, + { + "epoch": 1.5671062839410395, + "grad_norm": 0.04684958979487419, + "learning_rate": 1.101393068870969e-05, + "loss": 0.0006, + "step": 2020 + }, + { + "epoch": 1.5671062839410395, + "eval_Qnli-dev_cosine_accuracy": 0.685546875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8127551078796387, + "eval_Qnli-dev_cosine_ap": 0.725531069812458, + "eval_Qnli-dev_cosine_f1": 0.689655172413793, + "eval_Qnli-dev_cosine_f1_threshold": 0.7286459803581238, + "eval_Qnli-dev_cosine_precision": 0.5813953488372093, + "eval_Qnli-dev_cosine_recall": 0.847457627118644, + "eval_Qnli-dev_dot_accuracy": 0.65625, + "eval_Qnli-dev_dot_accuracy_threshold": 385.49615478515625, + "eval_Qnli-dev_dot_ap": 0.6464360273948009, + "eval_Qnli-dev_dot_f1": 0.689655172413793, + "eval_Qnli-dev_dot_f1_threshold": 308.6428527832031, + "eval_Qnli-dev_dot_precision": 0.5472636815920398, + "eval_Qnli-dev_dot_recall": 0.9322033898305084, + "eval_Qnli-dev_euclidean_accuracy": 0.6875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.204526901245117, + "eval_Qnli-dev_euclidean_ap": 0.726181492828112, + "eval_Qnli-dev_euclidean_f1": 0.688, + "eval_Qnli-dev_euclidean_f1_threshold": 17.032333374023438, + "eval_Qnli-dev_euclidean_precision": 0.5526992287917738, + "eval_Qnli-dev_euclidean_recall": 0.9110169491525424, + "eval_Qnli-dev_manhattan_accuracy": 0.685546875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 247.1206512451172, + "eval_Qnli-dev_manhattan_ap": 0.7267829940658428, + "eval_Qnli-dev_manhattan_f1": 0.6842923794712287, + "eval_Qnli-dev_manhattan_f1_threshold": 319.94061279296875, + "eval_Qnli-dev_manhattan_precision": 0.5405405405405406, + "eval_Qnli-dev_manhattan_recall": 0.9322033898305084, + "eval_Qnli-dev_max_accuracy": 0.6875, + "eval_Qnli-dev_max_accuracy_threshold": 385.49615478515625, + "eval_Qnli-dev_max_ap": 0.7267829940658428, + "eval_Qnli-dev_max_f1": 0.689655172413793, + "eval_Qnli-dev_max_f1_threshold": 319.94061279296875, + "eval_Qnli-dev_max_precision": 0.5813953488372093, + "eval_Qnli-dev_max_recall": 0.9322033898305084, + "eval_allNLI-dev_cosine_accuracy": 0.728515625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8414983153343201, + "eval_allNLI-dev_cosine_ap": 0.6145838814543069, + "eval_allNLI-dev_cosine_f1": 0.6263982102908278, + "eval_allNLI-dev_cosine_f1_threshold": 0.765398383140564, + "eval_allNLI-dev_cosine_precision": 0.5109489051094891, + "eval_allNLI-dev_cosine_recall": 0.8092485549132948, + "eval_allNLI-dev_dot_accuracy": 0.69921875, + "eval_allNLI-dev_dot_accuracy_threshold": 374.797119140625, + "eval_allNLI-dev_dot_ap": 0.5337365353789146, + "eval_allNLI-dev_dot_f1": 0.5944700460829493, + "eval_allNLI-dev_dot_f1_threshold": 339.7652587890625, + "eval_allNLI-dev_dot_precision": 0.4942528735632184, + "eval_allNLI-dev_dot_recall": 0.7456647398843931, + "eval_allNLI-dev_euclidean_accuracy": 0.73046875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.778360366821289, + "eval_allNLI-dev_euclidean_ap": 0.6151725642670107, + "eval_allNLI-dev_euclidean_f1": 0.6391304347826088, + "eval_allNLI-dev_euclidean_f1_threshold": 14.77199935913086, + "eval_allNLI-dev_euclidean_precision": 0.5121951219512195, + "eval_allNLI-dev_euclidean_recall": 0.8497109826589595, + "eval_allNLI-dev_manhattan_accuracy": 0.73046875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 218.05601501464844, + "eval_allNLI-dev_manhattan_ap": 0.6136640953944457, + "eval_allNLI-dev_manhattan_f1": 0.6359447004608295, + "eval_allNLI-dev_manhattan_f1_threshold": 256.1634216308594, + "eval_allNLI-dev_manhattan_precision": 0.5287356321839081, + "eval_allNLI-dev_manhattan_recall": 0.7976878612716763, + "eval_allNLI-dev_max_accuracy": 0.73046875, + "eval_allNLI-dev_max_accuracy_threshold": 374.797119140625, + "eval_allNLI-dev_max_ap": 0.6151725642670107, + "eval_allNLI-dev_max_f1": 0.6391304347826088, + "eval_allNLI-dev_max_f1_threshold": 339.7652587890625, + "eval_allNLI-dev_max_precision": 0.5287356321839081, + "eval_allNLI-dev_max_recall": 0.8497109826589595, + "eval_sequential_score": 0.7267829940658428, + "eval_sts-test_pearson_cosine": 0.8348389897427306, + "eval_sts-test_pearson_dot": 0.8018483996500129, + "eval_sts-test_pearson_euclidean": 0.8705508637456714, + "eval_sts-test_pearson_manhattan": 0.8681967553299961, + "eval_sts-test_pearson_max": 0.8705508637456714, + "eval_sts-test_spearman_cosine": 0.8755758821146743, + "eval_sts-test_spearman_dot": 0.7957245058660671, + "eval_sts-test_spearman_euclidean": 0.8699214259840865, + "eval_sts-test_spearman_manhattan": 0.8694451383090979, + "eval_sts-test_spearman_max": 0.8755758821146743, + "eval_vitaminc-pairs_loss": 3.9671263694763184, + "eval_vitaminc-pairs_runtime": 5.4881, + "eval_vitaminc-pairs_samples_per_second": 23.323, + "eval_vitaminc-pairs_steps_per_second": 0.364, + "step": 2020 + }, + { + "epoch": 1.5671062839410395, + "eval_negation-triplets_loss": 1.7057411670684814, + "eval_negation-triplets_runtime": 1.0774, + "eval_negation-triplets_samples_per_second": 118.799, + "eval_negation-triplets_steps_per_second": 1.856, + "step": 2020 + }, + { + "epoch": 1.5671062839410395, + "eval_scitail-pairs-pos_loss": 0.12558433413505554, + "eval_scitail-pairs-pos_runtime": 1.3346, + "eval_scitail-pairs-pos_samples_per_second": 95.911, + "eval_scitail-pairs-pos_steps_per_second": 1.499, + "step": 2020 + }, + { + "epoch": 1.5671062839410395, + "eval_scitail-pairs-qa_loss": 0.004487679805606604, + "eval_scitail-pairs-qa_runtime": 0.7853, + "eval_scitail-pairs-qa_samples_per_second": 162.988, + "eval_scitail-pairs-qa_steps_per_second": 2.547, + "step": 2020 + }, + { + "epoch": 1.5671062839410395, + "eval_xsum-pairs_loss": 0.08499541878700256, + "eval_xsum-pairs_runtime": 3.9399, + "eval_xsum-pairs_samples_per_second": 32.488, + "eval_xsum-pairs_steps_per_second": 0.508, + "step": 2020 + }, + { + "epoch": 1.5671062839410395, + "eval_sciq_pairs_loss": 0.042903732508420944, + "eval_sciq_pairs_runtime": 6.06, + "eval_sciq_pairs_samples_per_second": 21.122, + "eval_sciq_pairs_steps_per_second": 0.33, + "step": 2020 + }, + { + "epoch": 1.5671062839410395, + "eval_qasc_pairs_loss": 0.16728368401527405, + "eval_qasc_pairs_runtime": 0.9328, + "eval_qasc_pairs_samples_per_second": 137.215, + "eval_qasc_pairs_steps_per_second": 2.144, + "step": 2020 + }, + { + "epoch": 1.5671062839410395, + "eval_openbookqa_pairs_loss": 1.0172346830368042, + "eval_openbookqa_pairs_runtime": 0.8973, + "eval_openbookqa_pairs_samples_per_second": 142.651, + "eval_openbookqa_pairs_steps_per_second": 2.229, + "step": 2020 + }, + { + "epoch": 1.5671062839410395, + "eval_msmarco_pairs_loss": 0.33615532517433167, + "eval_msmarco_pairs_runtime": 1.9829, + "eval_msmarco_pairs_samples_per_second": 64.554, + "eval_msmarco_pairs_steps_per_second": 1.009, + "step": 2020 + }, + { + "epoch": 1.5671062839410395, + "eval_nq_pairs_loss": 0.5002948045730591, + "eval_nq_pairs_runtime": 3.3501, + "eval_nq_pairs_samples_per_second": 38.207, + "eval_nq_pairs_steps_per_second": 0.597, + "step": 2020 + }, + { + "epoch": 1.5671062839410395, + "eval_trivia_pairs_loss": 0.666401207447052, + "eval_trivia_pairs_runtime": 4.5031, + "eval_trivia_pairs_samples_per_second": 28.425, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 2020 + }, + { + "epoch": 1.5671062839410395, + "eval_gooaq_pairs_loss": 0.5669600963592529, + "eval_gooaq_pairs_runtime": 1.4816, + "eval_gooaq_pairs_samples_per_second": 86.391, + "eval_gooaq_pairs_steps_per_second": 1.35, + "step": 2020 + }, + { + "epoch": 1.5671062839410395, + "eval_paws-pos_loss": 0.03100033476948738, + "eval_paws-pos_runtime": 1.039, + "eval_paws-pos_samples_per_second": 123.192, + "eval_paws-pos_steps_per_second": 1.925, + "step": 2020 + }, + { + "epoch": 1.5671062839410395, + "eval_global_dataset_loss": 0.3974339962005615, + "eval_global_dataset_runtime": 19.4595, + "eval_global_dataset_samples_per_second": 21.378, + "eval_global_dataset_steps_per_second": 0.36, + "step": 2020 + }, + { + "epoch": 1.5678820791311094, + "grad_norm": 11.136065483093262, + "learning_rate": 1.0989522922995105e-05, + "loss": 0.5958, + "step": 2021 + }, + { + "epoch": 1.5686578743211792, + "grad_norm": 4.509794235229492, + "learning_rate": 1.096514249234363e-05, + "loss": 0.1131, + "step": 2022 + }, + { + "epoch": 1.5694336695112492, + "grad_norm": 5.513092517852783, + "learning_rate": 1.0940789477433487e-05, + "loss": 0.1063, + "step": 2023 + }, + { + "epoch": 1.570209464701319, + "grad_norm": 0.5959425568580627, + "learning_rate": 1.0916463958852169e-05, + "loss": 0.0157, + "step": 2024 + }, + { + "epoch": 1.5709852598913887, + "grad_norm": 2.8120079040527344, + "learning_rate": 1.0892166017096194e-05, + "loss": 0.0353, + "step": 2025 + }, + { + "epoch": 1.5717610550814585, + "grad_norm": 4.04110050201416, + "learning_rate": 1.086789573257082e-05, + "loss": 0.2481, + "step": 2026 + }, + { + "epoch": 1.5725368502715282, + "grad_norm": 6.810428619384766, + "learning_rate": 1.084365318558977e-05, + "loss": 0.1929, + "step": 2027 + }, + { + "epoch": 1.573312645461598, + "grad_norm": 3.8570098876953125, + "learning_rate": 1.0819438456375e-05, + "loss": 0.0888, + "step": 2028 + }, + { + "epoch": 1.574088440651668, + "grad_norm": 0.6034532785415649, + "learning_rate": 1.0795251625056399e-05, + "loss": 0.0069, + "step": 2029 + }, + { + "epoch": 1.5748642358417377, + "grad_norm": 3.7032084465026855, + "learning_rate": 1.077109277167155e-05, + "loss": 0.1091, + "step": 2030 + }, + { + "epoch": 1.5756400310318077, + "grad_norm": 0.9299270510673523, + "learning_rate": 1.0746961976165436e-05, + "loss": 0.0268, + "step": 2031 + }, + { + "epoch": 1.5764158262218775, + "grad_norm": 3.5875091552734375, + "learning_rate": 1.0722859318390208e-05, + "loss": 0.079, + "step": 2032 + }, + { + "epoch": 1.5771916214119472, + "grad_norm": 8.305997848510742, + "learning_rate": 1.0698784878104907e-05, + "loss": 0.5232, + "step": 2033 + }, + { + "epoch": 1.577967416602017, + "grad_norm": 7.639830589294434, + "learning_rate": 1.0674738734975173e-05, + "loss": 0.2957, + "step": 2034 + }, + { + "epoch": 1.5787432117920868, + "grad_norm": 2.6087515354156494, + "learning_rate": 1.065072096857304e-05, + "loss": 0.0371, + "step": 2035 + }, + { + "epoch": 1.5795190069821567, + "grad_norm": 7.454002857208252, + "learning_rate": 1.0626731658376625e-05, + "loss": 0.275, + "step": 2036 + }, + { + "epoch": 1.5802948021722265, + "grad_norm": 14.170170783996582, + "learning_rate": 1.060277088376987e-05, + "loss": 1.2525, + "step": 2037 + }, + { + "epoch": 1.5810705973622965, + "grad_norm": 7.296122074127197, + "learning_rate": 1.0578838724042301e-05, + "loss": 0.4272, + "step": 2038 + }, + { + "epoch": 1.5818463925523663, + "grad_norm": 1.941819667816162, + "learning_rate": 1.0554935258388758e-05, + "loss": 0.0417, + "step": 2039 + }, + { + "epoch": 1.582622187742436, + "grad_norm": 1.6679277420043945, + "learning_rate": 1.0531060565909108e-05, + "loss": 0.0235, + "step": 2040 + }, + { + "epoch": 1.582622187742436, + "eval_Qnli-dev_cosine_accuracy": 0.685546875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8150365948677063, + "eval_Qnli-dev_cosine_ap": 0.7296418061042323, + "eval_Qnli-dev_cosine_f1": 0.6938053097345133, + "eval_Qnli-dev_cosine_f1_threshold": 0.72687828540802, + "eval_Qnli-dev_cosine_precision": 0.5957446808510638, + "eval_Qnli-dev_cosine_recall": 0.8305084745762712, + "eval_Qnli-dev_dot_accuracy": 0.666015625, + "eval_Qnli-dev_dot_accuracy_threshold": 395.3042907714844, + "eval_Qnli-dev_dot_ap": 0.6498722210876892, + "eval_Qnli-dev_dot_f1": 0.6907378335949764, + "eval_Qnli-dev_dot_f1_threshold": 308.93328857421875, + "eval_Qnli-dev_dot_precision": 0.5486284289276808, + "eval_Qnli-dev_dot_recall": 0.9322033898305084, + "eval_Qnli-dev_euclidean_accuracy": 0.689453125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.69180965423584, + "eval_Qnli-dev_euclidean_ap": 0.7297428124292622, + "eval_Qnli-dev_euclidean_f1": 0.6915254237288135, + "eval_Qnli-dev_euclidean_f1_threshold": 16.685787200927734, + "eval_Qnli-dev_euclidean_precision": 0.576271186440678, + "eval_Qnli-dev_euclidean_recall": 0.864406779661017, + "eval_Qnli-dev_manhattan_accuracy": 0.6875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 251.98843383789062, + "eval_Qnli-dev_manhattan_ap": 0.7320937566127766, + "eval_Qnli-dev_manhattan_f1": 0.6870967741935484, + "eval_Qnli-dev_manhattan_f1_threshold": 316.79962158203125, + "eval_Qnli-dev_manhattan_precision": 0.5546875, + "eval_Qnli-dev_manhattan_recall": 0.902542372881356, + "eval_Qnli-dev_max_accuracy": 0.689453125, + "eval_Qnli-dev_max_accuracy_threshold": 395.3042907714844, + "eval_Qnli-dev_max_ap": 0.7320937566127766, + "eval_Qnli-dev_max_f1": 0.6938053097345133, + "eval_Qnli-dev_max_f1_threshold": 316.79962158203125, + "eval_Qnli-dev_max_precision": 0.5957446808510638, + "eval_Qnli-dev_max_recall": 0.9322033898305084, + "eval_allNLI-dev_cosine_accuracy": 0.732421875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8285204768180847, + "eval_allNLI-dev_cosine_ap": 0.6141849579590164, + "eval_allNLI-dev_cosine_f1": 0.6266666666666667, + "eval_allNLI-dev_cosine_f1_threshold": 0.7529633045196533, + "eval_allNLI-dev_cosine_precision": 0.5090252707581228, + "eval_allNLI-dev_cosine_recall": 0.815028901734104, + "eval_allNLI-dev_dot_accuracy": 0.7109375, + "eval_allNLI-dev_dot_accuracy_threshold": 359.93914794921875, + "eval_allNLI-dev_dot_ap": 0.5294600982940967, + "eval_allNLI-dev_dot_f1": 0.5924453280318092, + "eval_allNLI-dev_dot_f1_threshold": 307.938720703125, + "eval_allNLI-dev_dot_precision": 0.45151515151515154, + "eval_allNLI-dev_dot_recall": 0.861271676300578, + "eval_allNLI-dev_euclidean_accuracy": 0.732421875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.995444297790527, + "eval_allNLI-dev_euclidean_ap": 0.6184983909825983, + "eval_allNLI-dev_euclidean_f1": 0.6435185185185186, + "eval_allNLI-dev_euclidean_f1_threshold": 14.250646591186523, + "eval_allNLI-dev_euclidean_precision": 0.5366795366795367, + "eval_allNLI-dev_euclidean_recall": 0.8034682080924855, + "eval_allNLI-dev_manhattan_accuracy": 0.734375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 218.90469360351562, + "eval_allNLI-dev_manhattan_ap": 0.6165483157481356, + "eval_allNLI-dev_manhattan_f1": 0.639080459770115, + "eval_allNLI-dev_manhattan_f1_threshold": 262.6552734375, + "eval_allNLI-dev_manhattan_precision": 0.5305343511450382, + "eval_allNLI-dev_manhattan_recall": 0.8034682080924855, + "eval_allNLI-dev_max_accuracy": 0.734375, + "eval_allNLI-dev_max_accuracy_threshold": 359.93914794921875, + "eval_allNLI-dev_max_ap": 0.6184983909825983, + "eval_allNLI-dev_max_f1": 0.6435185185185186, + "eval_allNLI-dev_max_f1_threshold": 307.938720703125, + "eval_allNLI-dev_max_precision": 0.5366795366795367, + "eval_allNLI-dev_max_recall": 0.861271676300578, + "eval_sequential_score": 0.7320937566127766, + "eval_sts-test_pearson_cosine": 0.833375433898315, + "eval_sts-test_pearson_dot": 0.7964125377064872, + "eval_sts-test_pearson_euclidean": 0.8688418950926684, + "eval_sts-test_pearson_manhattan": 0.8663011919412575, + "eval_sts-test_pearson_max": 0.8688418950926684, + "eval_sts-test_spearman_cosine": 0.874226452578678, + "eval_sts-test_spearman_dot": 0.7883410847516228, + "eval_sts-test_spearman_euclidean": 0.8690521327315587, + "eval_sts-test_spearman_manhattan": 0.867681177838833, + "eval_sts-test_spearman_max": 0.874226452578678, + "eval_vitaminc-pairs_loss": 3.9999189376831055, + "eval_vitaminc-pairs_runtime": 5.5407, + "eval_vitaminc-pairs_samples_per_second": 23.102, + "eval_vitaminc-pairs_steps_per_second": 0.361, + "step": 2040 + }, + { + "epoch": 1.582622187742436, + "eval_negation-triplets_loss": 1.693063497543335, + "eval_negation-triplets_runtime": 1.102, + "eval_negation-triplets_samples_per_second": 116.15, + "eval_negation-triplets_steps_per_second": 1.815, + "step": 2040 + }, + { + "epoch": 1.582622187742436, + "eval_scitail-pairs-pos_loss": 0.12587006390094757, + "eval_scitail-pairs-pos_runtime": 1.362, + "eval_scitail-pairs-pos_samples_per_second": 93.98, + "eval_scitail-pairs-pos_steps_per_second": 1.468, + "step": 2040 + }, + { + "epoch": 1.582622187742436, + "eval_scitail-pairs-qa_loss": 0.004333545453846455, + "eval_scitail-pairs-qa_runtime": 0.7916, + "eval_scitail-pairs-qa_samples_per_second": 161.708, + "eval_scitail-pairs-qa_steps_per_second": 2.527, + "step": 2040 + }, + { + "epoch": 1.582622187742436, + "eval_xsum-pairs_loss": 0.08355499058961868, + "eval_xsum-pairs_runtime": 3.9416, + "eval_xsum-pairs_samples_per_second": 32.474, + "eval_xsum-pairs_steps_per_second": 0.507, + "step": 2040 + }, + { + "epoch": 1.582622187742436, + "eval_sciq_pairs_loss": 0.04478560760617256, + "eval_sciq_pairs_runtime": 6.0998, + "eval_sciq_pairs_samples_per_second": 20.984, + "eval_sciq_pairs_steps_per_second": 0.328, + "step": 2040 + }, + { + "epoch": 1.582622187742436, + "eval_qasc_pairs_loss": 0.18044772744178772, + "eval_qasc_pairs_runtime": 0.9337, + "eval_qasc_pairs_samples_per_second": 137.091, + "eval_qasc_pairs_steps_per_second": 2.142, + "step": 2040 + }, + { + "epoch": 1.582622187742436, + "eval_openbookqa_pairs_loss": 0.9951231479644775, + "eval_openbookqa_pairs_runtime": 0.883, + "eval_openbookqa_pairs_samples_per_second": 144.954, + "eval_openbookqa_pairs_steps_per_second": 2.265, + "step": 2040 + }, + { + "epoch": 1.582622187742436, + "eval_msmarco_pairs_loss": 0.32004624605178833, + "eval_msmarco_pairs_runtime": 1.9957, + "eval_msmarco_pairs_samples_per_second": 64.137, + "eval_msmarco_pairs_steps_per_second": 1.002, + "step": 2040 + }, + { + "epoch": 1.582622187742436, + "eval_nq_pairs_loss": 0.5070789456367493, + "eval_nq_pairs_runtime": 3.3664, + "eval_nq_pairs_samples_per_second": 38.022, + "eval_nq_pairs_steps_per_second": 0.594, + "step": 2040 + }, + { + "epoch": 1.582622187742436, + "eval_trivia_pairs_loss": 0.6948876976966858, + "eval_trivia_pairs_runtime": 4.5044, + "eval_trivia_pairs_samples_per_second": 28.416, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 2040 + }, + { + "epoch": 1.582622187742436, + "eval_gooaq_pairs_loss": 0.5867024064064026, + "eval_gooaq_pairs_runtime": 1.487, + "eval_gooaq_pairs_samples_per_second": 86.077, + "eval_gooaq_pairs_steps_per_second": 1.345, + "step": 2040 + }, + { + "epoch": 1.582622187742436, + "eval_paws-pos_loss": 0.030770210549235344, + "eval_paws-pos_runtime": 1.0575, + "eval_paws-pos_samples_per_second": 121.046, + "eval_paws-pos_steps_per_second": 1.891, + "step": 2040 + }, + { + "epoch": 1.582622187742436, + "eval_global_dataset_loss": 0.4081159830093384, + "eval_global_dataset_runtime": 19.4931, + "eval_global_dataset_samples_per_second": 21.341, + "eval_global_dataset_steps_per_second": 0.359, + "step": 2040 + }, + { + "epoch": 1.5833979829325058, + "grad_norm": 4.776727199554443, + "learning_rate": 1.0507214725608025e-05, + "loss": 0.1329, + "step": 2041 + }, + { + "epoch": 1.5841737781225755, + "grad_norm": 9.830643653869629, + "learning_rate": 1.0483397816394697e-05, + "loss": 0.5301, + "step": 2042 + }, + { + "epoch": 1.5849495733126453, + "grad_norm": 10.892820358276367, + "learning_rate": 1.0459609917082572e-05, + "loss": 0.6176, + "step": 2043 + }, + { + "epoch": 1.5857253685027153, + "grad_norm": 3.6035313606262207, + "learning_rate": 1.0435851106389107e-05, + "loss": 0.1119, + "step": 2044 + }, + { + "epoch": 1.5865011636927853, + "grad_norm": 8.48658275604248, + "learning_rate": 1.0412121462935499e-05, + "loss": 0.4047, + "step": 2045 + }, + { + "epoch": 1.587276958882855, + "grad_norm": 6.983282089233398, + "learning_rate": 1.0388421065246432e-05, + "loss": 0.2318, + "step": 2046 + }, + { + "epoch": 1.5880527540729248, + "grad_norm": 6.448197841644287, + "learning_rate": 1.0364749991749793e-05, + "loss": 0.2851, + "step": 2047 + }, + { + "epoch": 1.5888285492629945, + "grad_norm": 12.016427993774414, + "learning_rate": 1.0341108320776456e-05, + "loss": 0.7802, + "step": 2048 + }, + { + "epoch": 1.5896043444530643, + "grad_norm": 4.0701775550842285, + "learning_rate": 1.0317496130559974e-05, + "loss": 0.0684, + "step": 2049 + }, + { + "epoch": 1.590380139643134, + "grad_norm": 11.426847457885742, + "learning_rate": 1.0293913499236363e-05, + "loss": 0.4721, + "step": 2050 + }, + { + "epoch": 1.591155934833204, + "grad_norm": 2.2228753566741943, + "learning_rate": 1.0270360504843813e-05, + "loss": 0.0649, + "step": 2051 + }, + { + "epoch": 1.5919317300232738, + "grad_norm": 8.332490921020508, + "learning_rate": 1.0246837225322456e-05, + "loss": 0.4631, + "step": 2052 + }, + { + "epoch": 1.5927075252133438, + "grad_norm": 4.954179763793945, + "learning_rate": 1.0223343738514071e-05, + "loss": 0.1145, + "step": 2053 + }, + { + "epoch": 1.5934833204034136, + "grad_norm": 3.3984718322753906, + "learning_rate": 1.0199880122161862e-05, + "loss": 0.0762, + "step": 2054 + }, + { + "epoch": 1.5942591155934833, + "grad_norm": 4.7868475914001465, + "learning_rate": 1.0176446453910197e-05, + "loss": 0.1014, + "step": 2055 + }, + { + "epoch": 1.595034910783553, + "grad_norm": 4.564414978027344, + "learning_rate": 1.0153042811304319e-05, + "loss": 0.2135, + "step": 2056 + }, + { + "epoch": 1.5958107059736228, + "grad_norm": 6.792155742645264, + "learning_rate": 1.0129669271790127e-05, + "loss": 0.1438, + "step": 2057 + }, + { + "epoch": 1.5965865011636928, + "grad_norm": 4.789403438568115, + "learning_rate": 1.0106325912713908e-05, + "loss": 0.0918, + "step": 2058 + }, + { + "epoch": 1.5973622963537626, + "grad_norm": 7.525842189788818, + "learning_rate": 1.0083012811322061e-05, + "loss": 0.4185, + "step": 2059 + }, + { + "epoch": 1.5981380915438326, + "grad_norm": 2.8247182369232178, + "learning_rate": 1.0059730044760875e-05, + "loss": 0.1144, + "step": 2060 + }, + { + "epoch": 1.5981380915438326, + "eval_Qnli-dev_cosine_accuracy": 0.6875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8128195405006409, + "eval_Qnli-dev_cosine_ap": 0.730752926208778, + "eval_Qnli-dev_cosine_f1": 0.6915584415584415, + "eval_Qnli-dev_cosine_f1_threshold": 0.7115497589111328, + "eval_Qnli-dev_cosine_precision": 0.5605263157894737, + "eval_Qnli-dev_cosine_recall": 0.902542372881356, + "eval_Qnli-dev_dot_accuracy": 0.65234375, + "eval_Qnli-dev_dot_accuracy_threshold": 384.8974304199219, + "eval_Qnli-dev_dot_ap": 0.6453750276683526, + "eval_Qnli-dev_dot_f1": 0.6770670826833073, + "eval_Qnli-dev_dot_f1_threshold": 325.13623046875, + "eval_Qnli-dev_dot_precision": 0.5358024691358024, + "eval_Qnli-dev_dot_recall": 0.9194915254237288, + "eval_Qnli-dev_euclidean_accuracy": 0.68359375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.189104080200195, + "eval_Qnli-dev_euclidean_ap": 0.7304772220720139, + "eval_Qnli-dev_euclidean_f1": 0.6952380952380951, + "eval_Qnli-dev_euclidean_f1_threshold": 17.300424575805664, + "eval_Qnli-dev_euclidean_precision": 0.5558375634517766, + "eval_Qnli-dev_euclidean_recall": 0.9279661016949152, + "eval_Qnli-dev_manhattan_accuracy": 0.68359375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 243.80421447753906, + "eval_Qnli-dev_manhattan_ap": 0.7309786644288385, + "eval_Qnli-dev_manhattan_f1": 0.6868044515103339, + "eval_Qnli-dev_manhattan_f1_threshold": 314.5264587402344, + "eval_Qnli-dev_manhattan_precision": 0.549618320610687, + "eval_Qnli-dev_manhattan_recall": 0.9152542372881356, + "eval_Qnli-dev_max_accuracy": 0.6875, + "eval_Qnli-dev_max_accuracy_threshold": 384.8974304199219, + "eval_Qnli-dev_max_ap": 0.7309786644288385, + "eval_Qnli-dev_max_f1": 0.6952380952380951, + "eval_Qnli-dev_max_f1_threshold": 325.13623046875, + "eval_Qnli-dev_max_precision": 0.5605263157894737, + "eval_Qnli-dev_max_recall": 0.9279661016949152, + "eval_allNLI-dev_cosine_accuracy": 0.73828125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8228088617324829, + "eval_allNLI-dev_cosine_ap": 0.6175473521913242, + "eval_allNLI-dev_cosine_f1": 0.6339285714285714, + "eval_allNLI-dev_cosine_f1_threshold": 0.7681463956832886, + "eval_allNLI-dev_cosine_precision": 0.5163636363636364, + "eval_allNLI-dev_cosine_recall": 0.8208092485549133, + "eval_allNLI-dev_dot_accuracy": 0.693359375, + "eval_allNLI-dev_dot_accuracy_threshold": 378.00604248046875, + "eval_allNLI-dev_dot_ap": 0.5286115321533219, + "eval_allNLI-dev_dot_f1": 0.5846153846153845, + "eval_allNLI-dev_dot_f1_threshold": 317.15313720703125, + "eval_allNLI-dev_dot_precision": 0.43804034582132567, + "eval_allNLI-dev_dot_recall": 0.8786127167630058, + "eval_allNLI-dev_euclidean_accuracy": 0.732421875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.722639083862305, + "eval_allNLI-dev_euclidean_ap": 0.6215217652075724, + "eval_allNLI-dev_euclidean_f1": 0.6433566433566433, + "eval_allNLI-dev_euclidean_f1_threshold": 14.145479202270508, + "eval_allNLI-dev_euclidean_precision": 0.5390625, + "eval_allNLI-dev_euclidean_recall": 0.7976878612716763, + "eval_allNLI-dev_manhattan_accuracy": 0.736328125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 227.79635620117188, + "eval_allNLI-dev_manhattan_ap": 0.6230646458490248, + "eval_allNLI-dev_manhattan_f1": 0.6394230769230769, + "eval_allNLI-dev_manhattan_f1_threshold": 256.0809020996094, + "eval_allNLI-dev_manhattan_precision": 0.5473251028806584, + "eval_allNLI-dev_manhattan_recall": 0.7687861271676301, + "eval_allNLI-dev_max_accuracy": 0.73828125, + "eval_allNLI-dev_max_accuracy_threshold": 378.00604248046875, + "eval_allNLI-dev_max_ap": 0.6230646458490248, + "eval_allNLI-dev_max_f1": 0.6433566433566433, + "eval_allNLI-dev_max_f1_threshold": 317.15313720703125, + "eval_allNLI-dev_max_precision": 0.5473251028806584, + "eval_allNLI-dev_max_recall": 0.8786127167630058, + "eval_sequential_score": 0.7309786644288385, + "eval_sts-test_pearson_cosine": 0.8329477278019732, + "eval_sts-test_pearson_dot": 0.7708599746693708, + "eval_sts-test_pearson_euclidean": 0.8683688885137731, + "eval_sts-test_pearson_manhattan": 0.8661182709296134, + "eval_sts-test_pearson_max": 0.8683688885137731, + "eval_sts-test_spearman_cosine": 0.8722174342010338, + "eval_sts-test_spearman_dot": 0.7610662649638382, + "eval_sts-test_spearman_euclidean": 0.8679841886297144, + "eval_sts-test_spearman_manhattan": 0.8678856911033919, + "eval_sts-test_spearman_max": 0.8722174342010338, + "eval_vitaminc-pairs_loss": 3.9375758171081543, + "eval_vitaminc-pairs_runtime": 5.5651, + "eval_vitaminc-pairs_samples_per_second": 23.001, + "eval_vitaminc-pairs_steps_per_second": 0.359, + "step": 2060 + }, + { + "epoch": 1.5981380915438326, + "eval_negation-triplets_loss": 1.7106106281280518, + "eval_negation-triplets_runtime": 1.1118, + "eval_negation-triplets_samples_per_second": 115.126, + "eval_negation-triplets_steps_per_second": 1.799, + "step": 2060 + }, + { + "epoch": 1.5981380915438326, + "eval_scitail-pairs-pos_loss": 0.11161491274833679, + "eval_scitail-pairs-pos_runtime": 1.3746, + "eval_scitail-pairs-pos_samples_per_second": 93.121, + "eval_scitail-pairs-pos_steps_per_second": 1.455, + "step": 2060 + }, + { + "epoch": 1.5981380915438326, + "eval_scitail-pairs-qa_loss": 0.004735540132969618, + "eval_scitail-pairs-qa_runtime": 0.8118, + "eval_scitail-pairs-qa_samples_per_second": 157.668, + "eval_scitail-pairs-qa_steps_per_second": 2.464, + "step": 2060 + }, + { + "epoch": 1.5981380915438326, + "eval_xsum-pairs_loss": 0.08274101465940475, + "eval_xsum-pairs_runtime": 3.9834, + "eval_xsum-pairs_samples_per_second": 32.133, + "eval_xsum-pairs_steps_per_second": 0.502, + "step": 2060 + }, + { + "epoch": 1.5981380915438326, + "eval_sciq_pairs_loss": 0.04313517361879349, + "eval_sciq_pairs_runtime": 6.1934, + "eval_sciq_pairs_samples_per_second": 20.667, + "eval_sciq_pairs_steps_per_second": 0.323, + "step": 2060 + }, + { + "epoch": 1.5981380915438326, + "eval_qasc_pairs_loss": 0.1808406114578247, + "eval_qasc_pairs_runtime": 0.9495, + "eval_qasc_pairs_samples_per_second": 134.809, + "eval_qasc_pairs_steps_per_second": 2.106, + "step": 2060 + }, + { + "epoch": 1.5981380915438326, + "eval_openbookqa_pairs_loss": 1.0177208185195923, + "eval_openbookqa_pairs_runtime": 0.8969, + "eval_openbookqa_pairs_samples_per_second": 142.709, + "eval_openbookqa_pairs_steps_per_second": 2.23, + "step": 2060 + }, + { + "epoch": 1.5981380915438326, + "eval_msmarco_pairs_loss": 0.31818193197250366, + "eval_msmarco_pairs_runtime": 2.0093, + "eval_msmarco_pairs_samples_per_second": 63.703, + "eval_msmarco_pairs_steps_per_second": 0.995, + "step": 2060 + }, + { + "epoch": 1.5981380915438326, + "eval_nq_pairs_loss": 0.50083327293396, + "eval_nq_pairs_runtime": 3.3773, + "eval_nq_pairs_samples_per_second": 37.901, + "eval_nq_pairs_steps_per_second": 0.592, + "step": 2060 + }, + { + "epoch": 1.5981380915438326, + "eval_trivia_pairs_loss": 0.687404990196228, + "eval_trivia_pairs_runtime": 4.5208, + "eval_trivia_pairs_samples_per_second": 28.314, + "eval_trivia_pairs_steps_per_second": 0.442, + "step": 2060 + }, + { + "epoch": 1.5981380915438326, + "eval_gooaq_pairs_loss": 0.5568317174911499, + "eval_gooaq_pairs_runtime": 1.4954, + "eval_gooaq_pairs_samples_per_second": 85.593, + "eval_gooaq_pairs_steps_per_second": 1.337, + "step": 2060 + }, + { + "epoch": 1.5981380915438326, + "eval_paws-pos_loss": 0.031026028096675873, + "eval_paws-pos_runtime": 1.0588, + "eval_paws-pos_samples_per_second": 120.893, + "eval_paws-pos_steps_per_second": 1.889, + "step": 2060 + }, + { + "epoch": 1.5981380915438326, + "eval_global_dataset_loss": 0.4095609784126282, + "eval_global_dataset_runtime": 19.522, + "eval_global_dataset_samples_per_second": 21.309, + "eval_global_dataset_steps_per_second": 0.359, + "step": 2060 + }, + { + "epoch": 1.5989138867339023, + "grad_norm": 7.268509387969971, + "learning_rate": 1.0036477690076246e-05, + "loss": 0.3928, + "step": 2061 + }, + { + "epoch": 1.599689681923972, + "grad_norm": 8.954904556274414, + "learning_rate": 1.0013255824213449e-05, + "loss": 0.4804, + "step": 2062 + }, + { + "epoch": 1.6004654771140419, + "grad_norm": 0.1638353168964386, + "learning_rate": 9.99006452401684e-06, + "loss": 0.0027, + "step": 2063 + }, + { + "epoch": 1.6012412723041116, + "grad_norm": 0.08627258986234665, + "learning_rate": 9.966903866229657e-06, + "loss": 0.0009, + "step": 2064 + }, + { + "epoch": 1.6020170674941814, + "grad_norm": 3.7455804347991943, + "learning_rate": 9.943773927493714e-06, + "loss": 0.1276, + "step": 2065 + }, + { + "epoch": 1.6027928626842514, + "grad_norm": 5.271412372589111, + "learning_rate": 9.920674784349187e-06, + "loss": 0.1217, + "step": 2066 + }, + { + "epoch": 1.6035686578743211, + "grad_norm": 7.600584506988525, + "learning_rate": 9.897606513234344e-06, + "loss": 0.284, + "step": 2067 + }, + { + "epoch": 1.604344453064391, + "grad_norm": 5.208740711212158, + "learning_rate": 9.874569190485293e-06, + "loss": 0.2112, + "step": 2068 + }, + { + "epoch": 1.6051202482544609, + "grad_norm": 2.9017889499664307, + "learning_rate": 9.851562892335717e-06, + "loss": 0.0506, + "step": 2069 + }, + { + "epoch": 1.6058960434445306, + "grad_norm": 0.2717197835445404, + "learning_rate": 9.828587694916656e-06, + "loss": 0.0036, + "step": 2070 + }, + { + "epoch": 1.6066718386346004, + "grad_norm": 0.5529122352600098, + "learning_rate": 9.805643674256208e-06, + "loss": 0.0172, + "step": 2071 + }, + { + "epoch": 1.6074476338246702, + "grad_norm": 7.0041279792785645, + "learning_rate": 9.782730906279321e-06, + "loss": 0.1879, + "step": 2072 + }, + { + "epoch": 1.6082234290147401, + "grad_norm": 3.984844923019409, + "learning_rate": 9.759849466807521e-06, + "loss": 0.1004, + "step": 2073 + }, + { + "epoch": 1.60899922420481, + "grad_norm": 4.56865119934082, + "learning_rate": 9.736999431558668e-06, + "loss": 0.0891, + "step": 2074 + }, + { + "epoch": 1.6097750193948799, + "grad_norm": 5.845067024230957, + "learning_rate": 9.71418087614668e-06, + "loss": 0.402, + "step": 2075 + }, + { + "epoch": 1.6105508145849496, + "grad_norm": 3.987060070037842, + "learning_rate": 9.691393876081328e-06, + "loss": 0.1241, + "step": 2076 + }, + { + "epoch": 1.6113266097750194, + "grad_norm": 8.923416137695312, + "learning_rate": 9.668638506767949e-06, + "loss": 0.42, + "step": 2077 + }, + { + "epoch": 1.6121024049650892, + "grad_norm": 0.4656422734260559, + "learning_rate": 9.645914843507219e-06, + "loss": 0.0065, + "step": 2078 + }, + { + "epoch": 1.612878200155159, + "grad_norm": 4.503904342651367, + "learning_rate": 9.623222961494874e-06, + "loss": 0.2506, + "step": 2079 + }, + { + "epoch": 1.613653995345229, + "grad_norm": 2.8720030784606934, + "learning_rate": 9.600562935821514e-06, + "loss": 0.0461, + "step": 2080 + }, + { + "epoch": 1.613653995345229, + "eval_Qnli-dev_cosine_accuracy": 0.685546875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8059393763542175, + "eval_Qnli-dev_cosine_ap": 0.7276140334259783, + "eval_Qnli-dev_cosine_f1": 0.6919104991394148, + "eval_Qnli-dev_cosine_f1_threshold": 0.7289411425590515, + "eval_Qnli-dev_cosine_precision": 0.5826086956521739, + "eval_Qnli-dev_cosine_recall": 0.8516949152542372, + "eval_Qnli-dev_dot_accuracy": 0.654296875, + "eval_Qnli-dev_dot_accuracy_threshold": 380.04339599609375, + "eval_Qnli-dev_dot_ap": 0.6434304112943701, + "eval_Qnli-dev_dot_f1": 0.6853582554517135, + "eval_Qnli-dev_dot_f1_threshold": 315.35467529296875, + "eval_Qnli-dev_dot_precision": 0.541871921182266, + "eval_Qnli-dev_dot_recall": 0.9322033898305084, + "eval_Qnli-dev_euclidean_accuracy": 0.68359375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.735767364501953, + "eval_Qnli-dev_euclidean_ap": 0.7284353266883888, + "eval_Qnli-dev_euclidean_f1": 0.6943999999999998, + "eval_Qnli-dev_euclidean_f1_threshold": 17.37596893310547, + "eval_Qnli-dev_euclidean_precision": 0.5578406169665809, + "eval_Qnli-dev_euclidean_recall": 0.9194915254237288, + "eval_Qnli-dev_manhattan_accuracy": 0.689453125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 248.78868103027344, + "eval_Qnli-dev_manhattan_ap": 0.7280532572340896, + "eval_Qnli-dev_manhattan_f1": 0.688, + "eval_Qnli-dev_manhattan_f1_threshold": 314.9635009765625, + "eval_Qnli-dev_manhattan_precision": 0.5526992287917738, + "eval_Qnli-dev_manhattan_recall": 0.9110169491525424, + "eval_Qnli-dev_max_accuracy": 0.689453125, + "eval_Qnli-dev_max_accuracy_threshold": 380.04339599609375, + "eval_Qnli-dev_max_ap": 0.7284353266883888, + "eval_Qnli-dev_max_f1": 0.6943999999999998, + "eval_Qnli-dev_max_f1_threshold": 315.35467529296875, + "eval_Qnli-dev_max_precision": 0.5826086956521739, + "eval_Qnli-dev_max_recall": 0.9322033898305084, + "eval_allNLI-dev_cosine_accuracy": 0.7421875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8305503129959106, + "eval_allNLI-dev_cosine_ap": 0.6173572729396761, + "eval_allNLI-dev_cosine_f1": 0.6292682926829267, + "eval_allNLI-dev_cosine_f1_threshold": 0.7874393463134766, + "eval_allNLI-dev_cosine_precision": 0.5443037974683544, + "eval_allNLI-dev_cosine_recall": 0.7456647398843931, + "eval_allNLI-dev_dot_accuracy": 0.708984375, + "eval_allNLI-dev_dot_accuracy_threshold": 371.9136962890625, + "eval_allNLI-dev_dot_ap": 0.53629060164935, + "eval_allNLI-dev_dot_f1": 0.5892116182572614, + "eval_allNLI-dev_dot_f1_threshold": 326.8874816894531, + "eval_allNLI-dev_dot_precision": 0.459546925566343, + "eval_allNLI-dev_dot_recall": 0.8208092485549133, + "eval_allNLI-dev_euclidean_accuracy": 0.736328125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 12.008539199829102, + "eval_allNLI-dev_euclidean_ap": 0.6210496140941586, + "eval_allNLI-dev_euclidean_f1": 0.639080459770115, + "eval_allNLI-dev_euclidean_f1_threshold": 14.25057601928711, + "eval_allNLI-dev_euclidean_precision": 0.5305343511450382, + "eval_allNLI-dev_euclidean_recall": 0.8034682080924855, + "eval_allNLI-dev_manhattan_accuracy": 0.7421875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 225.76556396484375, + "eval_allNLI-dev_manhattan_ap": 0.6217450189079348, + "eval_allNLI-dev_manhattan_f1": 0.6380090497737557, + "eval_allNLI-dev_manhattan_f1_threshold": 264.8858642578125, + "eval_allNLI-dev_manhattan_precision": 0.5241635687732342, + "eval_allNLI-dev_manhattan_recall": 0.815028901734104, + "eval_allNLI-dev_max_accuracy": 0.7421875, + "eval_allNLI-dev_max_accuracy_threshold": 371.9136962890625, + "eval_allNLI-dev_max_ap": 0.6217450189079348, + "eval_allNLI-dev_max_f1": 0.639080459770115, + "eval_allNLI-dev_max_f1_threshold": 326.8874816894531, + "eval_allNLI-dev_max_precision": 0.5443037974683544, + "eval_allNLI-dev_max_recall": 0.8208092485549133, + "eval_sequential_score": 0.7284353266883888, + "eval_sts-test_pearson_cosine": 0.832423140066678, + "eval_sts-test_pearson_dot": 0.7771130575493548, + "eval_sts-test_pearson_euclidean": 0.8690902050301916, + "eval_sts-test_pearson_manhattan": 0.8669669886507452, + "eval_sts-test_pearson_max": 0.8690902050301916, + "eval_sts-test_spearman_cosine": 0.8732609441141584, + "eval_sts-test_spearman_dot": 0.7687142107843671, + "eval_sts-test_spearman_euclidean": 0.8692794691394271, + "eval_sts-test_spearman_manhattan": 0.8693619009629099, + "eval_sts-test_spearman_max": 0.8732609441141584, + "eval_vitaminc-pairs_loss": 4.0197577476501465, + "eval_vitaminc-pairs_runtime": 5.6738, + "eval_vitaminc-pairs_samples_per_second": 22.56, + "eval_vitaminc-pairs_steps_per_second": 0.352, + "step": 2080 + }, + { + "epoch": 1.613653995345229, + "eval_negation-triplets_loss": 1.7220062017440796, + "eval_negation-triplets_runtime": 1.2743, + "eval_negation-triplets_samples_per_second": 100.449, + "eval_negation-triplets_steps_per_second": 1.57, + "step": 2080 + }, + { + "epoch": 1.613653995345229, + "eval_scitail-pairs-pos_loss": 0.10904178023338318, + "eval_scitail-pairs-pos_runtime": 1.4676, + "eval_scitail-pairs-pos_samples_per_second": 87.215, + "eval_scitail-pairs-pos_steps_per_second": 1.363, + "step": 2080 + }, + { + "epoch": 1.613653995345229, + "eval_scitail-pairs-qa_loss": 0.005222574342042208, + "eval_scitail-pairs-qa_runtime": 0.8214, + "eval_scitail-pairs-qa_samples_per_second": 155.831, + "eval_scitail-pairs-qa_steps_per_second": 2.435, + "step": 2080 + }, + { + "epoch": 1.613653995345229, + "eval_xsum-pairs_loss": 0.0951642394065857, + "eval_xsum-pairs_runtime": 3.9711, + "eval_xsum-pairs_samples_per_second": 32.233, + "eval_xsum-pairs_steps_per_second": 0.504, + "step": 2080 + }, + { + "epoch": 1.613653995345229, + "eval_sciq_pairs_loss": 0.043161485344171524, + "eval_sciq_pairs_runtime": 6.3166, + "eval_sciq_pairs_samples_per_second": 20.264, + "eval_sciq_pairs_steps_per_second": 0.317, + "step": 2080 + }, + { + "epoch": 1.613653995345229, + "eval_qasc_pairs_loss": 0.1928309053182602, + "eval_qasc_pairs_runtime": 0.9677, + "eval_qasc_pairs_samples_per_second": 132.27, + "eval_qasc_pairs_steps_per_second": 2.067, + "step": 2080 + }, + { + "epoch": 1.613653995345229, + "eval_openbookqa_pairs_loss": 1.04000723361969, + "eval_openbookqa_pairs_runtime": 0.9122, + "eval_openbookqa_pairs_samples_per_second": 140.321, + "eval_openbookqa_pairs_steps_per_second": 2.193, + "step": 2080 + }, + { + "epoch": 1.613653995345229, + "eval_msmarco_pairs_loss": 0.33264267444610596, + "eval_msmarco_pairs_runtime": 1.9959, + "eval_msmarco_pairs_samples_per_second": 64.131, + "eval_msmarco_pairs_steps_per_second": 1.002, + "step": 2080 + }, + { + "epoch": 1.613653995345229, + "eval_nq_pairs_loss": 0.5309322476387024, + "eval_nq_pairs_runtime": 3.3826, + "eval_nq_pairs_samples_per_second": 37.84, + "eval_nq_pairs_steps_per_second": 0.591, + "step": 2080 + }, + { + "epoch": 1.613653995345229, + "eval_trivia_pairs_loss": 0.7191007733345032, + "eval_trivia_pairs_runtime": 4.5283, + "eval_trivia_pairs_samples_per_second": 28.267, + "eval_trivia_pairs_steps_per_second": 0.442, + "step": 2080 + }, + { + "epoch": 1.613653995345229, + "eval_gooaq_pairs_loss": 0.5855757594108582, + "eval_gooaq_pairs_runtime": 1.5301, + "eval_gooaq_pairs_samples_per_second": 83.654, + "eval_gooaq_pairs_steps_per_second": 1.307, + "step": 2080 + }, + { + "epoch": 1.613653995345229, + "eval_paws-pos_loss": 0.030025482177734375, + "eval_paws-pos_runtime": 1.095, + "eval_paws-pos_samples_per_second": 116.898, + "eval_paws-pos_steps_per_second": 1.827, + "step": 2080 + }, + { + "epoch": 1.613653995345229, + "eval_global_dataset_loss": 0.42721396684646606, + "eval_global_dataset_runtime": 19.509, + "eval_global_dataset_samples_per_second": 21.323, + "eval_global_dataset_steps_per_second": 0.359, + "step": 2080 + }, + { + "epoch": 1.6144297905352987, + "grad_norm": 2.6430935859680176, + "learning_rate": 9.577934841472284e-06, + "loss": 0.1089, + "step": 2081 + }, + { + "epoch": 1.6152055857253687, + "grad_norm": 0.5106087923049927, + "learning_rate": 9.555338753326692e-06, + "loss": 0.0273, + "step": 2082 + }, + { + "epoch": 1.6159813809154384, + "grad_norm": 5.704498767852783, + "learning_rate": 9.532774746158332e-06, + "loss": 0.2, + "step": 2083 + }, + { + "epoch": 1.6167571761055082, + "grad_norm": 2.7096312046051025, + "learning_rate": 9.51024289463462e-06, + "loss": 0.0443, + "step": 2084 + }, + { + "epoch": 1.617532971295578, + "grad_norm": 5.4050445556640625, + "learning_rate": 9.487743273316578e-06, + "loss": 0.0916, + "step": 2085 + }, + { + "epoch": 1.6183087664856477, + "grad_norm": 13.555615425109863, + "learning_rate": 9.465275956658571e-06, + "loss": 0.7323, + "step": 2086 + }, + { + "epoch": 1.6190845616757175, + "grad_norm": 8.599193572998047, + "learning_rate": 9.442841019008056e-06, + "loss": 0.34, + "step": 2087 + }, + { + "epoch": 1.6198603568657874, + "grad_norm": 6.1068525314331055, + "learning_rate": 9.420438534605356e-06, + "loss": 0.1962, + "step": 2088 + }, + { + "epoch": 1.6206361520558572, + "grad_norm": 6.415519714355469, + "learning_rate": 9.398068577583396e-06, + "loss": 0.2534, + "step": 2089 + }, + { + "epoch": 1.6214119472459272, + "grad_norm": 3.4375174045562744, + "learning_rate": 9.375731221967464e-06, + "loss": 0.0798, + "step": 2090 + }, + { + "epoch": 1.622187742435997, + "grad_norm": 0.5146510004997253, + "learning_rate": 9.353426541674958e-06, + "loss": 0.0055, + "step": 2091 + }, + { + "epoch": 1.6229635376260667, + "grad_norm": 4.018589496612549, + "learning_rate": 9.33115461051516e-06, + "loss": 0.0617, + "step": 2092 + }, + { + "epoch": 1.6237393328161365, + "grad_norm": 2.867698907852173, + "learning_rate": 9.308915502188974e-06, + "loss": 0.0402, + "step": 2093 + }, + { + "epoch": 1.6245151280062062, + "grad_norm": 6.602728843688965, + "learning_rate": 9.2867092902887e-06, + "loss": 0.2549, + "step": 2094 + }, + { + "epoch": 1.6252909231962762, + "grad_norm": 0.38559412956237793, + "learning_rate": 9.264536048297754e-06, + "loss": 0.0059, + "step": 2095 + }, + { + "epoch": 1.626066718386346, + "grad_norm": 4.081066608428955, + "learning_rate": 9.24239584959048e-06, + "loss": 0.0971, + "step": 2096 + }, + { + "epoch": 1.626842513576416, + "grad_norm": 2.7574284076690674, + "learning_rate": 9.22028876743185e-06, + "loss": 0.0448, + "step": 2097 + }, + { + "epoch": 1.6276183087664857, + "grad_norm": 6.365083694458008, + "learning_rate": 9.198214874977267e-06, + "loss": 0.2226, + "step": 2098 + }, + { + "epoch": 1.6283941039565555, + "grad_norm": 10.855636596679688, + "learning_rate": 9.176174245272302e-06, + "loss": 0.6508, + "step": 2099 + }, + { + "epoch": 1.6291698991466252, + "grad_norm": 9.044904708862305, + "learning_rate": 9.154166951252459e-06, + "loss": 0.5164, + "step": 2100 + }, + { + "epoch": 1.6291698991466252, + "eval_Qnli-dev_cosine_accuracy": 0.6796875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8041599988937378, + "eval_Qnli-dev_cosine_ap": 0.7158126906566802, + "eval_Qnli-dev_cosine_f1": 0.6853582554517135, + "eval_Qnli-dev_cosine_f1_threshold": 0.6541811227798462, + "eval_Qnli-dev_cosine_precision": 0.541871921182266, + "eval_Qnli-dev_cosine_recall": 0.9322033898305084, + "eval_Qnli-dev_dot_accuracy": 0.65234375, + "eval_Qnli-dev_dot_accuracy_threshold": 361.11041259765625, + "eval_Qnli-dev_dot_ap": 0.6402432856196079, + "eval_Qnli-dev_dot_f1": 0.6852713178294574, + "eval_Qnli-dev_dot_f1_threshold": 302.9774475097656, + "eval_Qnli-dev_dot_precision": 0.5403422982885085, + "eval_Qnli-dev_dot_recall": 0.9364406779661016, + "eval_Qnli-dev_euclidean_accuracy": 0.6796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.470650672912598, + "eval_Qnli-dev_euclidean_ap": 0.717528592823043, + "eval_Qnli-dev_euclidean_f1": 0.692179700499168, + "eval_Qnli-dev_euclidean_f1_threshold": 17.064395904541016, + "eval_Qnli-dev_euclidean_precision": 0.5698630136986301, + "eval_Qnli-dev_euclidean_recall": 0.8813559322033898, + "eval_Qnli-dev_manhattan_accuracy": 0.6796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 257.53155517578125, + "eval_Qnli-dev_manhattan_ap": 0.7179147660989839, + "eval_Qnli-dev_manhattan_f1": 0.6840390879478827, + "eval_Qnli-dev_manhattan_f1_threshold": 316.1563720703125, + "eval_Qnli-dev_manhattan_precision": 0.5555555555555556, + "eval_Qnli-dev_manhattan_recall": 0.8898305084745762, + "eval_Qnli-dev_max_accuracy": 0.6796875, + "eval_Qnli-dev_max_accuracy_threshold": 361.11041259765625, + "eval_Qnli-dev_max_ap": 0.7179147660989839, + "eval_Qnli-dev_max_f1": 0.692179700499168, + "eval_Qnli-dev_max_f1_threshold": 316.1563720703125, + "eval_Qnli-dev_max_precision": 0.5698630136986301, + "eval_Qnli-dev_max_recall": 0.9364406779661016, + "eval_allNLI-dev_cosine_accuracy": 0.73828125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8334616422653198, + "eval_allNLI-dev_cosine_ap": 0.6132635354612769, + "eval_allNLI-dev_cosine_f1": 0.632016632016632, + "eval_allNLI-dev_cosine_f1_threshold": 0.7310935258865356, + "eval_allNLI-dev_cosine_precision": 0.4935064935064935, + "eval_allNLI-dev_cosine_recall": 0.8786127167630058, + "eval_allNLI-dev_dot_accuracy": 0.70703125, + "eval_allNLI-dev_dot_accuracy_threshold": 365.2249450683594, + "eval_allNLI-dev_dot_ap": 0.5390503358631051, + "eval_allNLI-dev_dot_f1": 0.5877712031558185, + "eval_allNLI-dev_dot_f1_threshold": 304.40081787109375, + "eval_allNLI-dev_dot_precision": 0.44610778443113774, + "eval_allNLI-dev_dot_recall": 0.861271676300578, + "eval_allNLI-dev_euclidean_accuracy": 0.732421875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 12.50822639465332, + "eval_allNLI-dev_euclidean_ap": 0.6155864057262482, + "eval_allNLI-dev_euclidean_f1": 0.6371681415929203, + "eval_allNLI-dev_euclidean_f1_threshold": 14.766093254089355, + "eval_allNLI-dev_euclidean_precision": 0.5161290322580645, + "eval_allNLI-dev_euclidean_recall": 0.8323699421965318, + "eval_allNLI-dev_manhattan_accuracy": 0.740234375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 228.49171447753906, + "eval_allNLI-dev_manhattan_ap": 0.6175557289371132, + "eval_allNLI-dev_manhattan_f1": 0.6341463414634148, + "eval_allNLI-dev_manhattan_f1_threshold": 235.85476684570312, + "eval_allNLI-dev_manhattan_precision": 0.5969387755102041, + "eval_allNLI-dev_manhattan_recall": 0.6763005780346821, + "eval_allNLI-dev_max_accuracy": 0.740234375, + "eval_allNLI-dev_max_accuracy_threshold": 365.2249450683594, + "eval_allNLI-dev_max_ap": 0.6175557289371132, + "eval_allNLI-dev_max_f1": 0.6371681415929203, + "eval_allNLI-dev_max_f1_threshold": 304.40081787109375, + "eval_allNLI-dev_max_precision": 0.5969387755102041, + "eval_allNLI-dev_max_recall": 0.8786127167630058, + "eval_sequential_score": 0.7179147660989839, + "eval_sts-test_pearson_cosine": 0.83206016171404, + "eval_sts-test_pearson_dot": 0.7958776356396632, + "eval_sts-test_pearson_euclidean": 0.8694039751901117, + "eval_sts-test_pearson_manhattan": 0.8671245471058429, + "eval_sts-test_pearson_max": 0.8694039751901117, + "eval_sts-test_spearman_cosine": 0.8750202504532385, + "eval_sts-test_spearman_dot": 0.7886669634370751, + "eval_sts-test_spearman_euclidean": 0.8714036325126312, + "eval_sts-test_spearman_manhattan": 0.8706156523020514, + "eval_sts-test_spearman_max": 0.8750202504532385, + "eval_vitaminc-pairs_loss": 4.294209003448486, + "eval_vitaminc-pairs_runtime": 5.5424, + "eval_vitaminc-pairs_samples_per_second": 23.095, + "eval_vitaminc-pairs_steps_per_second": 0.361, + "step": 2100 + }, + { + "epoch": 1.6291698991466252, + "eval_negation-triplets_loss": 1.72343111038208, + "eval_negation-triplets_runtime": 1.1121, + "eval_negation-triplets_samples_per_second": 115.102, + "eval_negation-triplets_steps_per_second": 1.798, + "step": 2100 + }, + { + "epoch": 1.6291698991466252, + "eval_scitail-pairs-pos_loss": 0.10624809563159943, + "eval_scitail-pairs-pos_runtime": 1.3311, + "eval_scitail-pairs-pos_samples_per_second": 96.161, + "eval_scitail-pairs-pos_steps_per_second": 1.503, + "step": 2100 + }, + { + "epoch": 1.6291698991466252, + "eval_scitail-pairs-qa_loss": 0.005223955027759075, + "eval_scitail-pairs-qa_runtime": 0.7925, + "eval_scitail-pairs-qa_samples_per_second": 161.518, + "eval_scitail-pairs-qa_steps_per_second": 2.524, + "step": 2100 + }, + { + "epoch": 1.6291698991466252, + "eval_xsum-pairs_loss": 0.06834965944290161, + "eval_xsum-pairs_runtime": 3.9615, + "eval_xsum-pairs_samples_per_second": 32.311, + "eval_xsum-pairs_steps_per_second": 0.505, + "step": 2100 + }, + { + "epoch": 1.6291698991466252, + "eval_sciq_pairs_loss": 0.041589293628931046, + "eval_sciq_pairs_runtime": 6.2757, + "eval_sciq_pairs_samples_per_second": 20.396, + "eval_sciq_pairs_steps_per_second": 0.319, + "step": 2100 + }, + { + "epoch": 1.6291698991466252, + "eval_qasc_pairs_loss": 0.19304241240024567, + "eval_qasc_pairs_runtime": 0.9609, + "eval_qasc_pairs_samples_per_second": 133.208, + "eval_qasc_pairs_steps_per_second": 2.081, + "step": 2100 + }, + { + "epoch": 1.6291698991466252, + "eval_openbookqa_pairs_loss": 1.0549601316452026, + "eval_openbookqa_pairs_runtime": 0.9041, + "eval_openbookqa_pairs_samples_per_second": 141.584, + "eval_openbookqa_pairs_steps_per_second": 2.212, + "step": 2100 + }, + { + "epoch": 1.6291698991466252, + "eval_msmarco_pairs_loss": 0.34974053502082825, + "eval_msmarco_pairs_runtime": 1.9889, + "eval_msmarco_pairs_samples_per_second": 64.356, + "eval_msmarco_pairs_steps_per_second": 1.006, + "step": 2100 + }, + { + "epoch": 1.6291698991466252, + "eval_nq_pairs_loss": 0.5477519035339355, + "eval_nq_pairs_runtime": 3.376, + "eval_nq_pairs_samples_per_second": 37.915, + "eval_nq_pairs_steps_per_second": 0.592, + "step": 2100 + }, + { + "epoch": 1.6291698991466252, + "eval_trivia_pairs_loss": 0.6962348818778992, + "eval_trivia_pairs_runtime": 4.5332, + "eval_trivia_pairs_samples_per_second": 28.236, + "eval_trivia_pairs_steps_per_second": 0.441, + "step": 2100 + }, + { + "epoch": 1.6291698991466252, + "eval_gooaq_pairs_loss": 0.5942674875259399, + "eval_gooaq_pairs_runtime": 1.4946, + "eval_gooaq_pairs_samples_per_second": 85.643, + "eval_gooaq_pairs_steps_per_second": 1.338, + "step": 2100 + }, + { + "epoch": 1.6291698991466252, + "eval_paws-pos_loss": 0.030590102076530457, + "eval_paws-pos_runtime": 1.0562, + "eval_paws-pos_samples_per_second": 121.193, + "eval_paws-pos_steps_per_second": 1.894, + "step": 2100 + }, + { + "epoch": 1.6291698991466252, + "eval_global_dataset_loss": 0.4558424651622772, + "eval_global_dataset_runtime": 19.51, + "eval_global_dataset_samples_per_second": 21.322, + "eval_global_dataset_steps_per_second": 0.359, + "step": 2100 + }, + { + "epoch": 1.629945694336695, + "grad_norm": 2.6818180084228516, + "learning_rate": 9.132193065742909e-06, + "loss": 0.0615, + "step": 2101 + }, + { + "epoch": 1.630721489526765, + "grad_norm": 5.820191860198975, + "learning_rate": 9.110252661458291e-06, + "loss": 0.1838, + "step": 2102 + }, + { + "epoch": 1.6314972847168348, + "grad_norm": 0.4047784209251404, + "learning_rate": 9.088345811002436e-06, + "loss": 0.0246, + "step": 2103 + }, + { + "epoch": 1.6322730799069047, + "grad_norm": 9.811863899230957, + "learning_rate": 9.066472586868155e-06, + "loss": 0.3865, + "step": 2104 + }, + { + "epoch": 1.6330488750969745, + "grad_norm": 5.504253387451172, + "learning_rate": 9.044633061436973e-06, + "loss": 0.1233, + "step": 2105 + }, + { + "epoch": 1.6338246702870443, + "grad_norm": 12.337666511535645, + "learning_rate": 9.022827306978913e-06, + "loss": 0.6508, + "step": 2106 + }, + { + "epoch": 1.634600465477114, + "grad_norm": 6.9936842918396, + "learning_rate": 9.001055395652223e-06, + "loss": 0.1268, + "step": 2107 + }, + { + "epoch": 1.6353762606671838, + "grad_norm": 4.6181464195251465, + "learning_rate": 8.979317399503192e-06, + "loss": 0.1091, + "step": 2108 + }, + { + "epoch": 1.6361520558572535, + "grad_norm": 3.334299325942993, + "learning_rate": 8.957613390465837e-06, + "loss": 0.0771, + "step": 2109 + }, + { + "epoch": 1.6369278510473235, + "grad_norm": 2.442931652069092, + "learning_rate": 8.93594344036176e-06, + "loss": 0.0359, + "step": 2110 + }, + { + "epoch": 1.6377036462373933, + "grad_norm": 0.5620980858802795, + "learning_rate": 8.914307620899805e-06, + "loss": 0.0077, + "step": 2111 + }, + { + "epoch": 1.6384794414274633, + "grad_norm": 4.801235198974609, + "learning_rate": 8.892706003675911e-06, + "loss": 0.1159, + "step": 2112 + }, + { + "epoch": 1.639255236617533, + "grad_norm": 2.9653725624084473, + "learning_rate": 8.871138660172811e-06, + "loss": 0.0382, + "step": 2113 + }, + { + "epoch": 1.6400310318076028, + "grad_norm": 5.701256275177002, + "learning_rate": 8.849605661759838e-06, + "loss": 0.1688, + "step": 2114 + }, + { + "epoch": 1.6408068269976726, + "grad_norm": 9.519471168518066, + "learning_rate": 8.828107079692665e-06, + "loss": 0.5371, + "step": 2115 + }, + { + "epoch": 1.6415826221877423, + "grad_norm": 0.01797030307352543, + "learning_rate": 8.806642985113085e-06, + "loss": 0.0003, + "step": 2116 + }, + { + "epoch": 1.6423584173778123, + "grad_norm": 3.0608532428741455, + "learning_rate": 8.785213449048753e-06, + "loss": 0.0402, + "step": 2117 + }, + { + "epoch": 1.643134212567882, + "grad_norm": 2.7267723083496094, + "learning_rate": 8.763818542412979e-06, + "loss": 0.0278, + "step": 2118 + }, + { + "epoch": 1.643910007757952, + "grad_norm": 4.9376091957092285, + "learning_rate": 8.742458336004465e-06, + "loss": 0.0849, + "step": 2119 + }, + { + "epoch": 1.6446858029480218, + "grad_norm": 7.8673014640808105, + "learning_rate": 8.7211329005071e-06, + "loss": 0.3249, + "step": 2120 + }, + { + "epoch": 1.6446858029480218, + "eval_Qnli-dev_cosine_accuracy": 0.681640625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8021085262298584, + "eval_Qnli-dev_cosine_ap": 0.7163871787998357, + "eval_Qnli-dev_cosine_f1": 0.6897689768976898, + "eval_Qnli-dev_cosine_f1_threshold": 0.7058418989181519, + "eval_Qnli-dev_cosine_precision": 0.5648648648648649, + "eval_Qnli-dev_cosine_recall": 0.885593220338983, + "eval_Qnli-dev_dot_accuracy": 0.65625, + "eval_Qnli-dev_dot_accuracy_threshold": 390.9899597167969, + "eval_Qnli-dev_dot_ap": 0.6359672970538122, + "eval_Qnli-dev_dot_f1": 0.6791277258566978, + "eval_Qnli-dev_dot_f1_threshold": 317.3963623046875, + "eval_Qnli-dev_dot_precision": 0.5369458128078818, + "eval_Qnli-dev_dot_recall": 0.923728813559322, + "eval_Qnli-dev_euclidean_accuracy": 0.6796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 14.212274551391602, + "eval_Qnli-dev_euclidean_ap": 0.7169622800892286, + "eval_Qnli-dev_euclidean_f1": 0.6856187290969901, + "eval_Qnli-dev_euclidean_f1_threshold": 16.67272186279297, + "eval_Qnli-dev_euclidean_precision": 0.5662983425414365, + "eval_Qnli-dev_euclidean_recall": 0.8686440677966102, + "eval_Qnli-dev_manhattan_accuracy": 0.673828125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 253.67626953125, + "eval_Qnli-dev_manhattan_ap": 0.7176221436747259, + "eval_Qnli-dev_manhattan_f1": 0.6847133757961782, + "eval_Qnli-dev_manhattan_f1_threshold": 320.0645751953125, + "eval_Qnli-dev_manhattan_precision": 0.548469387755102, + "eval_Qnli-dev_manhattan_recall": 0.9110169491525424, + "eval_Qnli-dev_max_accuracy": 0.681640625, + "eval_Qnli-dev_max_accuracy_threshold": 390.9899597167969, + "eval_Qnli-dev_max_ap": 0.7176221436747259, + "eval_Qnli-dev_max_f1": 0.6897689768976898, + "eval_Qnli-dev_max_f1_threshold": 320.0645751953125, + "eval_Qnli-dev_max_precision": 0.5662983425414365, + "eval_Qnli-dev_max_recall": 0.923728813559322, + "eval_allNLI-dev_cosine_accuracy": 0.732421875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8434829711914062, + "eval_allNLI-dev_cosine_ap": 0.6113789443383647, + "eval_allNLI-dev_cosine_f1": 0.6266094420600858, + "eval_allNLI-dev_cosine_f1_threshold": 0.7573399543762207, + "eval_allNLI-dev_cosine_precision": 0.49829351535836175, + "eval_allNLI-dev_cosine_recall": 0.8439306358381503, + "eval_allNLI-dev_dot_accuracy": 0.712890625, + "eval_allNLI-dev_dot_accuracy_threshold": 372.7173767089844, + "eval_allNLI-dev_dot_ap": 0.5349878734021147, + "eval_allNLI-dev_dot_f1": 0.5907335907335908, + "eval_allNLI-dev_dot_f1_threshold": 310.47711181640625, + "eval_allNLI-dev_dot_precision": 0.4434782608695652, + "eval_allNLI-dev_dot_recall": 0.884393063583815, + "eval_allNLI-dev_euclidean_accuracy": 0.732421875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.85063648223877, + "eval_allNLI-dev_euclidean_ap": 0.6152313469489301, + "eval_allNLI-dev_euclidean_f1": 0.6338028169014085, + "eval_allNLI-dev_euclidean_f1_threshold": 14.019867897033691, + "eval_allNLI-dev_euclidean_precision": 0.5335968379446641, + "eval_allNLI-dev_euclidean_recall": 0.7803468208092486, + "eval_allNLI-dev_manhattan_accuracy": 0.73828125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 232.6871337890625, + "eval_allNLI-dev_manhattan_ap": 0.6172880842753407, + "eval_allNLI-dev_manhattan_f1": 0.631868131868132, + "eval_allNLI-dev_manhattan_f1_threshold": 232.6871337890625, + "eval_allNLI-dev_manhattan_precision": 0.6020942408376964, + "eval_allNLI-dev_manhattan_recall": 0.6647398843930635, + "eval_allNLI-dev_max_accuracy": 0.73828125, + "eval_allNLI-dev_max_accuracy_threshold": 372.7173767089844, + "eval_allNLI-dev_max_ap": 0.6172880842753407, + "eval_allNLI-dev_max_f1": 0.6338028169014085, + "eval_allNLI-dev_max_f1_threshold": 310.47711181640625, + "eval_allNLI-dev_max_precision": 0.6020942408376964, + "eval_allNLI-dev_max_recall": 0.884393063583815, + "eval_sequential_score": 0.7176221436747259, + "eval_sts-test_pearson_cosine": 0.8279197237273547, + "eval_sts-test_pearson_dot": 0.7839873573596241, + "eval_sts-test_pearson_euclidean": 0.8659697879304455, + "eval_sts-test_pearson_manhattan": 0.863585773239862, + "eval_sts-test_pearson_max": 0.8659697879304455, + "eval_sts-test_spearman_cosine": 0.871735418364587, + "eval_sts-test_spearman_dot": 0.774668858376368, + "eval_sts-test_spearman_euclidean": 0.8694058467015964, + "eval_sts-test_spearman_manhattan": 0.8679773416867215, + "eval_sts-test_spearman_max": 0.871735418364587, + "eval_vitaminc-pairs_loss": 4.215254306793213, + "eval_vitaminc-pairs_runtime": 5.5348, + "eval_vitaminc-pairs_samples_per_second": 23.126, + "eval_vitaminc-pairs_steps_per_second": 0.361, + "step": 2120 + }, + { + "epoch": 1.6446858029480218, + "eval_negation-triplets_loss": 1.6907252073287964, + "eval_negation-triplets_runtime": 1.1142, + "eval_negation-triplets_samples_per_second": 114.886, + "eval_negation-triplets_steps_per_second": 1.795, + "step": 2120 + }, + { + "epoch": 1.6446858029480218, + "eval_scitail-pairs-pos_loss": 0.09465485066175461, + "eval_scitail-pairs-pos_runtime": 1.4238, + "eval_scitail-pairs-pos_samples_per_second": 89.903, + "eval_scitail-pairs-pos_steps_per_second": 1.405, + "step": 2120 + }, + { + "epoch": 1.6446858029480218, + "eval_scitail-pairs-qa_loss": 0.004804396536201239, + "eval_scitail-pairs-qa_runtime": 0.8329, + "eval_scitail-pairs-qa_samples_per_second": 153.673, + "eval_scitail-pairs-qa_steps_per_second": 2.401, + "step": 2120 + }, + { + "epoch": 1.6446858029480218, + "eval_xsum-pairs_loss": 0.06843353062868118, + "eval_xsum-pairs_runtime": 3.962, + "eval_xsum-pairs_samples_per_second": 32.307, + "eval_xsum-pairs_steps_per_second": 0.505, + "step": 2120 + }, + { + "epoch": 1.6446858029480218, + "eval_sciq_pairs_loss": 0.04222538694739342, + "eval_sciq_pairs_runtime": 6.1306, + "eval_sciq_pairs_samples_per_second": 20.879, + "eval_sciq_pairs_steps_per_second": 0.326, + "step": 2120 + }, + { + "epoch": 1.6446858029480218, + "eval_qasc_pairs_loss": 0.20571990311145782, + "eval_qasc_pairs_runtime": 0.9351, + "eval_qasc_pairs_samples_per_second": 136.879, + "eval_qasc_pairs_steps_per_second": 2.139, + "step": 2120 + }, + { + "epoch": 1.6446858029480218, + "eval_openbookqa_pairs_loss": 1.0152397155761719, + "eval_openbookqa_pairs_runtime": 0.8882, + "eval_openbookqa_pairs_samples_per_second": 144.105, + "eval_openbookqa_pairs_steps_per_second": 2.252, + "step": 2120 + }, + { + "epoch": 1.6446858029480218, + "eval_msmarco_pairs_loss": 0.35138875246047974, + "eval_msmarco_pairs_runtime": 1.9871, + "eval_msmarco_pairs_samples_per_second": 64.416, + "eval_msmarco_pairs_steps_per_second": 1.006, + "step": 2120 + }, + { + "epoch": 1.6446858029480218, + "eval_nq_pairs_loss": 0.5176401734352112, + "eval_nq_pairs_runtime": 3.3649, + "eval_nq_pairs_samples_per_second": 38.039, + "eval_nq_pairs_steps_per_second": 0.594, + "step": 2120 + }, + { + "epoch": 1.6446858029480218, + "eval_trivia_pairs_loss": 0.6926418542861938, + "eval_trivia_pairs_runtime": 4.5245, + "eval_trivia_pairs_samples_per_second": 28.291, + "eval_trivia_pairs_steps_per_second": 0.442, + "step": 2120 + }, + { + "epoch": 1.6446858029480218, + "eval_gooaq_pairs_loss": 0.5924417972564697, + "eval_gooaq_pairs_runtime": 1.4805, + "eval_gooaq_pairs_samples_per_second": 86.459, + "eval_gooaq_pairs_steps_per_second": 1.351, + "step": 2120 + }, + { + "epoch": 1.6446858029480218, + "eval_paws-pos_loss": 0.031221268698573112, + "eval_paws-pos_runtime": 1.0505, + "eval_paws-pos_samples_per_second": 121.851, + "eval_paws-pos_steps_per_second": 1.904, + "step": 2120 + }, + { + "epoch": 1.6446858029480218, + "eval_global_dataset_loss": 0.4482298791408539, + "eval_global_dataset_runtime": 19.4865, + "eval_global_dataset_samples_per_second": 21.348, + "eval_global_dataset_steps_per_second": 0.359, + "step": 2120 + }, + { + "epoch": 1.6454615981380916, + "grad_norm": 10.168702125549316, + "learning_rate": 8.699842306489705e-06, + "loss": 0.3991, + "step": 2121 + }, + { + "epoch": 1.6462373933281613, + "grad_norm": 5.346624851226807, + "learning_rate": 8.678586624405804e-06, + "loss": 0.0819, + "step": 2122 + }, + { + "epoch": 1.647013188518231, + "grad_norm": 2.552154302597046, + "learning_rate": 8.657365924593391e-06, + "loss": 0.0421, + "step": 2123 + }, + { + "epoch": 1.6477889837083008, + "grad_norm": 0.41933292150497437, + "learning_rate": 8.636180277274708e-06, + "loss": 0.0079, + "step": 2124 + }, + { + "epoch": 1.6485647788983708, + "grad_norm": 2.4424939155578613, + "learning_rate": 8.615029752555982e-06, + "loss": 0.0442, + "step": 2125 + }, + { + "epoch": 1.6493405740884408, + "grad_norm": 4.427201271057129, + "learning_rate": 8.593914420427254e-06, + "loss": 0.0872, + "step": 2126 + }, + { + "epoch": 1.6501163692785106, + "grad_norm": 4.333737850189209, + "learning_rate": 8.572834350762063e-06, + "loss": 0.0816, + "step": 2127 + }, + { + "epoch": 1.6508921644685803, + "grad_norm": 8.300909042358398, + "learning_rate": 8.551789613317292e-06, + "loss": 0.3955, + "step": 2128 + }, + { + "epoch": 1.65166795965865, + "grad_norm": 11.166008949279785, + "learning_rate": 8.530780277732884e-06, + "loss": 0.6127, + "step": 2129 + }, + { + "epoch": 1.6524437548487199, + "grad_norm": 0.663284957408905, + "learning_rate": 8.50980641353165e-06, + "loss": 0.0457, + "step": 2130 + }, + { + "epoch": 1.6532195500387896, + "grad_norm": 9.437458992004395, + "learning_rate": 8.488868090119e-06, + "loss": 0.6146, + "step": 2131 + }, + { + "epoch": 1.6539953452288596, + "grad_norm": 8.388361930847168, + "learning_rate": 8.467965376782764e-06, + "loss": 0.2562, + "step": 2132 + }, + { + "epoch": 1.6547711404189294, + "grad_norm": 9.942096710205078, + "learning_rate": 8.447098342692902e-06, + "loss": 0.3282, + "step": 2133 + }, + { + "epoch": 1.6555469356089993, + "grad_norm": 13.442931175231934, + "learning_rate": 8.426267056901331e-06, + "loss": 1.5108, + "step": 2134 + }, + { + "epoch": 1.656322730799069, + "grad_norm": 9.218644142150879, + "learning_rate": 8.405471588341652e-06, + "loss": 0.6743, + "step": 2135 + }, + { + "epoch": 1.6570985259891389, + "grad_norm": 4.336494445800781, + "learning_rate": 8.38471200582896e-06, + "loss": 0.1116, + "step": 2136 + }, + { + "epoch": 1.6578743211792086, + "grad_norm": 6.513510704040527, + "learning_rate": 8.363988378059582e-06, + "loss": 0.334, + "step": 2137 + }, + { + "epoch": 1.6586501163692784, + "grad_norm": 7.969923496246338, + "learning_rate": 8.343300773610885e-06, + "loss": 0.4487, + "step": 2138 + }, + { + "epoch": 1.6594259115593484, + "grad_norm": 5.289208889007568, + "learning_rate": 8.322649260941006e-06, + "loss": 0.1806, + "step": 2139 + }, + { + "epoch": 1.6602017067494181, + "grad_norm": 7.457812309265137, + "learning_rate": 8.302033908388671e-06, + "loss": 0.3436, + "step": 2140 + }, + { + "epoch": 1.6602017067494181, + "eval_Qnli-dev_cosine_accuracy": 0.6796875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8082764148712158, + "eval_Qnli-dev_cosine_ap": 0.716349847036448, + "eval_Qnli-dev_cosine_f1": 0.687211093990755, + "eval_Qnli-dev_cosine_f1_threshold": 0.6642000675201416, + "eval_Qnli-dev_cosine_precision": 0.5399515738498789, + "eval_Qnli-dev_cosine_recall": 0.9449152542372882, + "eval_Qnli-dev_dot_accuracy": 0.662109375, + "eval_Qnli-dev_dot_accuracy_threshold": 403.91180419921875, + "eval_Qnli-dev_dot_ap": 0.6394812110894501, + "eval_Qnli-dev_dot_f1": 0.6782884310618067, + "eval_Qnli-dev_dot_f1_threshold": 324.0391540527344, + "eval_Qnli-dev_dot_precision": 0.5417721518987342, + "eval_Qnli-dev_dot_recall": 0.9067796610169492, + "eval_Qnli-dev_euclidean_accuracy": 0.681640625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 14.026934623718262, + "eval_Qnli-dev_euclidean_ap": 0.7162481221298695, + "eval_Qnli-dev_euclidean_f1": 0.6923076923076923, + "eval_Qnli-dev_euclidean_f1_threshold": 17.261743545532227, + "eval_Qnli-dev_euclidean_precision": 0.5567010309278351, + "eval_Qnli-dev_euclidean_recall": 0.9152542372881356, + "eval_Qnli-dev_manhattan_accuracy": 0.673828125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 245.589599609375, + "eval_Qnli-dev_manhattan_ap": 0.7181295462432675, + "eval_Qnli-dev_manhattan_f1": 0.6892109500805152, + "eval_Qnli-dev_manhattan_f1_threshold": 313.6116638183594, + "eval_Qnli-dev_manhattan_precision": 0.5558441558441558, + "eval_Qnli-dev_manhattan_recall": 0.9067796610169492, + "eval_Qnli-dev_max_accuracy": 0.681640625, + "eval_Qnli-dev_max_accuracy_threshold": 403.91180419921875, + "eval_Qnli-dev_max_ap": 0.7181295462432675, + "eval_Qnli-dev_max_f1": 0.6923076923076923, + "eval_Qnli-dev_max_f1_threshold": 324.0391540527344, + "eval_Qnli-dev_max_precision": 0.5567010309278351, + "eval_Qnli-dev_max_recall": 0.9449152542372882, + "eval_allNLI-dev_cosine_accuracy": 0.734375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8394966721534729, + "eval_allNLI-dev_cosine_ap": 0.6113298350868344, + "eval_allNLI-dev_cosine_f1": 0.6349206349206349, + "eval_allNLI-dev_cosine_f1_threshold": 0.77137291431427, + "eval_allNLI-dev_cosine_precision": 0.5223880597014925, + "eval_allNLI-dev_cosine_recall": 0.8092485549132948, + "eval_allNLI-dev_dot_accuracy": 0.6953125, + "eval_allNLI-dev_dot_accuracy_threshold": 377.90240478515625, + "eval_allNLI-dev_dot_ap": 0.5250521827869041, + "eval_allNLI-dev_dot_f1": 0.5902912621359223, + "eval_allNLI-dev_dot_f1_threshold": 317.43182373046875, + "eval_allNLI-dev_dot_precision": 0.4444444444444444, + "eval_allNLI-dev_dot_recall": 0.8786127167630058, + "eval_allNLI-dev_euclidean_accuracy": 0.7265625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.836989402770996, + "eval_allNLI-dev_euclidean_ap": 0.616040839722229, + "eval_allNLI-dev_euclidean_f1": 0.6389496717724288, + "eval_allNLI-dev_euclidean_f1_threshold": 14.760133743286133, + "eval_allNLI-dev_euclidean_precision": 0.5140845070422535, + "eval_allNLI-dev_euclidean_recall": 0.8439306358381503, + "eval_allNLI-dev_manhattan_accuracy": 0.736328125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 233.19607543945312, + "eval_allNLI-dev_manhattan_ap": 0.6172051988439028, + "eval_allNLI-dev_manhattan_f1": 0.6321525885558583, + "eval_allNLI-dev_manhattan_f1_threshold": 233.74215698242188, + "eval_allNLI-dev_manhattan_precision": 0.5979381443298969, + "eval_allNLI-dev_manhattan_recall": 0.6705202312138728, + "eval_allNLI-dev_max_accuracy": 0.736328125, + "eval_allNLI-dev_max_accuracy_threshold": 377.90240478515625, + "eval_allNLI-dev_max_ap": 0.6172051988439028, + "eval_allNLI-dev_max_f1": 0.6389496717724288, + "eval_allNLI-dev_max_f1_threshold": 317.43182373046875, + "eval_allNLI-dev_max_precision": 0.5979381443298969, + "eval_allNLI-dev_max_recall": 0.8786127167630058, + "eval_sequential_score": 0.7181295462432675, + "eval_sts-test_pearson_cosine": 0.831421226462288, + "eval_sts-test_pearson_dot": 0.7898346781039345, + "eval_sts-test_pearson_euclidean": 0.8676114259798706, + "eval_sts-test_pearson_manhattan": 0.8642624716496483, + "eval_sts-test_pearson_max": 0.8676114259798706, + "eval_sts-test_spearman_cosine": 0.8736787418905085, + "eval_sts-test_spearman_dot": 0.7787073017068508, + "eval_sts-test_spearman_euclidean": 0.8696291554959393, + "eval_sts-test_spearman_manhattan": 0.8676987650845599, + "eval_sts-test_spearman_max": 0.8736787418905085, + "eval_vitaminc-pairs_loss": 4.0517659187316895, + "eval_vitaminc-pairs_runtime": 5.5747, + "eval_vitaminc-pairs_samples_per_second": 22.961, + "eval_vitaminc-pairs_steps_per_second": 0.359, + "step": 2140 + }, + { + "epoch": 1.6602017067494181, + "eval_negation-triplets_loss": 1.6389212608337402, + "eval_negation-triplets_runtime": 1.1097, + "eval_negation-triplets_samples_per_second": 115.348, + "eval_negation-triplets_steps_per_second": 1.802, + "step": 2140 + }, + { + "epoch": 1.6602017067494181, + "eval_scitail-pairs-pos_loss": 0.07726699858903885, + "eval_scitail-pairs-pos_runtime": 1.379, + "eval_scitail-pairs-pos_samples_per_second": 92.819, + "eval_scitail-pairs-pos_steps_per_second": 1.45, + "step": 2140 + }, + { + "epoch": 1.6602017067494181, + "eval_scitail-pairs-qa_loss": 0.003857515985146165, + "eval_scitail-pairs-qa_runtime": 0.7999, + "eval_scitail-pairs-qa_samples_per_second": 160.029, + "eval_scitail-pairs-qa_steps_per_second": 2.5, + "step": 2140 + }, + { + "epoch": 1.6602017067494181, + "eval_xsum-pairs_loss": 0.060311902314424515, + "eval_xsum-pairs_runtime": 3.9483, + "eval_xsum-pairs_samples_per_second": 32.419, + "eval_xsum-pairs_steps_per_second": 0.507, + "step": 2140 + }, + { + "epoch": 1.6602017067494181, + "eval_sciq_pairs_loss": 0.04089152440428734, + "eval_sciq_pairs_runtime": 6.2533, + "eval_sciq_pairs_samples_per_second": 20.469, + "eval_sciq_pairs_steps_per_second": 0.32, + "step": 2140 + }, + { + "epoch": 1.6602017067494181, + "eval_qasc_pairs_loss": 0.17752566933631897, + "eval_qasc_pairs_runtime": 0.9693, + "eval_qasc_pairs_samples_per_second": 132.05, + "eval_qasc_pairs_steps_per_second": 2.063, + "step": 2140 + }, + { + "epoch": 1.6602017067494181, + "eval_openbookqa_pairs_loss": 1.015750527381897, + "eval_openbookqa_pairs_runtime": 0.8887, + "eval_openbookqa_pairs_samples_per_second": 144.029, + "eval_openbookqa_pairs_steps_per_second": 2.25, + "step": 2140 + }, + { + "epoch": 1.6602017067494181, + "eval_msmarco_pairs_loss": 0.3456163704395294, + "eval_msmarco_pairs_runtime": 1.9934, + "eval_msmarco_pairs_samples_per_second": 64.212, + "eval_msmarco_pairs_steps_per_second": 1.003, + "step": 2140 + }, + { + "epoch": 1.6602017067494181, + "eval_nq_pairs_loss": 0.513410747051239, + "eval_nq_pairs_runtime": 3.3742, + "eval_nq_pairs_samples_per_second": 37.935, + "eval_nq_pairs_steps_per_second": 0.593, + "step": 2140 + }, + { + "epoch": 1.6602017067494181, + "eval_trivia_pairs_loss": 0.680845320224762, + "eval_trivia_pairs_runtime": 4.5129, + "eval_trivia_pairs_samples_per_second": 28.363, + "eval_trivia_pairs_steps_per_second": 0.443, + "step": 2140 + }, + { + "epoch": 1.6602017067494181, + "eval_gooaq_pairs_loss": 0.6120270490646362, + "eval_gooaq_pairs_runtime": 1.4984, + "eval_gooaq_pairs_samples_per_second": 85.426, + "eval_gooaq_pairs_steps_per_second": 1.335, + "step": 2140 + }, + { + "epoch": 1.6602017067494181, + "eval_paws-pos_loss": 0.031594689935445786, + "eval_paws-pos_runtime": 1.0749, + "eval_paws-pos_samples_per_second": 119.083, + "eval_paws-pos_steps_per_second": 1.861, + "step": 2140 + }, + { + "epoch": 1.6602017067494181, + "eval_global_dataset_loss": 0.42831510305404663, + "eval_global_dataset_runtime": 19.528, + "eval_global_dataset_samples_per_second": 21.303, + "eval_global_dataset_steps_per_second": 0.358, + "step": 2140 + }, + { + "epoch": 1.6609775019394881, + "grad_norm": 1.4193707704544067, + "learning_rate": 8.281454784172927e-06, + "loss": 0.0283, + "step": 2141 + }, + { + "epoch": 1.6617532971295579, + "grad_norm": 1.739780068397522, + "learning_rate": 8.260911956392959e-06, + "loss": 0.0383, + "step": 2142 + }, + { + "epoch": 1.6625290923196276, + "grad_norm": 4.73090124130249, + "learning_rate": 8.240405493027821e-06, + "loss": 0.0708, + "step": 2143 + }, + { + "epoch": 1.6633048875096974, + "grad_norm": 2.2647974491119385, + "learning_rate": 8.219935461936253e-06, + "loss": 0.035, + "step": 2144 + }, + { + "epoch": 1.6640806826997672, + "grad_norm": 3.511270046234131, + "learning_rate": 8.199501930856411e-06, + "loss": 0.0565, + "step": 2145 + }, + { + "epoch": 1.664856477889837, + "grad_norm": 5.970034122467041, + "learning_rate": 8.179104967405689e-06, + "loss": 0.1883, + "step": 2146 + }, + { + "epoch": 1.665632273079907, + "grad_norm": 0.539351224899292, + "learning_rate": 8.158744639080454e-06, + "loss": 0.0069, + "step": 2147 + }, + { + "epoch": 1.6664080682699767, + "grad_norm": 3.5285160541534424, + "learning_rate": 8.138421013255866e-06, + "loss": 0.0591, + "step": 2148 + }, + { + "epoch": 1.6671838634600467, + "grad_norm": 7.970695495605469, + "learning_rate": 8.118134157185612e-06, + "loss": 0.3678, + "step": 2149 + }, + { + "epoch": 1.6679596586501164, + "grad_norm": 3.7049474716186523, + "learning_rate": 8.097884138001717e-06, + "loss": 0.0581, + "step": 2150 + }, + { + "epoch": 1.6687354538401862, + "grad_norm": 9.41329288482666, + "learning_rate": 8.077671022714288e-06, + "loss": 0.4059, + "step": 2151 + }, + { + "epoch": 1.669511249030256, + "grad_norm": 3.3245697021484375, + "learning_rate": 8.057494878211328e-06, + "loss": 0.0897, + "step": 2152 + }, + { + "epoch": 1.6702870442203257, + "grad_norm": 6.682858467102051, + "learning_rate": 8.037355771258498e-06, + "loss": 0.4522, + "step": 2153 + }, + { + "epoch": 1.6710628394103957, + "grad_norm": 7.449135780334473, + "learning_rate": 8.017253768498898e-06, + "loss": 0.2594, + "step": 2154 + }, + { + "epoch": 1.6718386346004654, + "grad_norm": 3.6802499294281006, + "learning_rate": 7.997188936452829e-06, + "loss": 0.1329, + "step": 2155 + }, + { + "epoch": 1.6726144297905354, + "grad_norm": 6.7140045166015625, + "learning_rate": 7.977161341517606e-06, + "loss": 0.4699, + "step": 2156 + }, + { + "epoch": 1.6733902249806052, + "grad_norm": 3.3418357372283936, + "learning_rate": 7.957171049967313e-06, + "loss": 0.048, + "step": 2157 + }, + { + "epoch": 1.674166020170675, + "grad_norm": 4.464799880981445, + "learning_rate": 7.937218127952595e-06, + "loss": 0.1194, + "step": 2158 + }, + { + "epoch": 1.6749418153607447, + "grad_norm": 10.836454391479492, + "learning_rate": 7.917302641500433e-06, + "loss": 0.6736, + "step": 2159 + }, + { + "epoch": 1.6757176105508145, + "grad_norm": 4.374032974243164, + "learning_rate": 7.897424656513944e-06, + "loss": 0.1029, + "step": 2160 + }, + { + "epoch": 1.6757176105508145, + "eval_Qnli-dev_cosine_accuracy": 0.677734375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8089977502822876, + "eval_Qnli-dev_cosine_ap": 0.7164770850451074, + "eval_Qnli-dev_cosine_f1": 0.6919431279620852, + "eval_Qnli-dev_cosine_f1_threshold": 0.6832389831542969, + "eval_Qnli-dev_cosine_precision": 0.5516372795969773, + "eval_Qnli-dev_cosine_recall": 0.9279661016949152, + "eval_Qnli-dev_dot_accuracy": 0.654296875, + "eval_Qnli-dev_dot_accuracy_threshold": 408.4425048828125, + "eval_Qnli-dev_dot_ap": 0.6385939581485318, + "eval_Qnli-dev_dot_f1": 0.6775244299674267, + "eval_Qnli-dev_dot_f1_threshold": 328.9086608886719, + "eval_Qnli-dev_dot_precision": 0.5502645502645502, + "eval_Qnli-dev_dot_recall": 0.8813559322033898, + "eval_Qnli-dev_euclidean_accuracy": 0.67578125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.56654167175293, + "eval_Qnli-dev_euclidean_ap": 0.7177549509346787, + "eval_Qnli-dev_euclidean_f1": 0.69281045751634, + "eval_Qnli-dev_euclidean_f1_threshold": 16.92102813720703, + "eval_Qnli-dev_euclidean_precision": 0.5638297872340425, + "eval_Qnli-dev_euclidean_recall": 0.8983050847457628, + "eval_Qnli-dev_manhattan_accuracy": 0.68359375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 249.37521362304688, + "eval_Qnli-dev_manhattan_ap": 0.7190034321128935, + "eval_Qnli-dev_manhattan_f1": 0.6864686468646866, + "eval_Qnli-dev_manhattan_f1_threshold": 308.01031494140625, + "eval_Qnli-dev_manhattan_precision": 0.5621621621621622, + "eval_Qnli-dev_manhattan_recall": 0.8813559322033898, + "eval_Qnli-dev_max_accuracy": 0.68359375, + "eval_Qnli-dev_max_accuracy_threshold": 408.4425048828125, + "eval_Qnli-dev_max_ap": 0.7190034321128935, + "eval_Qnli-dev_max_f1": 0.69281045751634, + "eval_Qnli-dev_max_f1_threshold": 328.9086608886719, + "eval_Qnli-dev_max_precision": 0.5638297872340425, + "eval_Qnli-dev_max_recall": 0.9279661016949152, + "eval_allNLI-dev_cosine_accuracy": 0.740234375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.831339955329895, + "eval_allNLI-dev_cosine_ap": 0.6099436344176263, + "eval_allNLI-dev_cosine_f1": 0.6332574031890661, + "eval_allNLI-dev_cosine_f1_threshold": 0.7633983492851257, + "eval_allNLI-dev_cosine_precision": 0.5225563909774437, + "eval_allNLI-dev_cosine_recall": 0.8034682080924855, + "eval_allNLI-dev_dot_accuracy": 0.697265625, + "eval_allNLI-dev_dot_accuracy_threshold": 369.04949951171875, + "eval_allNLI-dev_dot_ap": 0.5277008797716145, + "eval_allNLI-dev_dot_f1": 0.5896414342629482, + "eval_allNLI-dev_dot_f1_threshold": 314.61968994140625, + "eval_allNLI-dev_dot_precision": 0.44984802431610943, + "eval_allNLI-dev_dot_recall": 0.8554913294797688, + "eval_allNLI-dev_euclidean_accuracy": 0.736328125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.960643768310547, + "eval_allNLI-dev_euclidean_ap": 0.6144779239815363, + "eval_allNLI-dev_euclidean_f1": 0.64, + "eval_allNLI-dev_euclidean_f1_threshold": 14.733768463134766, + "eval_allNLI-dev_euclidean_precision": 0.51985559566787, + "eval_allNLI-dev_euclidean_recall": 0.8323699421965318, + "eval_allNLI-dev_manhattan_accuracy": 0.740234375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 233.75531005859375, + "eval_allNLI-dev_manhattan_ap": 0.6147640556335215, + "eval_allNLI-dev_manhattan_f1": 0.6396396396396398, + "eval_allNLI-dev_manhattan_f1_threshold": 269.86029052734375, + "eval_allNLI-dev_manhattan_precision": 0.5239852398523985, + "eval_allNLI-dev_manhattan_recall": 0.8208092485549133, + "eval_allNLI-dev_max_accuracy": 0.740234375, + "eval_allNLI-dev_max_accuracy_threshold": 369.04949951171875, + "eval_allNLI-dev_max_ap": 0.6147640556335215, + "eval_allNLI-dev_max_f1": 0.64, + "eval_allNLI-dev_max_f1_threshold": 314.61968994140625, + "eval_allNLI-dev_max_precision": 0.5239852398523985, + "eval_allNLI-dev_max_recall": 0.8554913294797688, + "eval_sequential_score": 0.7190034321128935, + "eval_sts-test_pearson_cosine": 0.8330143603254968, + "eval_sts-test_pearson_dot": 0.7976138121899649, + "eval_sts-test_pearson_euclidean": 0.8699429843395153, + "eval_sts-test_pearson_manhattan": 0.866792135118187, + "eval_sts-test_pearson_max": 0.8699429843395153, + "eval_sts-test_spearman_cosine": 0.8764818714015431, + "eval_sts-test_spearman_dot": 0.7907289673082065, + "eval_sts-test_spearman_euclidean": 0.8726754186101134, + "eval_sts-test_spearman_manhattan": 0.8711434486789015, + "eval_sts-test_spearman_max": 0.8764818714015431, + "eval_vitaminc-pairs_loss": 3.9939541816711426, + "eval_vitaminc-pairs_runtime": 5.5862, + "eval_vitaminc-pairs_samples_per_second": 22.914, + "eval_vitaminc-pairs_steps_per_second": 0.358, + "step": 2160 + }, + { + "epoch": 1.6757176105508145, + "eval_negation-triplets_loss": 1.6463830471038818, + "eval_negation-triplets_runtime": 1.1062, + "eval_negation-triplets_samples_per_second": 115.71, + "eval_negation-triplets_steps_per_second": 1.808, + "step": 2160 + }, + { + "epoch": 1.6757176105508145, + "eval_scitail-pairs-pos_loss": 0.07163487374782562, + "eval_scitail-pairs-pos_runtime": 1.4015, + "eval_scitail-pairs-pos_samples_per_second": 91.33, + "eval_scitail-pairs-pos_steps_per_second": 1.427, + "step": 2160 + }, + { + "epoch": 1.6757176105508145, + "eval_scitail-pairs-qa_loss": 0.0033316684421151876, + "eval_scitail-pairs-qa_runtime": 0.8019, + "eval_scitail-pairs-qa_samples_per_second": 159.623, + "eval_scitail-pairs-qa_steps_per_second": 2.494, + "step": 2160 + }, + { + "epoch": 1.6757176105508145, + "eval_xsum-pairs_loss": 0.05443917214870453, + "eval_xsum-pairs_runtime": 3.9457, + "eval_xsum-pairs_samples_per_second": 32.441, + "eval_xsum-pairs_steps_per_second": 0.507, + "step": 2160 + }, + { + "epoch": 1.6757176105508145, + "eval_sciq_pairs_loss": 0.03834564983844757, + "eval_sciq_pairs_runtime": 6.2874, + "eval_sciq_pairs_samples_per_second": 20.358, + "eval_sciq_pairs_steps_per_second": 0.318, + "step": 2160 + }, + { + "epoch": 1.6757176105508145, + "eval_qasc_pairs_loss": 0.17700031399726868, + "eval_qasc_pairs_runtime": 0.957, + "eval_qasc_pairs_samples_per_second": 133.75, + "eval_qasc_pairs_steps_per_second": 2.09, + "step": 2160 + }, + { + "epoch": 1.6757176105508145, + "eval_openbookqa_pairs_loss": 1.0174033641815186, + "eval_openbookqa_pairs_runtime": 0.9152, + "eval_openbookqa_pairs_samples_per_second": 139.865, + "eval_openbookqa_pairs_steps_per_second": 2.185, + "step": 2160 + }, + { + "epoch": 1.6757176105508145, + "eval_msmarco_pairs_loss": 0.3528894782066345, + "eval_msmarco_pairs_runtime": 1.9947, + "eval_msmarco_pairs_samples_per_second": 64.17, + "eval_msmarco_pairs_steps_per_second": 1.003, + "step": 2160 + }, + { + "epoch": 1.6757176105508145, + "eval_nq_pairs_loss": 0.528623104095459, + "eval_nq_pairs_runtime": 3.3752, + "eval_nq_pairs_samples_per_second": 37.924, + "eval_nq_pairs_steps_per_second": 0.593, + "step": 2160 + }, + { + "epoch": 1.6757176105508145, + "eval_trivia_pairs_loss": 0.6889814734458923, + "eval_trivia_pairs_runtime": 4.5217, + "eval_trivia_pairs_samples_per_second": 28.308, + "eval_trivia_pairs_steps_per_second": 0.442, + "step": 2160 + }, + { + "epoch": 1.6757176105508145, + "eval_gooaq_pairs_loss": 0.6055079698562622, + "eval_gooaq_pairs_runtime": 1.5079, + "eval_gooaq_pairs_samples_per_second": 84.885, + "eval_gooaq_pairs_steps_per_second": 1.326, + "step": 2160 + }, + { + "epoch": 1.6757176105508145, + "eval_paws-pos_loss": 0.03229650482535362, + "eval_paws-pos_runtime": 1.0875, + "eval_paws-pos_samples_per_second": 117.705, + "eval_paws-pos_steps_per_second": 1.839, + "step": 2160 + }, + { + "epoch": 1.6757176105508145, + "eval_global_dataset_loss": 0.42307451367378235, + "eval_global_dataset_runtime": 19.5593, + "eval_global_dataset_samples_per_second": 21.269, + "eval_global_dataset_steps_per_second": 0.358, + "step": 2160 + }, + { + "epoch": 1.6764934057408845, + "grad_norm": 6.188950061798096, + "learning_rate": 7.877584238772115e-06, + "loss": 0.1754, + "step": 2161 + }, + { + "epoch": 1.6772692009309542, + "grad_norm": 4.110795021057129, + "learning_rate": 7.857781453929653e-06, + "loss": 0.1009, + "step": 2162 + }, + { + "epoch": 1.6780449961210242, + "grad_norm": 4.659564971923828, + "learning_rate": 7.8380163675167e-06, + "loss": 0.1037, + "step": 2163 + }, + { + "epoch": 1.678820791311094, + "grad_norm": 7.453348636627197, + "learning_rate": 7.818289044938685e-06, + "loss": 0.4123, + "step": 2164 + }, + { + "epoch": 1.6795965865011637, + "grad_norm": 6.408524513244629, + "learning_rate": 7.798599551476036e-06, + "loss": 0.2947, + "step": 2165 + }, + { + "epoch": 1.6803723816912335, + "grad_norm": 6.922446250915527, + "learning_rate": 7.778947952284026e-06, + "loss": 0.2313, + "step": 2166 + }, + { + "epoch": 1.6811481768813032, + "grad_norm": 5.302913665771484, + "learning_rate": 7.759334312392509e-06, + "loss": 0.1126, + "step": 2167 + }, + { + "epoch": 1.681923972071373, + "grad_norm": 6.297609806060791, + "learning_rate": 7.73975869670575e-06, + "loss": 0.1992, + "step": 2168 + }, + { + "epoch": 1.682699767261443, + "grad_norm": 5.348913192749023, + "learning_rate": 7.72022117000215e-06, + "loss": 0.1742, + "step": 2169 + }, + { + "epoch": 1.6834755624515128, + "grad_norm": 1.9818607568740845, + "learning_rate": 7.70072179693412e-06, + "loss": 0.0322, + "step": 2170 + }, + { + "epoch": 1.6842513576415827, + "grad_norm": 3.3500750064849854, + "learning_rate": 7.681260642027768e-06, + "loss": 0.0589, + "step": 2171 + }, + { + "epoch": 1.6850271528316525, + "grad_norm": 3.284574031829834, + "learning_rate": 7.661837769682764e-06, + "loss": 0.2137, + "step": 2172 + }, + { + "epoch": 1.6858029480217223, + "grad_norm": 8.348149299621582, + "learning_rate": 7.64245324417207e-06, + "loss": 0.3659, + "step": 2173 + }, + { + "epoch": 1.686578743211792, + "grad_norm": 5.099652290344238, + "learning_rate": 7.623107129641776e-06, + "loss": 0.1042, + "step": 2174 + }, + { + "epoch": 1.6873545384018618, + "grad_norm": 7.712127208709717, + "learning_rate": 7.603799490110853e-06, + "loss": 0.2597, + "step": 2175 + }, + { + "epoch": 1.6881303335919318, + "grad_norm": 9.221052169799805, + "learning_rate": 7.5845303894709605e-06, + "loss": 0.3855, + "step": 2176 + }, + { + "epoch": 1.6889061287820015, + "grad_norm": 6.962251663208008, + "learning_rate": 7.565299891486211e-06, + "loss": 0.4895, + "step": 2177 + }, + { + "epoch": 1.6896819239720715, + "grad_norm": 3.1279869079589844, + "learning_rate": 7.546108059793e-06, + "loss": 0.0613, + "step": 2178 + }, + { + "epoch": 1.6904577191621413, + "grad_norm": 5.356664180755615, + "learning_rate": 7.526954957899744e-06, + "loss": 0.1362, + "step": 2179 + }, + { + "epoch": 1.691233514352211, + "grad_norm": 15.599915504455566, + "learning_rate": 7.507840649186717e-06, + "loss": 1.7851, + "step": 2180 + }, + { + "epoch": 1.691233514352211, + "eval_Qnli-dev_cosine_accuracy": 0.671875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8171384930610657, + "eval_Qnli-dev_cosine_ap": 0.7160996252633772, + "eval_Qnli-dev_cosine_f1": 0.6883116883116883, + "eval_Qnli-dev_cosine_f1_threshold": 0.6971203684806824, + "eval_Qnli-dev_cosine_precision": 0.5578947368421052, + "eval_Qnli-dev_cosine_recall": 0.8983050847457628, + "eval_Qnli-dev_dot_accuracy": 0.654296875, + "eval_Qnli-dev_dot_accuracy_threshold": 401.2979736328125, + "eval_Qnli-dev_dot_ap": 0.635880478530489, + "eval_Qnli-dev_dot_f1": 0.6810477657935284, + "eval_Qnli-dev_dot_f1_threshold": 313.5450134277344, + "eval_Qnli-dev_dot_precision": 0.5351089588377724, + "eval_Qnli-dev_dot_recall": 0.9364406779661016, + "eval_Qnli-dev_euclidean_accuracy": 0.671875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.779165267944336, + "eval_Qnli-dev_euclidean_ap": 0.716578387674031, + "eval_Qnli-dev_euclidean_f1": 0.6889632107023412, + "eval_Qnli-dev_euclidean_f1_threshold": 16.82510757446289, + "eval_Qnli-dev_euclidean_precision": 0.569060773480663, + "eval_Qnli-dev_euclidean_recall": 0.8728813559322034, + "eval_Qnli-dev_manhattan_accuracy": 0.677734375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 249.25936889648438, + "eval_Qnli-dev_manhattan_ap": 0.7161592891102796, + "eval_Qnli-dev_manhattan_f1": 0.6864686468646866, + "eval_Qnli-dev_manhattan_f1_threshold": 309.66021728515625, + "eval_Qnli-dev_manhattan_precision": 0.5621621621621622, + "eval_Qnli-dev_manhattan_recall": 0.8813559322033898, + "eval_Qnli-dev_max_accuracy": 0.677734375, + "eval_Qnli-dev_max_accuracy_threshold": 401.2979736328125, + "eval_Qnli-dev_max_ap": 0.716578387674031, + "eval_Qnli-dev_max_f1": 0.6889632107023412, + "eval_Qnli-dev_max_f1_threshold": 313.5450134277344, + "eval_Qnli-dev_max_precision": 0.569060773480663, + "eval_Qnli-dev_max_recall": 0.9364406779661016, + "eval_allNLI-dev_cosine_accuracy": 0.734375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8405022621154785, + "eval_allNLI-dev_cosine_ap": 0.6085970652659555, + "eval_allNLI-dev_cosine_f1": 0.6271929824561404, + "eval_allNLI-dev_cosine_f1_threshold": 0.7585848569869995, + "eval_allNLI-dev_cosine_precision": 0.5053003533568905, + "eval_allNLI-dev_cosine_recall": 0.8265895953757225, + "eval_allNLI-dev_dot_accuracy": 0.705078125, + "eval_allNLI-dev_dot_accuracy_threshold": 374.7132568359375, + "eval_allNLI-dev_dot_ap": 0.5261134119711065, + "eval_allNLI-dev_dot_f1": 0.5816733067729084, + "eval_allNLI-dev_dot_f1_threshold": 317.80047607421875, + "eval_allNLI-dev_dot_precision": 0.44376899696048633, + "eval_allNLI-dev_dot_recall": 0.8439306358381503, + "eval_allNLI-dev_euclidean_accuracy": 0.734375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 12.094595909118652, + "eval_allNLI-dev_euclidean_ap": 0.6114937641196427, + "eval_allNLI-dev_euclidean_f1": 0.6363636363636364, + "eval_allNLI-dev_euclidean_f1_threshold": 14.910051345825195, + "eval_allNLI-dev_euclidean_precision": 0.5086505190311419, + "eval_allNLI-dev_euclidean_recall": 0.8497109826589595, + "eval_allNLI-dev_manhattan_accuracy": 0.736328125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 226.37252807617188, + "eval_allNLI-dev_manhattan_ap": 0.613958065242196, + "eval_allNLI-dev_manhattan_f1": 0.639269406392694, + "eval_allNLI-dev_manhattan_f1_threshold": 266.17620849609375, + "eval_allNLI-dev_manhattan_precision": 0.5283018867924528, + "eval_allNLI-dev_manhattan_recall": 0.8092485549132948, + "eval_allNLI-dev_max_accuracy": 0.736328125, + "eval_allNLI-dev_max_accuracy_threshold": 374.7132568359375, + "eval_allNLI-dev_max_ap": 0.613958065242196, + "eval_allNLI-dev_max_f1": 0.639269406392694, + "eval_allNLI-dev_max_f1_threshold": 317.80047607421875, + "eval_allNLI-dev_max_precision": 0.5283018867924528, + "eval_allNLI-dev_max_recall": 0.8497109826589595, + "eval_sequential_score": 0.716578387674031, + "eval_sts-test_pearson_cosine": 0.829787290775935, + "eval_sts-test_pearson_dot": 0.7951702962721949, + "eval_sts-test_pearson_euclidean": 0.8670950576910648, + "eval_sts-test_pearson_manhattan": 0.863952322401149, + "eval_sts-test_pearson_max": 0.8670950576910648, + "eval_sts-test_spearman_cosine": 0.874477551906345, + "eval_sts-test_spearman_dot": 0.7901195446305785, + "eval_sts-test_spearman_euclidean": 0.8702467229031405, + "eval_sts-test_spearman_manhattan": 0.8680876087947899, + "eval_sts-test_spearman_max": 0.874477551906345, + "eval_vitaminc-pairs_loss": 3.993006467819214, + "eval_vitaminc-pairs_runtime": 5.5505, + "eval_vitaminc-pairs_samples_per_second": 23.061, + "eval_vitaminc-pairs_steps_per_second": 0.36, + "step": 2180 + }, + { + "epoch": 1.691233514352211, + "eval_negation-triplets_loss": 1.6651763916015625, + "eval_negation-triplets_runtime": 1.0993, + "eval_negation-triplets_samples_per_second": 116.438, + "eval_negation-triplets_steps_per_second": 1.819, + "step": 2180 + }, + { + "epoch": 1.691233514352211, + "eval_scitail-pairs-pos_loss": 0.0819476842880249, + "eval_scitail-pairs-pos_runtime": 1.3817, + "eval_scitail-pairs-pos_samples_per_second": 92.641, + "eval_scitail-pairs-pos_steps_per_second": 1.448, + "step": 2180 + }, + { + "epoch": 1.691233514352211, + "eval_scitail-pairs-qa_loss": 0.0036004537250846624, + "eval_scitail-pairs-qa_runtime": 0.7978, + "eval_scitail-pairs-qa_samples_per_second": 160.433, + "eval_scitail-pairs-qa_steps_per_second": 2.507, + "step": 2180 + }, + { + "epoch": 1.691233514352211, + "eval_xsum-pairs_loss": 0.05180860683321953, + "eval_xsum-pairs_runtime": 3.9674, + "eval_xsum-pairs_samples_per_second": 32.263, + "eval_xsum-pairs_steps_per_second": 0.504, + "step": 2180 + }, + { + "epoch": 1.691233514352211, + "eval_sciq_pairs_loss": 0.03650512918829918, + "eval_sciq_pairs_runtime": 6.1398, + "eval_sciq_pairs_samples_per_second": 20.848, + "eval_sciq_pairs_steps_per_second": 0.326, + "step": 2180 + }, + { + "epoch": 1.691233514352211, + "eval_qasc_pairs_loss": 0.18227122724056244, + "eval_qasc_pairs_runtime": 0.9413, + "eval_qasc_pairs_samples_per_second": 135.986, + "eval_qasc_pairs_steps_per_second": 2.125, + "step": 2180 + }, + { + "epoch": 1.691233514352211, + "eval_openbookqa_pairs_loss": 0.994342029094696, + "eval_openbookqa_pairs_runtime": 0.8933, + "eval_openbookqa_pairs_samples_per_second": 143.282, + "eval_openbookqa_pairs_steps_per_second": 2.239, + "step": 2180 + }, + { + "epoch": 1.691233514352211, + "eval_msmarco_pairs_loss": 0.35695943236351013, + "eval_msmarco_pairs_runtime": 1.9976, + "eval_msmarco_pairs_samples_per_second": 64.078, + "eval_msmarco_pairs_steps_per_second": 1.001, + "step": 2180 + }, + { + "epoch": 1.691233514352211, + "eval_nq_pairs_loss": 0.51131272315979, + "eval_nq_pairs_runtime": 3.3675, + "eval_nq_pairs_samples_per_second": 38.011, + "eval_nq_pairs_steps_per_second": 0.594, + "step": 2180 + }, + { + "epoch": 1.691233514352211, + "eval_trivia_pairs_loss": 0.6971176266670227, + "eval_trivia_pairs_runtime": 4.5187, + "eval_trivia_pairs_samples_per_second": 28.327, + "eval_trivia_pairs_steps_per_second": 0.443, + "step": 2180 + }, + { + "epoch": 1.691233514352211, + "eval_gooaq_pairs_loss": 0.5807812213897705, + "eval_gooaq_pairs_runtime": 1.4967, + "eval_gooaq_pairs_samples_per_second": 85.519, + "eval_gooaq_pairs_steps_per_second": 1.336, + "step": 2180 + }, + { + "epoch": 1.691233514352211, + "eval_paws-pos_loss": 0.03196890652179718, + "eval_paws-pos_runtime": 1.08, + "eval_paws-pos_samples_per_second": 118.514, + "eval_paws-pos_steps_per_second": 1.852, + "step": 2180 + }, + { + "epoch": 1.691233514352211, + "eval_global_dataset_loss": 0.41824451088905334, + "eval_global_dataset_runtime": 19.5372, + "eval_global_dataset_samples_per_second": 21.293, + "eval_global_dataset_steps_per_second": 0.358, + "step": 2180 + }, + { + "epoch": 1.6920093095422808, + "grad_norm": 11.586441040039062, + "learning_rate": 7.488765196905814e-06, + "loss": 0.4452, + "step": 2181 + }, + { + "epoch": 1.6927851047323506, + "grad_norm": 0.6499935388565063, + "learning_rate": 7.4697286641803575e-06, + "loss": 0.0295, + "step": 2182 + }, + { + "epoch": 1.6935608999224203, + "grad_norm": 4.3818359375, + "learning_rate": 7.450731114004859e-06, + "loss": 0.1009, + "step": 2183 + }, + { + "epoch": 1.6943366951124903, + "grad_norm": 8.34776496887207, + "learning_rate": 7.431772609244853e-06, + "loss": 0.4534, + "step": 2184 + }, + { + "epoch": 1.6951124903025603, + "grad_norm": 5.9303059577941895, + "learning_rate": 7.412853212636648e-06, + "loss": 0.1395, + "step": 2185 + }, + { + "epoch": 1.69588828549263, + "grad_norm": 4.192590236663818, + "learning_rate": 7.393972986787163e-06, + "loss": 0.1014, + "step": 2186 + }, + { + "epoch": 1.6966640806826998, + "grad_norm": 3.3982746601104736, + "learning_rate": 7.375131994173669e-06, + "loss": 0.0782, + "step": 2187 + }, + { + "epoch": 1.6974398758727696, + "grad_norm": 8.83377742767334, + "learning_rate": 7.356330297143631e-06, + "loss": 0.4526, + "step": 2188 + }, + { + "epoch": 1.6982156710628393, + "grad_norm": 10.193682670593262, + "learning_rate": 7.337567957914459e-06, + "loss": 0.5857, + "step": 2189 + }, + { + "epoch": 1.698991466252909, + "grad_norm": 3.9441211223602295, + "learning_rate": 7.3188450385733424e-06, + "loss": 0.1253, + "step": 2190 + }, + { + "epoch": 1.699767261442979, + "grad_norm": 1.805309534072876, + "learning_rate": 7.300161601077002e-06, + "loss": 0.0309, + "step": 2191 + }, + { + "epoch": 1.7005430566330488, + "grad_norm": 3.9004831314086914, + "learning_rate": 7.28151770725154e-06, + "loss": 0.1035, + "step": 2192 + }, + { + "epoch": 1.7013188518231188, + "grad_norm": 3.3366801738739014, + "learning_rate": 7.262913418792167e-06, + "loss": 0.1006, + "step": 2193 + }, + { + "epoch": 1.7020946470131886, + "grad_norm": 3.2111222743988037, + "learning_rate": 7.244348797263065e-06, + "loss": 0.0382, + "step": 2194 + }, + { + "epoch": 1.7028704422032583, + "grad_norm": 3.3468010425567627, + "learning_rate": 7.225823904097126e-06, + "loss": 0.0877, + "step": 2195 + }, + { + "epoch": 1.703646237393328, + "grad_norm": 7.553779125213623, + "learning_rate": 7.207338800595793e-06, + "loss": 0.2736, + "step": 2196 + }, + { + "epoch": 1.7044220325833979, + "grad_norm": 4.968908786773682, + "learning_rate": 7.188893547928835e-06, + "loss": 0.2681, + "step": 2197 + }, + { + "epoch": 1.7051978277734678, + "grad_norm": 4.194811820983887, + "learning_rate": 7.170488207134155e-06, + "loss": 0.1156, + "step": 2198 + }, + { + "epoch": 1.7059736229635376, + "grad_norm": 3.3986384868621826, + "learning_rate": 7.152122839117564e-06, + "loss": 0.0779, + "step": 2199 + }, + { + "epoch": 1.7067494181536076, + "grad_norm": 6.786818981170654, + "learning_rate": 7.133797504652621e-06, + "loss": 0.1913, + "step": 2200 + }, + { + "epoch": 1.7067494181536076, + "eval_Qnli-dev_cosine_accuracy": 0.671875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8359333276748657, + "eval_Qnli-dev_cosine_ap": 0.7220408442169709, + "eval_Qnli-dev_cosine_f1": 0.688888888888889, + "eval_Qnli-dev_cosine_f1_threshold": 0.6887697577476501, + "eval_Qnli-dev_cosine_precision": 0.550761421319797, + "eval_Qnli-dev_cosine_recall": 0.9194915254237288, + "eval_Qnli-dev_dot_accuracy": 0.658203125, + "eval_Qnli-dev_dot_accuracy_threshold": 384.63787841796875, + "eval_Qnli-dev_dot_ap": 0.6447609323340185, + "eval_Qnli-dev_dot_f1": 0.6791277258566978, + "eval_Qnli-dev_dot_f1_threshold": 321.091552734375, + "eval_Qnli-dev_dot_precision": 0.5369458128078818, + "eval_Qnli-dev_dot_recall": 0.923728813559322, + "eval_Qnli-dev_euclidean_accuracy": 0.677734375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.798051834106445, + "eval_Qnli-dev_euclidean_ap": 0.7203967395518998, + "eval_Qnli-dev_euclidean_f1": 0.6902356902356903, + "eval_Qnli-dev_euclidean_f1_threshold": 16.77446746826172, + "eval_Qnli-dev_euclidean_precision": 0.5726256983240223, + "eval_Qnli-dev_euclidean_recall": 0.8686440677966102, + "eval_Qnli-dev_manhattan_accuracy": 0.671875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 251.71177673339844, + "eval_Qnli-dev_manhattan_ap": 0.7214639970269668, + "eval_Qnli-dev_manhattan_f1": 0.6873977086743045, + "eval_Qnli-dev_manhattan_f1_threshold": 312.7536315917969, + "eval_Qnli-dev_manhattan_precision": 0.56, + "eval_Qnli-dev_manhattan_recall": 0.8898305084745762, + "eval_Qnli-dev_max_accuracy": 0.677734375, + "eval_Qnli-dev_max_accuracy_threshold": 384.63787841796875, + "eval_Qnli-dev_max_ap": 0.7220408442169709, + "eval_Qnli-dev_max_f1": 0.6902356902356903, + "eval_Qnli-dev_max_f1_threshold": 321.091552734375, + "eval_Qnli-dev_max_precision": 0.5726256983240223, + "eval_Qnli-dev_max_recall": 0.923728813559322, + "eval_allNLI-dev_cosine_accuracy": 0.736328125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8356429934501648, + "eval_allNLI-dev_cosine_ap": 0.6088940283747784, + "eval_allNLI-dev_cosine_f1": 0.6266094420600858, + "eval_allNLI-dev_cosine_f1_threshold": 0.7475576400756836, + "eval_allNLI-dev_cosine_precision": 0.49829351535836175, + "eval_allNLI-dev_cosine_recall": 0.8439306358381503, + "eval_allNLI-dev_dot_accuracy": 0.703125, + "eval_allNLI-dev_dot_accuracy_threshold": 376.5745849609375, + "eval_allNLI-dev_dot_ap": 0.5245456023907981, + "eval_allNLI-dev_dot_f1": 0.5795918367346938, + "eval_allNLI-dev_dot_f1_threshold": 323.87872314453125, + "eval_allNLI-dev_dot_precision": 0.4479495268138801, + "eval_allNLI-dev_dot_recall": 0.8208092485549133, + "eval_allNLI-dev_euclidean_accuracy": 0.7265625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 12.278661727905273, + "eval_allNLI-dev_euclidean_ap": 0.6119173455586703, + "eval_allNLI-dev_euclidean_f1": 0.6410835214446953, + "eval_allNLI-dev_euclidean_f1_threshold": 14.594625473022461, + "eval_allNLI-dev_euclidean_precision": 0.5259259259259259, + "eval_allNLI-dev_euclidean_recall": 0.8208092485549133, + "eval_allNLI-dev_manhattan_accuracy": 0.734375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 230.93019104003906, + "eval_allNLI-dev_manhattan_ap": 0.6155539984509018, + "eval_allNLI-dev_manhattan_f1": 0.6334841628959276, + "eval_allNLI-dev_manhattan_f1_threshold": 269.4210510253906, + "eval_allNLI-dev_manhattan_precision": 0.5204460966542751, + "eval_allNLI-dev_manhattan_recall": 0.8092485549132948, + "eval_allNLI-dev_max_accuracy": 0.736328125, + "eval_allNLI-dev_max_accuracy_threshold": 376.5745849609375, + "eval_allNLI-dev_max_ap": 0.6155539984509018, + "eval_allNLI-dev_max_f1": 0.6410835214446953, + "eval_allNLI-dev_max_f1_threshold": 323.87872314453125, + "eval_allNLI-dev_max_precision": 0.5259259259259259, + "eval_allNLI-dev_max_recall": 0.8439306358381503, + "eval_sequential_score": 0.7220408442169709, + "eval_sts-test_pearson_cosine": 0.830362635891251, + "eval_sts-test_pearson_dot": 0.7971321328848736, + "eval_sts-test_pearson_euclidean": 0.8649695141471707, + "eval_sts-test_pearson_manhattan": 0.8618061837688793, + "eval_sts-test_pearson_max": 0.8649695141471707, + "eval_sts-test_spearman_cosine": 0.8717315250048459, + "eval_sts-test_spearman_dot": 0.7924249058611563, + "eval_sts-test_spearman_euclidean": 0.8670878208640446, + "eval_sts-test_spearman_manhattan": 0.865141633257394, + "eval_sts-test_spearman_max": 0.8717315250048459, + "eval_vitaminc-pairs_loss": 3.8740808963775635, + "eval_vitaminc-pairs_runtime": 5.5572, + "eval_vitaminc-pairs_samples_per_second": 23.033, + "eval_vitaminc-pairs_steps_per_second": 0.36, + "step": 2200 + }, + { + "epoch": 1.7067494181536076, + "eval_negation-triplets_loss": 1.6735501289367676, + "eval_negation-triplets_runtime": 1.1025, + "eval_negation-triplets_samples_per_second": 116.101, + "eval_negation-triplets_steps_per_second": 1.814, + "step": 2200 + }, + { + "epoch": 1.7067494181536076, + "eval_scitail-pairs-pos_loss": 0.08114144951105118, + "eval_scitail-pairs-pos_runtime": 1.3556, + "eval_scitail-pairs-pos_samples_per_second": 94.42, + "eval_scitail-pairs-pos_steps_per_second": 1.475, + "step": 2200 + }, + { + "epoch": 1.7067494181536076, + "eval_scitail-pairs-qa_loss": 0.0036077096592634916, + "eval_scitail-pairs-qa_runtime": 0.7964, + "eval_scitail-pairs-qa_samples_per_second": 160.732, + "eval_scitail-pairs-qa_steps_per_second": 2.511, + "step": 2200 + }, + { + "epoch": 1.7067494181536076, + "eval_xsum-pairs_loss": 0.05097520351409912, + "eval_xsum-pairs_runtime": 3.9685, + "eval_xsum-pairs_samples_per_second": 32.254, + "eval_xsum-pairs_steps_per_second": 0.504, + "step": 2200 + }, + { + "epoch": 1.7067494181536076, + "eval_sciq_pairs_loss": 0.03588932752609253, + "eval_sciq_pairs_runtime": 6.3447, + "eval_sciq_pairs_samples_per_second": 20.174, + "eval_sciq_pairs_steps_per_second": 0.315, + "step": 2200 + }, + { + "epoch": 1.7067494181536076, + "eval_qasc_pairs_loss": 0.18052475154399872, + "eval_qasc_pairs_runtime": 0.9888, + "eval_qasc_pairs_samples_per_second": 129.444, + "eval_qasc_pairs_steps_per_second": 2.023, + "step": 2200 + }, + { + "epoch": 1.7067494181536076, + "eval_openbookqa_pairs_loss": 0.979564905166626, + "eval_openbookqa_pairs_runtime": 0.9143, + "eval_openbookqa_pairs_samples_per_second": 139.993, + "eval_openbookqa_pairs_steps_per_second": 2.187, + "step": 2200 + }, + { + "epoch": 1.7067494181536076, + "eval_msmarco_pairs_loss": 0.3533523976802826, + "eval_msmarco_pairs_runtime": 2.002, + "eval_msmarco_pairs_samples_per_second": 63.937, + "eval_msmarco_pairs_steps_per_second": 0.999, + "step": 2200 + }, + { + "epoch": 1.7067494181536076, + "eval_nq_pairs_loss": 0.4935872256755829, + "eval_nq_pairs_runtime": 3.4001, + "eval_nq_pairs_samples_per_second": 37.646, + "eval_nq_pairs_steps_per_second": 0.588, + "step": 2200 + }, + { + "epoch": 1.7067494181536076, + "eval_trivia_pairs_loss": 0.706648588180542, + "eval_trivia_pairs_runtime": 4.5002, + "eval_trivia_pairs_samples_per_second": 28.443, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 2200 + }, + { + "epoch": 1.7067494181536076, + "eval_gooaq_pairs_loss": 0.5487090349197388, + "eval_gooaq_pairs_runtime": 1.4962, + "eval_gooaq_pairs_samples_per_second": 85.552, + "eval_gooaq_pairs_steps_per_second": 1.337, + "step": 2200 + }, + { + "epoch": 1.7067494181536076, + "eval_paws-pos_loss": 0.0318489708006382, + "eval_paws-pos_runtime": 1.0658, + "eval_paws-pos_samples_per_second": 120.102, + "eval_paws-pos_steps_per_second": 1.877, + "step": 2200 + }, + { + "epoch": 1.7067494181536076, + "eval_global_dataset_loss": 0.4033568799495697, + "eval_global_dataset_runtime": 19.4948, + "eval_global_dataset_samples_per_second": 21.339, + "eval_global_dataset_steps_per_second": 0.359, + "step": 2200 + }, + { + "epoch": 1.7075252133436774, + "grad_norm": 3.088318109512329, + "learning_rate": 7.1155122643803825e-06, + "loss": 0.0464, + "step": 2201 + }, + { + "epoch": 1.7083010085337471, + "grad_norm": 5.706811428070068, + "learning_rate": 7.097267178809253e-06, + "loss": 0.2556, + "step": 2202 + }, + { + "epoch": 1.7090768037238169, + "grad_norm": 4.838696002960205, + "learning_rate": 7.0790623083147445e-06, + "loss": 0.1116, + "step": 2203 + }, + { + "epoch": 1.7098525989138866, + "grad_norm": 10.371256828308105, + "learning_rate": 7.060897713139301e-06, + "loss": 0.5461, + "step": 2204 + }, + { + "epoch": 1.7106283941039564, + "grad_norm": 8.492655754089355, + "learning_rate": 7.0427734533920745e-06, + "loss": 0.3875, + "step": 2205 + }, + { + "epoch": 1.7114041892940264, + "grad_norm": 9.380661964416504, + "learning_rate": 7.024689589048762e-06, + "loss": 0.4466, + "step": 2206 + }, + { + "epoch": 1.7121799844840964, + "grad_norm": 5.090227127075195, + "learning_rate": 7.006646179951363e-06, + "loss": 0.1028, + "step": 2207 + }, + { + "epoch": 1.7129557796741661, + "grad_norm": 5.126872539520264, + "learning_rate": 6.988643285808034e-06, + "loss": 0.1039, + "step": 2208 + }, + { + "epoch": 1.7137315748642359, + "grad_norm": 9.21656608581543, + "learning_rate": 6.970680966192835e-06, + "loss": 0.3907, + "step": 2209 + }, + { + "epoch": 1.7145073700543056, + "grad_norm": 6.153556823730469, + "learning_rate": 6.9527592805455776e-06, + "loss": 0.2645, + "step": 2210 + }, + { + "epoch": 1.7152831652443754, + "grad_norm": 7.270167350769043, + "learning_rate": 6.934878288171593e-06, + "loss": 0.3612, + "step": 2211 + }, + { + "epoch": 1.7160589604344452, + "grad_norm": 7.52494478225708, + "learning_rate": 6.917038048241574e-06, + "loss": 0.199, + "step": 2212 + }, + { + "epoch": 1.7168347556245152, + "grad_norm": 4.230739593505859, + "learning_rate": 6.89923861979133e-06, + "loss": 0.0764, + "step": 2213 + }, + { + "epoch": 1.717610550814585, + "grad_norm": 3.2624244689941406, + "learning_rate": 6.881480061721654e-06, + "loss": 0.069, + "step": 2214 + }, + { + "epoch": 1.718386346004655, + "grad_norm": 0.3048231303691864, + "learning_rate": 6.863762432798062e-06, + "loss": 0.0032, + "step": 2215 + }, + { + "epoch": 1.7191621411947247, + "grad_norm": 5.956365585327148, + "learning_rate": 6.846085791650652e-06, + "loss": 0.1429, + "step": 2216 + }, + { + "epoch": 1.7199379363847944, + "grad_norm": 7.06648063659668, + "learning_rate": 6.828450196773871e-06, + "loss": 0.2846, + "step": 2217 + }, + { + "epoch": 1.7207137315748642, + "grad_norm": 2.3516721725463867, + "learning_rate": 6.8108557065263475e-06, + "loss": 0.0288, + "step": 2218 + }, + { + "epoch": 1.721489526764934, + "grad_norm": 7.515676975250244, + "learning_rate": 6.79330237913069e-06, + "loss": 0.3465, + "step": 2219 + }, + { + "epoch": 1.722265321955004, + "grad_norm": 0.16573144495487213, + "learning_rate": 6.775790272673294e-06, + "loss": 0.0016, + "step": 2220 + }, + { + "epoch": 1.722265321955004, + "eval_Qnli-dev_cosine_accuracy": 0.681640625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8273611068725586, + "eval_Qnli-dev_cosine_ap": 0.7253312666548202, + "eval_Qnli-dev_cosine_f1": 0.6926829268292684, + "eval_Qnli-dev_cosine_f1_threshold": 0.7055302858352661, + "eval_Qnli-dev_cosine_precision": 0.5620052770448549, + "eval_Qnli-dev_cosine_recall": 0.902542372881356, + "eval_Qnli-dev_dot_accuracy": 0.654296875, + "eval_Qnli-dev_dot_accuracy_threshold": 390.065673828125, + "eval_Qnli-dev_dot_ap": 0.6476764696800515, + "eval_Qnli-dev_dot_f1": 0.6830015313935681, + "eval_Qnli-dev_dot_f1_threshold": 315.0107421875, + "eval_Qnli-dev_dot_precision": 0.5347721822541966, + "eval_Qnli-dev_dot_recall": 0.9449152542372882, + "eval_Qnli-dev_euclidean_accuracy": 0.67578125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.690587043762207, + "eval_Qnli-dev_euclidean_ap": 0.723608071460659, + "eval_Qnli-dev_euclidean_f1": 0.6946308724832215, + "eval_Qnli-dev_euclidean_f1_threshold": 16.70089340209961, + "eval_Qnli-dev_euclidean_precision": 0.575, + "eval_Qnli-dev_euclidean_recall": 0.8771186440677966, + "eval_Qnli-dev_manhattan_accuracy": 0.677734375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 253.13702392578125, + "eval_Qnli-dev_manhattan_ap": 0.725136634158304, + "eval_Qnli-dev_manhattan_f1": 0.6911519198664441, + "eval_Qnli-dev_manhattan_f1_threshold": 306.06304931640625, + "eval_Qnli-dev_manhattan_precision": 0.5702479338842975, + "eval_Qnli-dev_manhattan_recall": 0.8771186440677966, + "eval_Qnli-dev_max_accuracy": 0.681640625, + "eval_Qnli-dev_max_accuracy_threshold": 390.065673828125, + "eval_Qnli-dev_max_ap": 0.7253312666548202, + "eval_Qnli-dev_max_f1": 0.6946308724832215, + "eval_Qnli-dev_max_f1_threshold": 315.0107421875, + "eval_Qnli-dev_max_precision": 0.575, + "eval_Qnli-dev_max_recall": 0.9449152542372882, + "eval_allNLI-dev_cosine_accuracy": 0.740234375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8340811133384705, + "eval_allNLI-dev_cosine_ap": 0.6098415137621321, + "eval_allNLI-dev_cosine_f1": 0.6344827586206897, + "eval_allNLI-dev_cosine_f1_threshold": 0.7721707224845886, + "eval_allNLI-dev_cosine_precision": 0.5267175572519084, + "eval_allNLI-dev_cosine_recall": 0.7976878612716763, + "eval_allNLI-dev_dot_accuracy": 0.69140625, + "eval_allNLI-dev_dot_accuracy_threshold": 382.8107604980469, + "eval_allNLI-dev_dot_ap": 0.5249056095199245, + "eval_allNLI-dev_dot_f1": 0.588957055214724, + "eval_allNLI-dev_dot_f1_threshold": 323.04412841796875, + "eval_allNLI-dev_dot_precision": 0.45569620253164556, + "eval_allNLI-dev_dot_recall": 0.8323699421965318, + "eval_allNLI-dev_euclidean_accuracy": 0.736328125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 11.967682838439941, + "eval_allNLI-dev_euclidean_ap": 0.6165173268463078, + "eval_allNLI-dev_euclidean_f1": 0.641860465116279, + "eval_allNLI-dev_euclidean_f1_threshold": 14.136090278625488, + "eval_allNLI-dev_euclidean_precision": 0.5369649805447471, + "eval_allNLI-dev_euclidean_recall": 0.7976878612716763, + "eval_allNLI-dev_manhattan_accuracy": 0.736328125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 226.24905395507812, + "eval_allNLI-dev_manhattan_ap": 0.6168461447991158, + "eval_allNLI-dev_manhattan_f1": 0.6407322654462242, + "eval_allNLI-dev_manhattan_f1_threshold": 261.7345886230469, + "eval_allNLI-dev_manhattan_precision": 0.5303030303030303, + "eval_allNLI-dev_manhattan_recall": 0.8092485549132948, + "eval_allNLI-dev_max_accuracy": 0.740234375, + "eval_allNLI-dev_max_accuracy_threshold": 382.8107604980469, + "eval_allNLI-dev_max_ap": 0.6168461447991158, + "eval_allNLI-dev_max_f1": 0.641860465116279, + "eval_allNLI-dev_max_f1_threshold": 323.04412841796875, + "eval_allNLI-dev_max_precision": 0.5369649805447471, + "eval_allNLI-dev_max_recall": 0.8323699421965318, + "eval_sequential_score": 0.7253312666548202, + "eval_sts-test_pearson_cosine": 0.8306760890624848, + "eval_sts-test_pearson_dot": 0.8001484328316538, + "eval_sts-test_pearson_euclidean": 0.8643988540233163, + "eval_sts-test_pearson_manhattan": 0.8612698760087611, + "eval_sts-test_pearson_max": 0.8643988540233163, + "eval_sts-test_spearman_cosine": 0.8708373500509844, + "eval_sts-test_spearman_dot": 0.7955648333654226, + "eval_sts-test_spearman_euclidean": 0.8657053648971414, + "eval_sts-test_spearman_manhattan": 0.863598430759573, + "eval_sts-test_spearman_max": 0.8708373500509844, + "eval_vitaminc-pairs_loss": 3.783086061477661, + "eval_vitaminc-pairs_runtime": 5.5533, + "eval_vitaminc-pairs_samples_per_second": 23.049, + "eval_vitaminc-pairs_steps_per_second": 0.36, + "step": 2220 + }, + { + "epoch": 1.722265321955004, + "eval_negation-triplets_loss": 1.6609992980957031, + "eval_negation-triplets_runtime": 1.1041, + "eval_negation-triplets_samples_per_second": 115.935, + "eval_negation-triplets_steps_per_second": 1.811, + "step": 2220 + }, + { + "epoch": 1.722265321955004, + "eval_scitail-pairs-pos_loss": 0.07123254239559174, + "eval_scitail-pairs-pos_runtime": 1.3605, + "eval_scitail-pairs-pos_samples_per_second": 94.085, + "eval_scitail-pairs-pos_steps_per_second": 1.47, + "step": 2220 + }, + { + "epoch": 1.722265321955004, + "eval_scitail-pairs-qa_loss": 0.0034449417144060135, + "eval_scitail-pairs-qa_runtime": 0.8192, + "eval_scitail-pairs-qa_samples_per_second": 156.252, + "eval_scitail-pairs-qa_steps_per_second": 2.441, + "step": 2220 + }, + { + "epoch": 1.722265321955004, + "eval_xsum-pairs_loss": 0.04525871202349663, + "eval_xsum-pairs_runtime": 3.9541, + "eval_xsum-pairs_samples_per_second": 32.371, + "eval_xsum-pairs_steps_per_second": 0.506, + "step": 2220 + }, + { + "epoch": 1.722265321955004, + "eval_sciq_pairs_loss": 0.035215750336647034, + "eval_sciq_pairs_runtime": 6.3464, + "eval_sciq_pairs_samples_per_second": 20.169, + "eval_sciq_pairs_steps_per_second": 0.315, + "step": 2220 + }, + { + "epoch": 1.722265321955004, + "eval_qasc_pairs_loss": 0.1782047301530838, + "eval_qasc_pairs_runtime": 0.9596, + "eval_qasc_pairs_samples_per_second": 133.383, + "eval_qasc_pairs_steps_per_second": 2.084, + "step": 2220 + }, + { + "epoch": 1.722265321955004, + "eval_openbookqa_pairs_loss": 1.0122265815734863, + "eval_openbookqa_pairs_runtime": 0.914, + "eval_openbookqa_pairs_samples_per_second": 140.046, + "eval_openbookqa_pairs_steps_per_second": 2.188, + "step": 2220 + }, + { + "epoch": 1.722265321955004, + "eval_msmarco_pairs_loss": 0.35276272892951965, + "eval_msmarco_pairs_runtime": 1.9942, + "eval_msmarco_pairs_samples_per_second": 64.185, + "eval_msmarco_pairs_steps_per_second": 1.003, + "step": 2220 + }, + { + "epoch": 1.722265321955004, + "eval_nq_pairs_loss": 0.5168060064315796, + "eval_nq_pairs_runtime": 3.3837, + "eval_nq_pairs_samples_per_second": 37.828, + "eval_nq_pairs_steps_per_second": 0.591, + "step": 2220 + }, + { + "epoch": 1.722265321955004, + "eval_trivia_pairs_loss": 0.7149307131767273, + "eval_trivia_pairs_runtime": 4.5259, + "eval_trivia_pairs_samples_per_second": 28.282, + "eval_trivia_pairs_steps_per_second": 0.442, + "step": 2220 + }, + { + "epoch": 1.722265321955004, + "eval_gooaq_pairs_loss": 0.554092288017273, + "eval_gooaq_pairs_runtime": 1.5004, + "eval_gooaq_pairs_samples_per_second": 85.312, + "eval_gooaq_pairs_steps_per_second": 1.333, + "step": 2220 + }, + { + "epoch": 1.722265321955004, + "eval_paws-pos_loss": 0.03179006651043892, + "eval_paws-pos_runtime": 1.0806, + "eval_paws-pos_samples_per_second": 118.452, + "eval_paws-pos_steps_per_second": 1.851, + "step": 2220 + }, + { + "epoch": 1.722265321955004, + "eval_global_dataset_loss": 0.38646379113197327, + "eval_global_dataset_runtime": 19.5426, + "eval_global_dataset_samples_per_second": 21.287, + "eval_global_dataset_steps_per_second": 0.358, + "step": 2220 + }, + { + "epoch": 1.7230411171450737, + "grad_norm": 4.681718349456787, + "learning_rate": 6.758319445104138e-06, + "loss": 0.2998, + "step": 2221 + }, + { + "epoch": 1.7238169123351437, + "grad_norm": 8.387868881225586, + "learning_rate": 6.740889954236618e-06, + "loss": 0.4367, + "step": 2222 + }, + { + "epoch": 1.7245927075252134, + "grad_norm": 3.1864960193634033, + "learning_rate": 6.723501857747326e-06, + "loss": 0.0745, + "step": 2223 + }, + { + "epoch": 1.7253685027152832, + "grad_norm": 2.2968714237213135, + "learning_rate": 6.706155213175887e-06, + "loss": 0.041, + "step": 2224 + }, + { + "epoch": 1.726144297905353, + "grad_norm": 11.861412048339844, + "learning_rate": 6.68885007792475e-06, + "loss": 0.4998, + "step": 2225 + }, + { + "epoch": 1.7269200930954227, + "grad_norm": 5.622377395629883, + "learning_rate": 6.671586509259012e-06, + "loss": 0.1575, + "step": 2226 + }, + { + "epoch": 1.7276958882854925, + "grad_norm": 5.34967565536499, + "learning_rate": 6.654364564306205e-06, + "loss": 0.2688, + "step": 2227 + }, + { + "epoch": 1.7284716834755625, + "grad_norm": 0.3765782415866852, + "learning_rate": 6.637184300056135e-06, + "loss": 0.0084, + "step": 2228 + }, + { + "epoch": 1.7292474786656322, + "grad_norm": 5.277986526489258, + "learning_rate": 6.620045773360669e-06, + "loss": 0.1028, + "step": 2229 + }, + { + "epoch": 1.7300232738557022, + "grad_norm": 10.694246292114258, + "learning_rate": 6.602949040933581e-06, + "loss": 0.5326, + "step": 2230 + }, + { + "epoch": 1.730799069045772, + "grad_norm": 7.292803764343262, + "learning_rate": 6.585894159350314e-06, + "loss": 0.3344, + "step": 2231 + }, + { + "epoch": 1.7315748642358417, + "grad_norm": 10.289847373962402, + "learning_rate": 6.5688811850478435e-06, + "loss": 0.446, + "step": 2232 + }, + { + "epoch": 1.7323506594259115, + "grad_norm": 4.341922760009766, + "learning_rate": 6.55191017432445e-06, + "loss": 0.0637, + "step": 2233 + }, + { + "epoch": 1.7331264546159812, + "grad_norm": 5.947787284851074, + "learning_rate": 6.534981183339562e-06, + "loss": 0.1474, + "step": 2234 + }, + { + "epoch": 1.7339022498060512, + "grad_norm": 7.201769828796387, + "learning_rate": 6.51809426811356e-06, + "loss": 0.203, + "step": 2235 + }, + { + "epoch": 1.734678044996121, + "grad_norm": 2.8695247173309326, + "learning_rate": 6.5012494845275855e-06, + "loss": 0.1049, + "step": 2236 + }, + { + "epoch": 1.735453840186191, + "grad_norm": 3.7639167308807373, + "learning_rate": 6.484446888323357e-06, + "loss": 0.101, + "step": 2237 + }, + { + "epoch": 1.7362296353762607, + "grad_norm": 4.888440132141113, + "learning_rate": 6.467686535102998e-06, + "loss": 0.1282, + "step": 2238 + }, + { + "epoch": 1.7370054305663305, + "grad_norm": 0.5418955683708191, + "learning_rate": 6.450968480328835e-06, + "loss": 0.025, + "step": 2239 + }, + { + "epoch": 1.7377812257564003, + "grad_norm": 14.925061225891113, + "learning_rate": 6.434292779323226e-06, + "loss": 1.8005, + "step": 2240 + }, + { + "epoch": 1.7377812257564003, + "eval_Qnli-dev_cosine_accuracy": 0.68359375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8109099864959717, + "eval_Qnli-dev_cosine_ap": 0.7271626333070804, + "eval_Qnli-dev_cosine_f1": 0.6915254237288135, + "eval_Qnli-dev_cosine_f1_threshold": 0.7126137018203735, + "eval_Qnli-dev_cosine_precision": 0.576271186440678, + "eval_Qnli-dev_cosine_recall": 0.864406779661017, + "eval_Qnli-dev_dot_accuracy": 0.654296875, + "eval_Qnli-dev_dot_accuracy_threshold": 380.5254821777344, + "eval_Qnli-dev_dot_ap": 0.6430362202824561, + "eval_Qnli-dev_dot_f1": 0.68259385665529, + "eval_Qnli-dev_dot_f1_threshold": 332.0799560546875, + "eval_Qnli-dev_dot_precision": 0.5714285714285714, + "eval_Qnli-dev_dot_recall": 0.847457627118644, + "eval_Qnli-dev_euclidean_accuracy": 0.67578125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.220438003540039, + "eval_Qnli-dev_euclidean_ap": 0.7268300849974108, + "eval_Qnli-dev_euclidean_f1": 0.6939443535188216, + "eval_Qnli-dev_euclidean_f1_threshold": 17.350887298583984, + "eval_Qnli-dev_euclidean_precision": 0.5653333333333334, + "eval_Qnli-dev_euclidean_recall": 0.8983050847457628, + "eval_Qnli-dev_manhattan_accuracy": 0.677734375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 253.2805938720703, + "eval_Qnli-dev_manhattan_ap": 0.7256025734573402, + "eval_Qnli-dev_manhattan_f1": 0.6939443535188216, + "eval_Qnli-dev_manhattan_f1_threshold": 316.2123107910156, + "eval_Qnli-dev_manhattan_precision": 0.5653333333333334, + "eval_Qnli-dev_manhattan_recall": 0.8983050847457628, + "eval_Qnli-dev_max_accuracy": 0.68359375, + "eval_Qnli-dev_max_accuracy_threshold": 380.5254821777344, + "eval_Qnli-dev_max_ap": 0.7271626333070804, + "eval_Qnli-dev_max_f1": 0.6939443535188216, + "eval_Qnli-dev_max_f1_threshold": 332.0799560546875, + "eval_Qnli-dev_max_precision": 0.576271186440678, + "eval_Qnli-dev_max_recall": 0.8983050847457628, + "eval_allNLI-dev_cosine_accuracy": 0.734375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8269100189208984, + "eval_allNLI-dev_cosine_ap": 0.6104376595665144, + "eval_allNLI-dev_cosine_f1": 0.6322869955156951, + "eval_allNLI-dev_cosine_f1_threshold": 0.7548410892486572, + "eval_allNLI-dev_cosine_precision": 0.5164835164835165, + "eval_allNLI-dev_cosine_recall": 0.815028901734104, + "eval_allNLI-dev_dot_accuracy": 0.697265625, + "eval_allNLI-dev_dot_accuracy_threshold": 355.90008544921875, + "eval_allNLI-dev_dot_ap": 0.5304746556582018, + "eval_allNLI-dev_dot_f1": 0.589430894308943, + "eval_allNLI-dev_dot_f1_threshold": 310.1934509277344, + "eval_allNLI-dev_dot_precision": 0.45454545454545453, + "eval_allNLI-dev_dot_recall": 0.838150289017341, + "eval_allNLI-dev_euclidean_accuracy": 0.73046875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 12.22577953338623, + "eval_allNLI-dev_euclidean_ap": 0.6157206517068545, + "eval_allNLI-dev_euclidean_f1": 0.6389496717724288, + "eval_allNLI-dev_euclidean_f1_threshold": 15.07485580444336, + "eval_allNLI-dev_euclidean_precision": 0.5140845070422535, + "eval_allNLI-dev_euclidean_recall": 0.8439306358381503, + "eval_allNLI-dev_manhattan_accuracy": 0.734375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 232.46633911132812, + "eval_allNLI-dev_manhattan_ap": 0.6170461539533935, + "eval_allNLI-dev_manhattan_f1": 0.6357308584686775, + "eval_allNLI-dev_manhattan_f1_threshold": 264.5990295410156, + "eval_allNLI-dev_manhattan_precision": 0.5310077519379846, + "eval_allNLI-dev_manhattan_recall": 0.791907514450867, + "eval_allNLI-dev_max_accuracy": 0.734375, + "eval_allNLI-dev_max_accuracy_threshold": 355.90008544921875, + "eval_allNLI-dev_max_ap": 0.6170461539533935, + "eval_allNLI-dev_max_f1": 0.6389496717724288, + "eval_allNLI-dev_max_f1_threshold": 310.1934509277344, + "eval_allNLI-dev_max_precision": 0.5310077519379846, + "eval_allNLI-dev_max_recall": 0.8439306358381503, + "eval_sequential_score": 0.7271626333070804, + "eval_sts-test_pearson_cosine": 0.8336975337402394, + "eval_sts-test_pearson_dot": 0.8087537812404475, + "eval_sts-test_pearson_euclidean": 0.8670868110665668, + "eval_sts-test_pearson_manhattan": 0.8639716895408258, + "eval_sts-test_pearson_max": 0.8670868110665668, + "eval_sts-test_spearman_cosine": 0.8717710851199161, + "eval_sts-test_spearman_dot": 0.8045200530337169, + "eval_sts-test_spearman_euclidean": 0.8664894517479802, + "eval_sts-test_spearman_manhattan": 0.8652269739128673, + "eval_sts-test_spearman_max": 0.8717710851199161, + "eval_vitaminc-pairs_loss": 3.8014743328094482, + "eval_vitaminc-pairs_runtime": 5.4679, + "eval_vitaminc-pairs_samples_per_second": 23.409, + "eval_vitaminc-pairs_steps_per_second": 0.366, + "step": 2240 + }, + { + "epoch": 1.7377812257564003, + "eval_negation-triplets_loss": 1.7285974025726318, + "eval_negation-triplets_runtime": 1.0812, + "eval_negation-triplets_samples_per_second": 118.385, + "eval_negation-triplets_steps_per_second": 1.85, + "step": 2240 + }, + { + "epoch": 1.7377812257564003, + "eval_scitail-pairs-pos_loss": 0.07390061765909195, + "eval_scitail-pairs-pos_runtime": 1.3265, + "eval_scitail-pairs-pos_samples_per_second": 96.493, + "eval_scitail-pairs-pos_steps_per_second": 1.508, + "step": 2240 + }, + { + "epoch": 1.7377812257564003, + "eval_scitail-pairs-qa_loss": 0.003566851606592536, + "eval_scitail-pairs-qa_runtime": 0.7836, + "eval_scitail-pairs-qa_samples_per_second": 163.342, + "eval_scitail-pairs-qa_steps_per_second": 2.552, + "step": 2240 + }, + { + "epoch": 1.7377812257564003, + "eval_xsum-pairs_loss": 0.041560105979442596, + "eval_xsum-pairs_runtime": 3.9313, + "eval_xsum-pairs_samples_per_second": 32.559, + "eval_xsum-pairs_steps_per_second": 0.509, + "step": 2240 + }, + { + "epoch": 1.7377812257564003, + "eval_sciq_pairs_loss": 0.037773385643959045, + "eval_sciq_pairs_runtime": 6.2845, + "eval_sciq_pairs_samples_per_second": 20.368, + "eval_sciq_pairs_steps_per_second": 0.318, + "step": 2240 + }, + { + "epoch": 1.7377812257564003, + "eval_qasc_pairs_loss": 0.16767503321170807, + "eval_qasc_pairs_runtime": 0.9658, + "eval_qasc_pairs_samples_per_second": 132.527, + "eval_qasc_pairs_steps_per_second": 2.071, + "step": 2240 + }, + { + "epoch": 1.7377812257564003, + "eval_openbookqa_pairs_loss": 1.0207995176315308, + "eval_openbookqa_pairs_runtime": 0.9075, + "eval_openbookqa_pairs_samples_per_second": 141.04, + "eval_openbookqa_pairs_steps_per_second": 2.204, + "step": 2240 + }, + { + "epoch": 1.7377812257564003, + "eval_msmarco_pairs_loss": 0.36745238304138184, + "eval_msmarco_pairs_runtime": 1.9828, + "eval_msmarco_pairs_samples_per_second": 64.556, + "eval_msmarco_pairs_steps_per_second": 1.009, + "step": 2240 + }, + { + "epoch": 1.7377812257564003, + "eval_nq_pairs_loss": 0.5393792986869812, + "eval_nq_pairs_runtime": 3.3581, + "eval_nq_pairs_samples_per_second": 38.117, + "eval_nq_pairs_steps_per_second": 0.596, + "step": 2240 + }, + { + "epoch": 1.7377812257564003, + "eval_trivia_pairs_loss": 0.7095509171485901, + "eval_trivia_pairs_runtime": 4.5113, + "eval_trivia_pairs_samples_per_second": 28.373, + "eval_trivia_pairs_steps_per_second": 0.443, + "step": 2240 + }, + { + "epoch": 1.7377812257564003, + "eval_gooaq_pairs_loss": 0.5693987011909485, + "eval_gooaq_pairs_runtime": 1.4705, + "eval_gooaq_pairs_samples_per_second": 87.043, + "eval_gooaq_pairs_steps_per_second": 1.36, + "step": 2240 + }, + { + "epoch": 1.7377812257564003, + "eval_paws-pos_loss": 0.03142717853188515, + "eval_paws-pos_runtime": 1.0411, + "eval_paws-pos_samples_per_second": 122.941, + "eval_paws-pos_steps_per_second": 1.921, + "step": 2240 + }, + { + "epoch": 1.7377812257564003, + "eval_global_dataset_loss": 0.3893611431121826, + "eval_global_dataset_runtime": 19.4688, + "eval_global_dataset_samples_per_second": 21.367, + "eval_global_dataset_steps_per_second": 0.36, + "step": 2240 + }, + { + "epoch": 1.73855702094647, + "grad_norm": 2.5471160411834717, + "learning_rate": 6.417659487268382e-06, + "loss": 0.0545, + "step": 2241 + }, + { + "epoch": 1.73933281613654, + "grad_norm": 3.7361276149749756, + "learning_rate": 6.4010686592061716e-06, + "loss": 0.0758, + "step": 2242 + }, + { + "epoch": 1.7401086113266098, + "grad_norm": 5.29871940612793, + "learning_rate": 6.384520350037936e-06, + "loss": 0.0897, + "step": 2243 + }, + { + "epoch": 1.7408844065166797, + "grad_norm": 6.844561576843262, + "learning_rate": 6.368014614524331e-06, + "loss": 0.2017, + "step": 2244 + }, + { + "epoch": 1.7416602017067495, + "grad_norm": 3.56931734085083, + "learning_rate": 6.351551507285111e-06, + "loss": 0.0589, + "step": 2245 + }, + { + "epoch": 1.7424359968968193, + "grad_norm": 3.1031105518341064, + "learning_rate": 6.335131082798993e-06, + "loss": 0.1044, + "step": 2246 + }, + { + "epoch": 1.743211792086889, + "grad_norm": 3.3660740852355957, + "learning_rate": 6.318753395403426e-06, + "loss": 0.0531, + "step": 2247 + }, + { + "epoch": 1.7439875872769588, + "grad_norm": 0.043400950729846954, + "learning_rate": 6.3024184992944535e-06, + "loss": 0.0004, + "step": 2248 + }, + { + "epoch": 1.7447633824670286, + "grad_norm": 5.363286018371582, + "learning_rate": 6.286126448526504e-06, + "loss": 0.1656, + "step": 2249 + }, + { + "epoch": 1.7455391776570985, + "grad_norm": 8.477604866027832, + "learning_rate": 6.269877297012235e-06, + "loss": 0.2754, + "step": 2250 + }, + { + "epoch": 1.7463149728471683, + "grad_norm": 9.264958381652832, + "learning_rate": 6.253671098522335e-06, + "loss": 0.3611, + "step": 2251 + }, + { + "epoch": 1.7470907680372383, + "grad_norm": 9.910209655761719, + "learning_rate": 6.237507906685368e-06, + "loss": 0.4672, + "step": 2252 + }, + { + "epoch": 1.747866563227308, + "grad_norm": 4.195262908935547, + "learning_rate": 6.221387774987566e-06, + "loss": 0.1178, + "step": 2253 + }, + { + "epoch": 1.7486423584173778, + "grad_norm": 6.8127546310424805, + "learning_rate": 6.205310756772689e-06, + "loss": 0.176, + "step": 2254 + }, + { + "epoch": 1.7494181536074476, + "grad_norm": 3.678555727005005, + "learning_rate": 6.189276905241807e-06, + "loss": 0.0772, + "step": 2255 + }, + { + "epoch": 1.7501939487975173, + "grad_norm": 13.364038467407227, + "learning_rate": 6.173286273453161e-06, + "loss": 1.4385, + "step": 2256 + }, + { + "epoch": 1.7509697439875873, + "grad_norm": 3.441401958465576, + "learning_rate": 6.157338914321966e-06, + "loss": 0.0545, + "step": 2257 + }, + { + "epoch": 1.751745539177657, + "grad_norm": 5.015955924987793, + "learning_rate": 6.141434880620249e-06, + "loss": 0.1787, + "step": 2258 + }, + { + "epoch": 1.752521334367727, + "grad_norm": 6.993857383728027, + "learning_rate": 6.12557422497665e-06, + "loss": 0.3527, + "step": 2259 + }, + { + "epoch": 1.7532971295577968, + "grad_norm": 6.21045446395874, + "learning_rate": 6.109756999876288e-06, + "loss": 0.3543, + "step": 2260 + }, + { + "epoch": 1.7532971295577968, + "eval_Qnli-dev_cosine_accuracy": 0.6875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8071527481079102, + "eval_Qnli-dev_cosine_ap": 0.7289214410536902, + "eval_Qnli-dev_cosine_f1": 0.694300518134715, + "eval_Qnli-dev_cosine_f1_threshold": 0.7246489524841309, + "eval_Qnli-dev_cosine_precision": 0.5860058309037901, + "eval_Qnli-dev_cosine_recall": 0.8516949152542372, + "eval_Qnli-dev_dot_accuracy": 0.66015625, + "eval_Qnli-dev_dot_accuracy_threshold": 379.10821533203125, + "eval_Qnli-dev_dot_ap": 0.6467386679320082, + "eval_Qnli-dev_dot_f1": 0.6812080536912752, + "eval_Qnli-dev_dot_f1_threshold": 328.3521423339844, + "eval_Qnli-dev_dot_precision": 0.5638888888888889, + "eval_Qnli-dev_dot_recall": 0.8601694915254238, + "eval_Qnli-dev_euclidean_accuracy": 0.6796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.111236572265625, + "eval_Qnli-dev_euclidean_ap": 0.7287866121535946, + "eval_Qnli-dev_euclidean_f1": 0.6978297161936562, + "eval_Qnli-dev_euclidean_f1_threshold": 16.946762084960938, + "eval_Qnli-dev_euclidean_precision": 0.5757575757575758, + "eval_Qnli-dev_euclidean_recall": 0.885593220338983, + "eval_Qnli-dev_manhattan_accuracy": 0.67578125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 244.04986572265625, + "eval_Qnli-dev_manhattan_ap": 0.72710784469852, + "eval_Qnli-dev_manhattan_f1": 0.6934189406099519, + "eval_Qnli-dev_manhattan_f1_threshold": 320.29962158203125, + "eval_Qnli-dev_manhattan_precision": 0.5581395348837209, + "eval_Qnli-dev_manhattan_recall": 0.9152542372881356, + "eval_Qnli-dev_max_accuracy": 0.6875, + "eval_Qnli-dev_max_accuracy_threshold": 379.10821533203125, + "eval_Qnli-dev_max_ap": 0.7289214410536902, + "eval_Qnli-dev_max_f1": 0.6978297161936562, + "eval_Qnli-dev_max_f1_threshold": 328.3521423339844, + "eval_Qnli-dev_max_precision": 0.5860058309037901, + "eval_Qnli-dev_max_recall": 0.9152542372881356, + "eval_allNLI-dev_cosine_accuracy": 0.73046875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8273772597312927, + "eval_allNLI-dev_cosine_ap": 0.6117315343644149, + "eval_allNLI-dev_cosine_f1": 0.6353467561521253, + "eval_allNLI-dev_cosine_f1_threshold": 0.7526835203170776, + "eval_allNLI-dev_cosine_precision": 0.5182481751824818, + "eval_allNLI-dev_cosine_recall": 0.8208092485549133, + "eval_allNLI-dev_dot_accuracy": 0.69921875, + "eval_allNLI-dev_dot_accuracy_threshold": 354.53826904296875, + "eval_allNLI-dev_dot_ap": 0.5302352728396749, + "eval_allNLI-dev_dot_f1": 0.5864077669902913, + "eval_allNLI-dev_dot_f1_threshold": 302.43927001953125, + "eval_allNLI-dev_dot_precision": 0.4415204678362573, + "eval_allNLI-dev_dot_recall": 0.8728323699421965, + "eval_allNLI-dev_euclidean_accuracy": 0.728515625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 12.168167114257812, + "eval_allNLI-dev_euclidean_ap": 0.6151682729558721, + "eval_allNLI-dev_euclidean_f1": 0.6406926406926406, + "eval_allNLI-dev_euclidean_f1_threshold": 15.183954238891602, + "eval_allNLI-dev_euclidean_precision": 0.5121107266435986, + "eval_allNLI-dev_euclidean_recall": 0.8554913294797688, + "eval_allNLI-dev_manhattan_accuracy": 0.736328125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 231.5113983154297, + "eval_allNLI-dev_manhattan_ap": 0.6170620067709898, + "eval_allNLI-dev_manhattan_f1": 0.6435185185185186, + "eval_allNLI-dev_manhattan_f1_threshold": 263.3918151855469, + "eval_allNLI-dev_manhattan_precision": 0.5366795366795367, + "eval_allNLI-dev_manhattan_recall": 0.8034682080924855, + "eval_allNLI-dev_max_accuracy": 0.736328125, + "eval_allNLI-dev_max_accuracy_threshold": 354.53826904296875, + "eval_allNLI-dev_max_ap": 0.6170620067709898, + "eval_allNLI-dev_max_f1": 0.6435185185185186, + "eval_allNLI-dev_max_f1_threshold": 302.43927001953125, + "eval_allNLI-dev_max_precision": 0.5366795366795367, + "eval_allNLI-dev_max_recall": 0.8728323699421965, + "eval_sequential_score": 0.7289214410536902, + "eval_sts-test_pearson_cosine": 0.835152844525731, + "eval_sts-test_pearson_dot": 0.8114569798150233, + "eval_sts-test_pearson_euclidean": 0.8689014516637018, + "eval_sts-test_pearson_manhattan": 0.8653302920204115, + "eval_sts-test_pearson_max": 0.8689014516637018, + "eval_sts-test_spearman_cosine": 0.8728781418242163, + "eval_sts-test_spearman_dot": 0.8066315965512088, + "eval_sts-test_spearman_euclidean": 0.8681406837908, + "eval_sts-test_spearman_manhattan": 0.8661865304596217, + "eval_sts-test_spearman_max": 0.8728781418242163, + "eval_vitaminc-pairs_loss": 3.8451149463653564, + "eval_vitaminc-pairs_runtime": 5.5134, + "eval_vitaminc-pairs_samples_per_second": 23.216, + "eval_vitaminc-pairs_steps_per_second": 0.363, + "step": 2260 + }, + { + "epoch": 1.7532971295577968, + "eval_negation-triplets_loss": 1.7055304050445557, + "eval_negation-triplets_runtime": 1.1336, + "eval_negation-triplets_samples_per_second": 112.914, + "eval_negation-triplets_steps_per_second": 1.764, + "step": 2260 + }, + { + "epoch": 1.7532971295577968, + "eval_scitail-pairs-pos_loss": 0.0811329334974289, + "eval_scitail-pairs-pos_runtime": 1.3667, + "eval_scitail-pairs-pos_samples_per_second": 93.657, + "eval_scitail-pairs-pos_steps_per_second": 1.463, + "step": 2260 + }, + { + "epoch": 1.7532971295577968, + "eval_scitail-pairs-qa_loss": 0.0040625412948429585, + "eval_scitail-pairs-qa_runtime": 0.7893, + "eval_scitail-pairs-qa_samples_per_second": 162.169, + "eval_scitail-pairs-qa_steps_per_second": 2.534, + "step": 2260 + }, + { + "epoch": 1.7532971295577968, + "eval_xsum-pairs_loss": 0.04030831158161163, + "eval_xsum-pairs_runtime": 3.9573, + "eval_xsum-pairs_samples_per_second": 32.345, + "eval_xsum-pairs_steps_per_second": 0.505, + "step": 2260 + }, + { + "epoch": 1.7532971295577968, + "eval_sciq_pairs_loss": 0.03895900771021843, + "eval_sciq_pairs_runtime": 6.3028, + "eval_sciq_pairs_samples_per_second": 20.308, + "eval_sciq_pairs_steps_per_second": 0.317, + "step": 2260 + }, + { + "epoch": 1.7532971295577968, + "eval_qasc_pairs_loss": 0.1679190695285797, + "eval_qasc_pairs_runtime": 0.9526, + "eval_qasc_pairs_samples_per_second": 134.372, + "eval_qasc_pairs_steps_per_second": 2.1, + "step": 2260 + }, + { + "epoch": 1.7532971295577968, + "eval_openbookqa_pairs_loss": 1.0061839818954468, + "eval_openbookqa_pairs_runtime": 0.9035, + "eval_openbookqa_pairs_samples_per_second": 141.663, + "eval_openbookqa_pairs_steps_per_second": 2.213, + "step": 2260 + }, + { + "epoch": 1.7532971295577968, + "eval_msmarco_pairs_loss": 0.3601939082145691, + "eval_msmarco_pairs_runtime": 1.9933, + "eval_msmarco_pairs_samples_per_second": 64.214, + "eval_msmarco_pairs_steps_per_second": 1.003, + "step": 2260 + }, + { + "epoch": 1.7532971295577968, + "eval_nq_pairs_loss": 0.5282629132270813, + "eval_nq_pairs_runtime": 3.3669, + "eval_nq_pairs_samples_per_second": 38.017, + "eval_nq_pairs_steps_per_second": 0.594, + "step": 2260 + }, + { + "epoch": 1.7532971295577968, + "eval_trivia_pairs_loss": 0.7078163623809814, + "eval_trivia_pairs_runtime": 4.5043, + "eval_trivia_pairs_samples_per_second": 28.417, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 2260 + }, + { + "epoch": 1.7532971295577968, + "eval_gooaq_pairs_loss": 0.5526044964790344, + "eval_gooaq_pairs_runtime": 1.4934, + "eval_gooaq_pairs_samples_per_second": 85.712, + "eval_gooaq_pairs_steps_per_second": 1.339, + "step": 2260 + }, + { + "epoch": 1.7532971295577968, + "eval_paws-pos_loss": 0.03132355585694313, + "eval_paws-pos_runtime": 1.0734, + "eval_paws-pos_samples_per_second": 119.251, + "eval_paws-pos_steps_per_second": 1.863, + "step": 2260 + }, + { + "epoch": 1.7532971295577968, + "eval_global_dataset_loss": 0.38668879866600037, + "eval_global_dataset_runtime": 19.5004, + "eval_global_dataset_samples_per_second": 21.333, + "eval_global_dataset_steps_per_second": 0.359, + "step": 2260 + }, + { + "epoch": 1.7540729247478666, + "grad_norm": 6.046392440795898, + "learning_rate": 6.0939832576605385e-06, + "loss": 0.2241, + "step": 2261 + }, + { + "epoch": 1.7548487199379363, + "grad_norm": 6.761354446411133, + "learning_rate": 6.07825305052691e-06, + "loss": 0.18, + "step": 2262 + }, + { + "epoch": 1.755624515128006, + "grad_norm": 9.496856689453125, + "learning_rate": 6.062566430528833e-06, + "loss": 0.3759, + "step": 2263 + }, + { + "epoch": 1.7564003103180759, + "grad_norm": 7.890334129333496, + "learning_rate": 6.046923449575511e-06, + "loss": 0.3328, + "step": 2264 + }, + { + "epoch": 1.7571761055081458, + "grad_norm": 14.678853034973145, + "learning_rate": 6.031324159431729e-06, + "loss": 1.8252, + "step": 2265 + }, + { + "epoch": 1.7579519006982158, + "grad_norm": 8.122284889221191, + "learning_rate": 6.015768611717705e-06, + "loss": 0.3433, + "step": 2266 + }, + { + "epoch": 1.7587276958882856, + "grad_norm": 1.8519117832183838, + "learning_rate": 6.000256857908894e-06, + "loss": 0.0342, + "step": 2267 + }, + { + "epoch": 1.7595034910783554, + "grad_norm": 7.873664855957031, + "learning_rate": 5.984788949335851e-06, + "loss": 0.3256, + "step": 2268 + }, + { + "epoch": 1.7602792862684251, + "grad_norm": 8.257996559143066, + "learning_rate": 5.969364937184015e-06, + "loss": 0.3362, + "step": 2269 + }, + { + "epoch": 1.7610550814584949, + "grad_norm": 4.054425239562988, + "learning_rate": 5.95398487249359e-06, + "loss": 0.1256, + "step": 2270 + }, + { + "epoch": 1.7618308766485646, + "grad_norm": 3.092909336090088, + "learning_rate": 5.938648806159331e-06, + "loss": 0.08, + "step": 2271 + }, + { + "epoch": 1.7626066718386346, + "grad_norm": 6.458966255187988, + "learning_rate": 5.923356788930414e-06, + "loss": 0.4165, + "step": 2272 + }, + { + "epoch": 1.7633824670287044, + "grad_norm": 4.16500997543335, + "learning_rate": 5.908108871410236e-06, + "loss": 0.0892, + "step": 2273 + }, + { + "epoch": 1.7641582622187744, + "grad_norm": 9.270672798156738, + "learning_rate": 5.8929051040562775e-06, + "loss": 0.3545, + "step": 2274 + }, + { + "epoch": 1.7649340574088441, + "grad_norm": 8.036388397216797, + "learning_rate": 5.877745537179902e-06, + "loss": 0.4287, + "step": 2275 + }, + { + "epoch": 1.7657098525989139, + "grad_norm": 11.200905799865723, + "learning_rate": 5.862630220946227e-06, + "loss": 0.643, + "step": 2276 + }, + { + "epoch": 1.7664856477889836, + "grad_norm": 3.4929072856903076, + "learning_rate": 5.847559205373917e-06, + "loss": 0.0887, + "step": 2277 + }, + { + "epoch": 1.7672614429790534, + "grad_norm": 4.263335704803467, + "learning_rate": 5.8325325403350536e-06, + "loss": 0.0946, + "step": 2278 + }, + { + "epoch": 1.7680372381691234, + "grad_norm": 2.291351795196533, + "learning_rate": 5.817550275554954e-06, + "loss": 0.0716, + "step": 2279 + }, + { + "epoch": 1.7688130333591932, + "grad_norm": 5.060395240783691, + "learning_rate": 5.802612460612017e-06, + "loss": 0.1916, + "step": 2280 + }, + { + "epoch": 1.7688130333591932, + "eval_Qnli-dev_cosine_accuracy": 0.685546875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.7943185567855835, + "eval_Qnli-dev_cosine_ap": 0.7240251503908031, + "eval_Qnli-dev_cosine_f1": 0.6895424836601307, + "eval_Qnli-dev_cosine_f1_threshold": 0.6986972093582153, + "eval_Qnli-dev_cosine_precision": 0.5611702127659575, + "eval_Qnli-dev_cosine_recall": 0.8940677966101694, + "eval_Qnli-dev_dot_accuracy": 0.65625, + "eval_Qnli-dev_dot_accuracy_threshold": 366.395751953125, + "eval_Qnli-dev_dot_ap": 0.6392079887988535, + "eval_Qnli-dev_dot_f1": 0.6813186813186812, + "eval_Qnli-dev_dot_f1_threshold": 313.6031188964844, + "eval_Qnli-dev_dot_precision": 0.5411471321695761, + "eval_Qnli-dev_dot_recall": 0.9194915254237288, + "eval_Qnli-dev_euclidean_accuracy": 0.6796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.396364212036133, + "eval_Qnli-dev_euclidean_ap": 0.7237697406323702, + "eval_Qnli-dev_euclidean_f1": 0.6986754966887417, + "eval_Qnli-dev_euclidean_f1_threshold": 16.969345092773438, + "eval_Qnli-dev_euclidean_precision": 0.5733695652173914, + "eval_Qnli-dev_euclidean_recall": 0.8940677966101694, + "eval_Qnli-dev_manhattan_accuracy": 0.677734375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 250.10076904296875, + "eval_Qnli-dev_manhattan_ap": 0.7252127112512553, + "eval_Qnli-dev_manhattan_f1": 0.6900958466453675, + "eval_Qnli-dev_manhattan_f1_threshold": 317.82275390625, + "eval_Qnli-dev_manhattan_precision": 0.5538461538461539, + "eval_Qnli-dev_manhattan_recall": 0.9152542372881356, + "eval_Qnli-dev_max_accuracy": 0.685546875, + "eval_Qnli-dev_max_accuracy_threshold": 366.395751953125, + "eval_Qnli-dev_max_ap": 0.7252127112512553, + "eval_Qnli-dev_max_f1": 0.6986754966887417, + "eval_Qnli-dev_max_f1_threshold": 317.82275390625, + "eval_Qnli-dev_max_precision": 0.5733695652173914, + "eval_Qnli-dev_max_recall": 0.9194915254237288, + "eval_allNLI-dev_cosine_accuracy": 0.734375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.828066885471344, + "eval_allNLI-dev_cosine_ap": 0.6125518735932471, + "eval_allNLI-dev_cosine_f1": 0.6306306306306307, + "eval_allNLI-dev_cosine_f1_threshold": 0.7572066187858582, + "eval_allNLI-dev_cosine_precision": 0.5166051660516605, + "eval_allNLI-dev_cosine_recall": 0.8092485549132948, + "eval_allNLI-dev_dot_accuracy": 0.701171875, + "eval_allNLI-dev_dot_accuracy_threshold": 358.7825622558594, + "eval_allNLI-dev_dot_ap": 0.5297035312799973, + "eval_allNLI-dev_dot_f1": 0.5925925925925926, + "eval_allNLI-dev_dot_f1_threshold": 334.0279235839844, + "eval_allNLI-dev_dot_precision": 0.4942084942084942, + "eval_allNLI-dev_dot_recall": 0.7398843930635838, + "eval_allNLI-dev_euclidean_accuracy": 0.73046875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 12.589889526367188, + "eval_allNLI-dev_euclidean_ap": 0.6148347374518279, + "eval_allNLI-dev_euclidean_f1": 0.6414253897550111, + "eval_allNLI-dev_euclidean_f1_threshold": 14.703290939331055, + "eval_allNLI-dev_euclidean_precision": 0.5217391304347826, + "eval_allNLI-dev_euclidean_recall": 0.8323699421965318, + "eval_allNLI-dev_manhattan_accuracy": 0.73828125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 232.22227478027344, + "eval_allNLI-dev_manhattan_ap": 0.6166607340624967, + "eval_allNLI-dev_manhattan_f1": 0.6367713004484306, + "eval_allNLI-dev_manhattan_f1_threshold": 268.90460205078125, + "eval_allNLI-dev_manhattan_precision": 0.5201465201465202, + "eval_allNLI-dev_manhattan_recall": 0.8208092485549133, + "eval_allNLI-dev_max_accuracy": 0.73828125, + "eval_allNLI-dev_max_accuracy_threshold": 358.7825622558594, + "eval_allNLI-dev_max_ap": 0.6166607340624967, + "eval_allNLI-dev_max_f1": 0.6414253897550111, + "eval_allNLI-dev_max_f1_threshold": 334.0279235839844, + "eval_allNLI-dev_max_precision": 0.5217391304347826, + "eval_allNLI-dev_max_recall": 0.8323699421965318, + "eval_sequential_score": 0.7252127112512553, + "eval_sts-test_pearson_cosine": 0.836779946984835, + "eval_sts-test_pearson_dot": 0.8117525296913375, + "eval_sts-test_pearson_euclidean": 0.8717266474320228, + "eval_sts-test_pearson_manhattan": 0.8681800234844868, + "eval_sts-test_pearson_max": 0.8717266474320228, + "eval_sts-test_spearman_cosine": 0.876974896048292, + "eval_sts-test_spearman_dot": 0.8095139805461674, + "eval_sts-test_spearman_euclidean": 0.8711759828459331, + "eval_sts-test_spearman_manhattan": 0.8698704990486232, + "eval_sts-test_spearman_max": 0.876974896048292, + "eval_vitaminc-pairs_loss": 3.8584303855895996, + "eval_vitaminc-pairs_runtime": 5.5209, + "eval_vitaminc-pairs_samples_per_second": 23.185, + "eval_vitaminc-pairs_steps_per_second": 0.362, + "step": 2280 + }, + { + "epoch": 1.7688130333591932, + "eval_negation-triplets_loss": 1.6840014457702637, + "eval_negation-triplets_runtime": 1.0991, + "eval_negation-triplets_samples_per_second": 116.463, + "eval_negation-triplets_steps_per_second": 1.82, + "step": 2280 + }, + { + "epoch": 1.7688130333591932, + "eval_scitail-pairs-pos_loss": 0.09057288616895676, + "eval_scitail-pairs-pos_runtime": 1.3472, + "eval_scitail-pairs-pos_samples_per_second": 95.015, + "eval_scitail-pairs-pos_steps_per_second": 1.485, + "step": 2280 + }, + { + "epoch": 1.7688130333591932, + "eval_scitail-pairs-qa_loss": 0.004001098219305277, + "eval_scitail-pairs-qa_runtime": 0.8167, + "eval_scitail-pairs-qa_samples_per_second": 156.732, + "eval_scitail-pairs-qa_steps_per_second": 2.449, + "step": 2280 + }, + { + "epoch": 1.7688130333591932, + "eval_xsum-pairs_loss": 0.039700236171483994, + "eval_xsum-pairs_runtime": 3.9395, + "eval_xsum-pairs_samples_per_second": 32.492, + "eval_xsum-pairs_steps_per_second": 0.508, + "step": 2280 + }, + { + "epoch": 1.7688130333591932, + "eval_sciq_pairs_loss": 0.03681030124425888, + "eval_sciq_pairs_runtime": 6.2977, + "eval_sciq_pairs_samples_per_second": 20.325, + "eval_sciq_pairs_steps_per_second": 0.318, + "step": 2280 + }, + { + "epoch": 1.7688130333591932, + "eval_qasc_pairs_loss": 0.17691761255264282, + "eval_qasc_pairs_runtime": 0.9451, + "eval_qasc_pairs_samples_per_second": 135.436, + "eval_qasc_pairs_steps_per_second": 2.116, + "step": 2280 + }, + { + "epoch": 1.7688130333591932, + "eval_openbookqa_pairs_loss": 1.0082783699035645, + "eval_openbookqa_pairs_runtime": 0.9003, + "eval_openbookqa_pairs_samples_per_second": 142.169, + "eval_openbookqa_pairs_steps_per_second": 2.221, + "step": 2280 + }, + { + "epoch": 1.7688130333591932, + "eval_msmarco_pairs_loss": 0.34672462940216064, + "eval_msmarco_pairs_runtime": 1.9896, + "eval_msmarco_pairs_samples_per_second": 64.334, + "eval_msmarco_pairs_steps_per_second": 1.005, + "step": 2280 + }, + { + "epoch": 1.7688130333591932, + "eval_nq_pairs_loss": 0.5206706523895264, + "eval_nq_pairs_runtime": 3.3846, + "eval_nq_pairs_samples_per_second": 37.818, + "eval_nq_pairs_steps_per_second": 0.591, + "step": 2280 + }, + { + "epoch": 1.7688130333591932, + "eval_trivia_pairs_loss": 0.7025155425071716, + "eval_trivia_pairs_runtime": 4.5069, + "eval_trivia_pairs_samples_per_second": 28.401, + "eval_trivia_pairs_steps_per_second": 0.444, + "step": 2280 + }, + { + "epoch": 1.7688130333591932, + "eval_gooaq_pairs_loss": 0.5543738007545471, + "eval_gooaq_pairs_runtime": 1.4902, + "eval_gooaq_pairs_samples_per_second": 85.897, + "eval_gooaq_pairs_steps_per_second": 1.342, + "step": 2280 + }, + { + "epoch": 1.7688130333591932, + "eval_paws-pos_loss": 0.030947351828217506, + "eval_paws-pos_runtime": 1.0662, + "eval_paws-pos_samples_per_second": 120.051, + "eval_paws-pos_steps_per_second": 1.876, + "step": 2280 + }, + { + "epoch": 1.7688130333591932, + "eval_global_dataset_loss": 0.387752503156662, + "eval_global_dataset_runtime": 19.498, + "eval_global_dataset_samples_per_second": 21.336, + "eval_global_dataset_steps_per_second": 0.359, + "step": 2280 + }, + { + "epoch": 1.7695888285492631, + "grad_norm": 1.9592446088790894, + "learning_rate": 5.787719144937529e-06, + "loss": 0.0553, + "step": 2281 + }, + { + "epoch": 1.770364623739333, + "grad_norm": 5.08129358291626, + "learning_rate": 5.772870377815542e-06, + "loss": 0.2103, + "step": 2282 + }, + { + "epoch": 1.7711404189294027, + "grad_norm": 6.345480442047119, + "learning_rate": 5.758066208382678e-06, + "loss": 0.1743, + "step": 2283 + }, + { + "epoch": 1.7719162141194724, + "grad_norm": 8.238494873046875, + "learning_rate": 5.743306685627989e-06, + "loss": 0.4008, + "step": 2284 + }, + { + "epoch": 1.7726920093095422, + "grad_norm": 2.5731921195983887, + "learning_rate": 5.728591858392779e-06, + "loss": 0.0345, + "step": 2285 + }, + { + "epoch": 1.773467804499612, + "grad_norm": 1.9975038766860962, + "learning_rate": 5.713921775370457e-06, + "loss": 0.0368, + "step": 2286 + }, + { + "epoch": 1.774243599689682, + "grad_norm": 7.497552394866943, + "learning_rate": 5.69929648510635e-06, + "loss": 0.3037, + "step": 2287 + }, + { + "epoch": 1.775019394879752, + "grad_norm": 6.714852333068848, + "learning_rate": 5.684716035997575e-06, + "loss": 0.251, + "step": 2288 + }, + { + "epoch": 1.7757951900698217, + "grad_norm": 2.3582918643951416, + "learning_rate": 5.670180476292852e-06, + "loss": 0.0423, + "step": 2289 + }, + { + "epoch": 1.7765709852598914, + "grad_norm": 5.7815093994140625, + "learning_rate": 5.655689854092375e-06, + "loss": 0.1528, + "step": 2290 + }, + { + "epoch": 1.7773467804499612, + "grad_norm": 8.394386291503906, + "learning_rate": 5.641244217347607e-06, + "loss": 0.2422, + "step": 2291 + }, + { + "epoch": 1.778122575640031, + "grad_norm": 4.447476387023926, + "learning_rate": 5.626843613861172e-06, + "loss": 0.0576, + "step": 2292 + }, + { + "epoch": 1.7788983708301007, + "grad_norm": 3.693026065826416, + "learning_rate": 5.612488091286652e-06, + "loss": 0.0926, + "step": 2293 + }, + { + "epoch": 1.7796741660201707, + "grad_norm": 9.303796768188477, + "learning_rate": 5.5981776971284686e-06, + "loss": 0.4563, + "step": 2294 + }, + { + "epoch": 1.7804499612102405, + "grad_norm": 4.628039360046387, + "learning_rate": 5.583912478741689e-06, + "loss": 0.1998, + "step": 2295 + }, + { + "epoch": 1.7812257564003104, + "grad_norm": 6.5070061683654785, + "learning_rate": 5.56969248333191e-06, + "loss": 0.2243, + "step": 2296 + }, + { + "epoch": 1.7820015515903802, + "grad_norm": 0.07618746161460876, + "learning_rate": 5.555517757955058e-06, + "loss": 0.0006, + "step": 2297 + }, + { + "epoch": 1.78277734678045, + "grad_norm": 10.463044166564941, + "learning_rate": 5.541388349517267e-06, + "loss": 0.4765, + "step": 2298 + }, + { + "epoch": 1.7835531419705197, + "grad_norm": 11.366082191467285, + "learning_rate": 5.527304304774702e-06, + "loss": 0.4909, + "step": 2299 + }, + { + "epoch": 1.7843289371605895, + "grad_norm": 9.398041725158691, + "learning_rate": 5.513265670333422e-06, + "loss": 0.4393, + "step": 2300 + }, + { + "epoch": 1.7843289371605895, + "eval_Qnli-dev_cosine_accuracy": 0.6796875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.7942582964897156, + "eval_Qnli-dev_cosine_ap": 0.721690179586384, + "eval_Qnli-dev_cosine_f1": 0.6928571428571428, + "eval_Qnli-dev_cosine_f1_threshold": 0.734821081161499, + "eval_Qnli-dev_cosine_precision": 0.5987654320987654, + "eval_Qnli-dev_cosine_recall": 0.8220338983050848, + "eval_Qnli-dev_dot_accuracy": 0.658203125, + "eval_Qnli-dev_dot_accuracy_threshold": 361.1005554199219, + "eval_Qnli-dev_dot_ap": 0.6332421423382695, + "eval_Qnli-dev_dot_f1": 0.6838709677419356, + "eval_Qnli-dev_dot_f1_threshold": 322.07086181640625, + "eval_Qnli-dev_dot_precision": 0.5520833333333334, + "eval_Qnli-dev_dot_recall": 0.8983050847457628, + "eval_Qnli-dev_euclidean_accuracy": 0.677734375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 14.433286666870117, + "eval_Qnli-dev_euclidean_ap": 0.721752901979977, + "eval_Qnli-dev_euclidean_f1": 0.6957983193277311, + "eval_Qnli-dev_euclidean_f1_threshold": 16.961063385009766, + "eval_Qnli-dev_euclidean_precision": 0.5766016713091922, + "eval_Qnli-dev_euclidean_recall": 0.8771186440677966, + "eval_Qnli-dev_manhattan_accuracy": 0.67578125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 249.80099487304688, + "eval_Qnli-dev_manhattan_ap": 0.7229679701178839, + "eval_Qnli-dev_manhattan_f1": 0.6881028938906752, + "eval_Qnli-dev_manhattan_f1_threshold": 318.8553771972656, + "eval_Qnli-dev_manhattan_precision": 0.5544041450777202, + "eval_Qnli-dev_manhattan_recall": 0.9067796610169492, + "eval_Qnli-dev_max_accuracy": 0.6796875, + "eval_Qnli-dev_max_accuracy_threshold": 361.1005554199219, + "eval_Qnli-dev_max_ap": 0.7229679701178839, + "eval_Qnli-dev_max_f1": 0.6957983193277311, + "eval_Qnli-dev_max_f1_threshold": 322.07086181640625, + "eval_Qnli-dev_max_precision": 0.5987654320987654, + "eval_Qnli-dev_max_recall": 0.9067796610169492, + "eval_allNLI-dev_cosine_accuracy": 0.734375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.826436460018158, + "eval_allNLI-dev_cosine_ap": 0.6123077878112371, + "eval_allNLI-dev_cosine_f1": 0.6294642857142856, + "eval_allNLI-dev_cosine_f1_threshold": 0.7496253252029419, + "eval_allNLI-dev_cosine_precision": 0.5127272727272727, + "eval_allNLI-dev_cosine_recall": 0.815028901734104, + "eval_allNLI-dev_dot_accuracy": 0.70703125, + "eval_allNLI-dev_dot_accuracy_threshold": 369.2552795410156, + "eval_allNLI-dev_dot_ap": 0.5269394782272924, + "eval_allNLI-dev_dot_f1": 0.5844748858447488, + "eval_allNLI-dev_dot_f1_threshold": 330.62750244140625, + "eval_allNLI-dev_dot_precision": 0.4830188679245283, + "eval_allNLI-dev_dot_recall": 0.7398843930635838, + "eval_allNLI-dev_euclidean_accuracy": 0.736328125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 12.827374458312988, + "eval_allNLI-dev_euclidean_ap": 0.6147864902995601, + "eval_allNLI-dev_euclidean_f1": 0.6471910112359551, + "eval_allNLI-dev_euclidean_f1_threshold": 14.888999938964844, + "eval_allNLI-dev_euclidean_precision": 0.5294117647058824, + "eval_allNLI-dev_euclidean_recall": 0.8323699421965318, + "eval_allNLI-dev_manhattan_accuracy": 0.736328125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 235.92343139648438, + "eval_allNLI-dev_manhattan_ap": 0.6150835597668937, + "eval_allNLI-dev_manhattan_f1": 0.6455981941309255, + "eval_allNLI-dev_manhattan_f1_threshold": 273.2347106933594, + "eval_allNLI-dev_manhattan_precision": 0.5296296296296297, + "eval_allNLI-dev_manhattan_recall": 0.8265895953757225, + "eval_allNLI-dev_max_accuracy": 0.736328125, + "eval_allNLI-dev_max_accuracy_threshold": 369.2552795410156, + "eval_allNLI-dev_max_ap": 0.6150835597668937, + "eval_allNLI-dev_max_f1": 0.6471910112359551, + "eval_allNLI-dev_max_f1_threshold": 330.62750244140625, + "eval_allNLI-dev_max_precision": 0.5296296296296297, + "eval_allNLI-dev_max_recall": 0.8323699421965318, + "eval_sequential_score": 0.7229679701178839, + "eval_sts-test_pearson_cosine": 0.838969359527302, + "eval_sts-test_pearson_dot": 0.8108614204914337, + "eval_sts-test_pearson_euclidean": 0.8739404125066487, + "eval_sts-test_pearson_manhattan": 0.870450000114995, + "eval_sts-test_pearson_max": 0.8739404125066487, + "eval_sts-test_spearman_cosine": 0.8785229316823433, + "eval_sts-test_spearman_dot": 0.8081709056892886, + "eval_sts-test_spearman_euclidean": 0.8724554661603747, + "eval_sts-test_spearman_manhattan": 0.8713478724409375, + "eval_sts-test_spearman_max": 0.8785229316823433, + "eval_vitaminc-pairs_loss": 3.8593368530273438, + "eval_vitaminc-pairs_runtime": 5.5175, + "eval_vitaminc-pairs_samples_per_second": 23.199, + "eval_vitaminc-pairs_steps_per_second": 0.362, + "step": 2300 + }, + { + "epoch": 1.7843289371605895, + "eval_negation-triplets_loss": 1.695976734161377, + "eval_negation-triplets_runtime": 1.091, + "eval_negation-triplets_samples_per_second": 117.328, + "eval_negation-triplets_steps_per_second": 1.833, + "step": 2300 + }, + { + "epoch": 1.7843289371605895, + "eval_scitail-pairs-pos_loss": 0.10571404546499252, + "eval_scitail-pairs-pos_runtime": 1.3474, + "eval_scitail-pairs-pos_samples_per_second": 95.001, + "eval_scitail-pairs-pos_steps_per_second": 1.484, + "step": 2300 + }, + { + "epoch": 1.7843289371605895, + "eval_scitail-pairs-qa_loss": 0.0035468959249556065, + "eval_scitail-pairs-qa_runtime": 0.7929, + "eval_scitail-pairs-qa_samples_per_second": 161.436, + "eval_scitail-pairs-qa_steps_per_second": 2.522, + "step": 2300 + }, + { + "epoch": 1.7843289371605895, + "eval_xsum-pairs_loss": 0.04150925576686859, + "eval_xsum-pairs_runtime": 3.9432, + "eval_xsum-pairs_samples_per_second": 32.461, + "eval_xsum-pairs_steps_per_second": 0.507, + "step": 2300 + }, + { + "epoch": 1.7843289371605895, + "eval_sciq_pairs_loss": 0.037853583693504333, + "eval_sciq_pairs_runtime": 6.3006, + "eval_sciq_pairs_samples_per_second": 20.316, + "eval_sciq_pairs_steps_per_second": 0.317, + "step": 2300 + }, + { + "epoch": 1.7843289371605895, + "eval_qasc_pairs_loss": 0.17679226398468018, + "eval_qasc_pairs_runtime": 0.9455, + "eval_qasc_pairs_samples_per_second": 135.374, + "eval_qasc_pairs_steps_per_second": 2.115, + "step": 2300 + }, + { + "epoch": 1.7843289371605895, + "eval_openbookqa_pairs_loss": 1.0365798473358154, + "eval_openbookqa_pairs_runtime": 0.9083, + "eval_openbookqa_pairs_samples_per_second": 140.924, + "eval_openbookqa_pairs_steps_per_second": 2.202, + "step": 2300 + }, + { + "epoch": 1.7843289371605895, + "eval_msmarco_pairs_loss": 0.3410050868988037, + "eval_msmarco_pairs_runtime": 1.9864, + "eval_msmarco_pairs_samples_per_second": 64.437, + "eval_msmarco_pairs_steps_per_second": 1.007, + "step": 2300 + }, + { + "epoch": 1.7843289371605895, + "eval_nq_pairs_loss": 0.5360149145126343, + "eval_nq_pairs_runtime": 3.3697, + "eval_nq_pairs_samples_per_second": 37.986, + "eval_nq_pairs_steps_per_second": 0.594, + "step": 2300 + }, + { + "epoch": 1.7843289371605895, + "eval_trivia_pairs_loss": 0.6922394633293152, + "eval_trivia_pairs_runtime": 4.5156, + "eval_trivia_pairs_samples_per_second": 28.346, + "eval_trivia_pairs_steps_per_second": 0.443, + "step": 2300 + }, + { + "epoch": 1.7843289371605895, + "eval_gooaq_pairs_loss": 0.5462455749511719, + "eval_gooaq_pairs_runtime": 1.4877, + "eval_gooaq_pairs_samples_per_second": 86.037, + "eval_gooaq_pairs_steps_per_second": 1.344, + "step": 2300 + }, + { + "epoch": 1.7843289371605895, + "eval_paws-pos_loss": 0.030734632164239883, + "eval_paws-pos_runtime": 1.0497, + "eval_paws-pos_samples_per_second": 121.939, + "eval_paws-pos_steps_per_second": 1.905, + "step": 2300 + }, + { + "epoch": 1.7843289371605895, + "eval_global_dataset_loss": 0.3971985876560211, + "eval_global_dataset_runtime": 19.4825, + "eval_global_dataset_samples_per_second": 21.353, + "eval_global_dataset_steps_per_second": 0.359, + "step": 2300 + }, + { + "epoch": 1.7851047323506595, + "grad_norm": 8.165913581848145, + "learning_rate": 5.499272492649213e-06, + "loss": 0.343, + "step": 2301 + }, + { + "epoch": 1.7858805275407292, + "grad_norm": 2.892871141433716, + "learning_rate": 5.485324818027441e-06, + "loss": 0.0522, + "step": 2302 + }, + { + "epoch": 1.7866563227307992, + "grad_norm": 7.177694320678711, + "learning_rate": 5.47142269262289e-06, + "loss": 0.142, + "step": 2303 + }, + { + "epoch": 1.787432117920869, + "grad_norm": 0.8981366753578186, + "learning_rate": 5.457566162439622e-06, + "loss": 0.007, + "step": 2304 + }, + { + "epoch": 1.7882079131109387, + "grad_norm": 2.9240400791168213, + "learning_rate": 5.44375527333081e-06, + "loss": 0.0585, + "step": 2305 + }, + { + "epoch": 1.7889837083010085, + "grad_norm": 7.580024719238281, + "learning_rate": 5.429990070998604e-06, + "loss": 0.2579, + "step": 2306 + }, + { + "epoch": 1.7897595034910783, + "grad_norm": 7.820934772491455, + "learning_rate": 5.416270600993964e-06, + "loss": 0.2591, + "step": 2307 + }, + { + "epoch": 1.790535298681148, + "grad_norm": 5.22636604309082, + "learning_rate": 5.402596908716521e-06, + "loss": 0.2753, + "step": 2308 + }, + { + "epoch": 1.791311093871218, + "grad_norm": 3.3353359699249268, + "learning_rate": 5.38896903941441e-06, + "loss": 0.0503, + "step": 2309 + }, + { + "epoch": 1.7920868890612878, + "grad_norm": 5.063332557678223, + "learning_rate": 5.375387038184147e-06, + "loss": 0.1909, + "step": 2310 + }, + { + "epoch": 1.7928626842513578, + "grad_norm": 9.62601089477539, + "learning_rate": 5.361850949970443e-06, + "loss": 0.5584, + "step": 2311 + }, + { + "epoch": 1.7936384794414275, + "grad_norm": 6.017707347869873, + "learning_rate": 5.348360819566106e-06, + "loss": 0.2152, + "step": 2312 + }, + { + "epoch": 1.7944142746314973, + "grad_norm": 0.6999373435974121, + "learning_rate": 5.334916691611835e-06, + "loss": 0.0302, + "step": 2313 + }, + { + "epoch": 1.795190069821567, + "grad_norm": 5.963460445404053, + "learning_rate": 5.3215186105961165e-06, + "loss": 0.348, + "step": 2314 + }, + { + "epoch": 1.7959658650116368, + "grad_norm": 0.8628275990486145, + "learning_rate": 5.3081666208550525e-06, + "loss": 0.0079, + "step": 2315 + }, + { + "epoch": 1.7967416602017068, + "grad_norm": 2.4215261936187744, + "learning_rate": 5.294860766572231e-06, + "loss": 0.0399, + "step": 2316 + }, + { + "epoch": 1.7975174553917765, + "grad_norm": 6.199039459228516, + "learning_rate": 5.281601091778566e-06, + "loss": 0.2051, + "step": 2317 + }, + { + "epoch": 1.7982932505818465, + "grad_norm": 12.852889060974121, + "learning_rate": 5.268387640352162e-06, + "loss": 0.7093, + "step": 2318 + }, + { + "epoch": 1.7990690457719163, + "grad_norm": 11.801790237426758, + "learning_rate": 5.2552204560181525e-06, + "loss": 0.3956, + "step": 2319 + }, + { + "epoch": 1.799844840961986, + "grad_norm": 3.3729662895202637, + "learning_rate": 5.242099582348584e-06, + "loss": 0.0822, + "step": 2320 + }, + { + "epoch": 1.799844840961986, + "eval_Qnli-dev_cosine_accuracy": 0.677734375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8217729330062866, + "eval_Qnli-dev_cosine_ap": 0.7211717123810969, + "eval_Qnli-dev_cosine_f1": 0.6918238993710691, + "eval_Qnli-dev_cosine_f1_threshold": 0.6766316890716553, + "eval_Qnli-dev_cosine_precision": 0.55, + "eval_Qnli-dev_cosine_recall": 0.9322033898305084, + "eval_Qnli-dev_dot_accuracy": 0.6484375, + "eval_Qnli-dev_dot_accuracy_threshold": 355.84197998046875, + "eval_Qnli-dev_dot_ap": 0.6392926584533516, + "eval_Qnli-dev_dot_f1": 0.6788990825688074, + "eval_Qnli-dev_dot_f1_threshold": 304.874755859375, + "eval_Qnli-dev_dot_precision": 0.5311004784688995, + "eval_Qnli-dev_dot_recall": 0.940677966101695, + "eval_Qnli-dev_euclidean_accuracy": 0.671875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 13.185789108276367, + "eval_Qnli-dev_euclidean_ap": 0.7196141268390591, + "eval_Qnli-dev_euclidean_f1": 0.6975206611570248, + "eval_Qnli-dev_euclidean_f1_threshold": 17.1843318939209, + "eval_Qnli-dev_euclidean_precision": 0.5718157181571816, + "eval_Qnli-dev_euclidean_recall": 0.8940677966101694, + "eval_Qnli-dev_manhattan_accuracy": 0.6796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 252.36135864257812, + "eval_Qnli-dev_manhattan_ap": 0.7199603742243552, + "eval_Qnli-dev_manhattan_f1": 0.6902086677367576, + "eval_Qnli-dev_manhattan_f1_threshold": 321.187744140625, + "eval_Qnli-dev_manhattan_precision": 0.5555555555555556, + "eval_Qnli-dev_manhattan_recall": 0.9110169491525424, + "eval_Qnli-dev_max_accuracy": 0.6796875, + "eval_Qnli-dev_max_accuracy_threshold": 355.84197998046875, + "eval_Qnli-dev_max_ap": 0.7211717123810969, + "eval_Qnli-dev_max_f1": 0.6975206611570248, + "eval_Qnli-dev_max_f1_threshold": 321.187744140625, + "eval_Qnli-dev_max_precision": 0.5718157181571816, + "eval_Qnli-dev_max_recall": 0.940677966101695, + "eval_allNLI-dev_cosine_accuracy": 0.734375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8276818990707397, + "eval_allNLI-dev_cosine_ap": 0.6090357221170635, + "eval_allNLI-dev_cosine_f1": 0.6280623608017818, + "eval_allNLI-dev_cosine_f1_threshold": 0.752535879611969, + "eval_allNLI-dev_cosine_precision": 0.5108695652173914, + "eval_allNLI-dev_cosine_recall": 0.815028901734104, + "eval_allNLI-dev_dot_accuracy": 0.697265625, + "eval_allNLI-dev_dot_accuracy_threshold": 378.18780517578125, + "eval_allNLI-dev_dot_ap": 0.5214706796665698, + "eval_allNLI-dev_dot_f1": 0.5809768637532133, + "eval_allNLI-dev_dot_f1_threshold": 354.50445556640625, + "eval_allNLI-dev_dot_precision": 0.5231481481481481, + "eval_allNLI-dev_dot_recall": 0.653179190751445, + "eval_allNLI-dev_euclidean_accuracy": 0.7265625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 12.777070999145508, + "eval_allNLI-dev_euclidean_ap": 0.6142501812814289, + "eval_allNLI-dev_euclidean_f1": 0.6410835214446953, + "eval_allNLI-dev_euclidean_f1_threshold": 14.825325012207031, + "eval_allNLI-dev_euclidean_precision": 0.5259259259259259, + "eval_allNLI-dev_euclidean_recall": 0.8208092485549133, + "eval_allNLI-dev_manhattan_accuracy": 0.73046875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 233.76202392578125, + "eval_allNLI-dev_manhattan_ap": 0.6152203204639695, + "eval_allNLI-dev_manhattan_f1": 0.6412556053811659, + "eval_allNLI-dev_manhattan_f1_threshold": 273.9834289550781, + "eval_allNLI-dev_manhattan_precision": 0.5238095238095238, + "eval_allNLI-dev_manhattan_recall": 0.8265895953757225, + "eval_allNLI-dev_max_accuracy": 0.734375, + "eval_allNLI-dev_max_accuracy_threshold": 378.18780517578125, + "eval_allNLI-dev_max_ap": 0.6152203204639695, + "eval_allNLI-dev_max_f1": 0.6412556053811659, + "eval_allNLI-dev_max_f1_threshold": 354.50445556640625, + "eval_allNLI-dev_max_precision": 0.5259259259259259, + "eval_allNLI-dev_max_recall": 0.8265895953757225, + "eval_sequential_score": 0.7211717123810969, + "eval_sts-test_pearson_cosine": 0.8379717757574963, + "eval_sts-test_pearson_dot": 0.8070552903221473, + "eval_sts-test_pearson_euclidean": 0.8728635232348738, + "eval_sts-test_pearson_manhattan": 0.8689292237453781, + "eval_sts-test_pearson_max": 0.8728635232348738, + "eval_sts-test_spearman_cosine": 0.8779696718377614, + "eval_sts-test_spearman_dot": 0.8050180898217418, + "eval_sts-test_spearman_euclidean": 0.8721625691545679, + "eval_sts-test_spearman_manhattan": 0.8700785029115704, + "eval_sts-test_spearman_max": 0.8779696718377614, + "eval_vitaminc-pairs_loss": 3.8130831718444824, + "eval_vitaminc-pairs_runtime": 5.5129, + "eval_vitaminc-pairs_samples_per_second": 23.218, + "eval_vitaminc-pairs_steps_per_second": 0.363, + "step": 2320 + }, + { + "epoch": 1.799844840961986, + "eval_negation-triplets_loss": 1.6576777696609497, + "eval_negation-triplets_runtime": 1.0905, + "eval_negation-triplets_samples_per_second": 117.375, + "eval_negation-triplets_steps_per_second": 1.834, + "step": 2320 + }, + { + "epoch": 1.799844840961986, + "eval_scitail-pairs-pos_loss": 0.0977407768368721, + "eval_scitail-pairs-pos_runtime": 1.3465, + "eval_scitail-pairs-pos_samples_per_second": 95.064, + "eval_scitail-pairs-pos_steps_per_second": 1.485, + "step": 2320 + }, + { + "epoch": 1.799844840961986, + "eval_scitail-pairs-qa_loss": 0.0036120759323239326, + "eval_scitail-pairs-qa_runtime": 0.7881, + "eval_scitail-pairs-qa_samples_per_second": 162.424, + "eval_scitail-pairs-qa_steps_per_second": 2.538, + "step": 2320 + }, + { + "epoch": 1.799844840961986, + "eval_xsum-pairs_loss": 0.05010436847805977, + "eval_xsum-pairs_runtime": 3.9363, + "eval_xsum-pairs_samples_per_second": 32.518, + "eval_xsum-pairs_steps_per_second": 0.508, + "step": 2320 + }, + { + "epoch": 1.799844840961986, + "eval_sciq_pairs_loss": 0.037255626171827316, + "eval_sciq_pairs_runtime": 6.0839, + "eval_sciq_pairs_samples_per_second": 21.039, + "eval_sciq_pairs_steps_per_second": 0.329, + "step": 2320 + }, + { + "epoch": 1.799844840961986, + "eval_qasc_pairs_loss": 0.18382717669010162, + "eval_qasc_pairs_runtime": 0.9699, + "eval_qasc_pairs_samples_per_second": 131.968, + "eval_qasc_pairs_steps_per_second": 2.062, + "step": 2320 + }, + { + "epoch": 1.799844840961986, + "eval_openbookqa_pairs_loss": 1.0195646286010742, + "eval_openbookqa_pairs_runtime": 0.8947, + "eval_openbookqa_pairs_samples_per_second": 143.058, + "eval_openbookqa_pairs_steps_per_second": 2.235, + "step": 2320 + }, + { + "epoch": 1.799844840961986, + "eval_msmarco_pairs_loss": 0.3330722153186798, + "eval_msmarco_pairs_runtime": 1.9845, + "eval_msmarco_pairs_samples_per_second": 64.5, + "eval_msmarco_pairs_steps_per_second": 1.008, + "step": 2320 + }, + { + "epoch": 1.799844840961986, + "eval_nq_pairs_loss": 0.5333605408668518, + "eval_nq_pairs_runtime": 3.3827, + "eval_nq_pairs_samples_per_second": 37.84, + "eval_nq_pairs_steps_per_second": 0.591, + "step": 2320 + }, + { + "epoch": 1.799844840961986, + "eval_trivia_pairs_loss": 0.6911617517471313, + "eval_trivia_pairs_runtime": 4.5192, + "eval_trivia_pairs_samples_per_second": 28.324, + "eval_trivia_pairs_steps_per_second": 0.443, + "step": 2320 + }, + { + "epoch": 1.799844840961986, + "eval_gooaq_pairs_loss": 0.5544390678405762, + "eval_gooaq_pairs_runtime": 1.4824, + "eval_gooaq_pairs_samples_per_second": 86.345, + "eval_gooaq_pairs_steps_per_second": 1.349, + "step": 2320 + }, + { + "epoch": 1.799844840961986, + "eval_paws-pos_loss": 0.030726628378033638, + "eval_paws-pos_runtime": 1.0464, + "eval_paws-pos_samples_per_second": 122.326, + "eval_paws-pos_steps_per_second": 1.911, + "step": 2320 + }, + { + "epoch": 1.799844840961986, + "eval_global_dataset_loss": 0.3957351744174957, + "eval_global_dataset_runtime": 19.4724, + "eval_global_dataset_samples_per_second": 21.364, + "eval_global_dataset_steps_per_second": 0.359, + "step": 2320 + }, + { + "epoch": 1.8006206361520558, + "grad_norm": 5.154975891113281, + "learning_rate": 5.2290250627622355e-06, + "loss": 0.2043, + "step": 2321 + }, + { + "epoch": 1.8013964313421256, + "grad_norm": 9.584165573120117, + "learning_rate": 5.215996940524511e-06, + "loss": 0.6568, + "step": 2322 + } + ], + "logging_steps": 1, + "max_steps": 2578, + "num_input_tokens_seen": 0, + "num_train_epochs": 2, + "save_steps": 258, + "stateful_callbacks": { + "TrainerControl": { + "args": { + "should_epoch_stop": false, + "should_evaluate": false, + "should_log": false, + "should_save": true, + "should_training_stop": false + }, + "attributes": {} + } + }, + "total_flos": 0.0, + "train_batch_size": 96, + "trial_name": null, + "trial_params": null +}