diff --git "a/checkpoint-774/trainer_state.json" "b/checkpoint-774/trainer_state.json" new file mode 100644--- /dev/null +++ "b/checkpoint-774/trainer_state.json" @@ -0,0 +1,12785 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 0.6004654771140419, + "eval_steps": 20, + "global_step": 774, + "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 + } + ], + "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 +}