diff --git "a/checkpoint-2354/trainer_state.json" "b/checkpoint-2354/trainer_state.json" new file mode 100644--- /dev/null +++ "b/checkpoint-2354/trainer_state.json" @@ -0,0 +1,3654 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 2.753216374269006, + "eval_steps": 214, + "global_step": 2354, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.012865497076023392, + "grad_norm": 0.07311347126960754, + "learning_rate": 3.5081502480510277e-07, + "loss": 0.1802, + "step": 11 + }, + { + "epoch": 0.025730994152046785, + "grad_norm": 0.3663940727710724, + "learning_rate": 7.016300496102055e-07, + "loss": 0.1573, + "step": 22 + }, + { + "epoch": 0.03859649122807018, + "grad_norm": 0.22339867055416107, + "learning_rate": 1.0524450744153084e-06, + "loss": 0.1184, + "step": 33 + }, + { + "epoch": 0.05146198830409357, + "grad_norm": 0.06488268822431564, + "learning_rate": 1.403260099220411e-06, + "loss": 0.1456, + "step": 44 + }, + { + "epoch": 0.06432748538011696, + "grad_norm": 0.7134692072868347, + "learning_rate": 1.754075124025514e-06, + "loss": 0.2102, + "step": 55 + }, + { + "epoch": 0.07719298245614035, + "grad_norm": 0.048255760222673416, + "learning_rate": 2.104890148830617e-06, + "loss": 0.151, + "step": 66 + }, + { + "epoch": 0.09005847953216374, + "grad_norm": 0.43056783080101013, + "learning_rate": 2.4557051736357195e-06, + "loss": 0.1356, + "step": 77 + }, + { + "epoch": 0.10292397660818714, + "grad_norm": 0.7663404941558838, + "learning_rate": 2.806520198440822e-06, + "loss": 0.2292, + "step": 88 + }, + { + "epoch": 0.11578947368421053, + "grad_norm": 0.48333388566970825, + "learning_rate": 3.1573352232459252e-06, + "loss": 0.126, + "step": 99 + }, + { + "epoch": 0.1286549707602339, + "grad_norm": 0.5180703997612, + "learning_rate": 3.508150248051028e-06, + "loss": 0.1942, + "step": 110 + }, + { + "epoch": 0.1415204678362573, + "grad_norm": 0.6447436809539795, + "learning_rate": 3.85896527285613e-06, + "loss": 0.2089, + "step": 121 + }, + { + "epoch": 0.1543859649122807, + "grad_norm": 0.5512965321540833, + "learning_rate": 4.209780297661234e-06, + "loss": 0.1225, + "step": 132 + }, + { + "epoch": 0.1672514619883041, + "grad_norm": 0.5181564688682556, + "learning_rate": 4.560595322466336e-06, + "loss": 0.2504, + "step": 143 + }, + { + "epoch": 0.18011695906432748, + "grad_norm": 1.549516201019287, + "learning_rate": 4.911410347271439e-06, + "loss": 0.1454, + "step": 154 + }, + { + "epoch": 0.19298245614035087, + "grad_norm": 0.06744886934757233, + "learning_rate": 5.262225372076542e-06, + "loss": 0.2052, + "step": 165 + }, + { + "epoch": 0.20584795321637428, + "grad_norm": 0.3412599265575409, + "learning_rate": 5.613040396881644e-06, + "loss": 0.1321, + "step": 176 + }, + { + "epoch": 0.21871345029239767, + "grad_norm": 0.4878319203853607, + "learning_rate": 5.963855421686748e-06, + "loss": 0.1975, + "step": 187 + }, + { + "epoch": 0.23157894736842105, + "grad_norm": 0.6517124176025391, + "learning_rate": 6.3146704464918505e-06, + "loss": 0.1615, + "step": 198 + }, + { + "epoch": 0.24444444444444444, + "grad_norm": 0.1128321960568428, + "learning_rate": 6.665485471296953e-06, + "loss": 0.2174, + "step": 209 + }, + { + "epoch": 0.25029239766081873, + "eval_Qnli-dev_cosine_accuracy": 0.697265625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6495540738105774, + "eval_Qnli-dev_cosine_ap": 0.7205205074888317, + "eval_Qnli-dev_cosine_f1": 0.6815286624203821, + "eval_Qnli-dev_cosine_f1_threshold": 0.4835514724254608, + "eval_Qnli-dev_cosine_precision": 0.5459183673469388, + "eval_Qnli-dev_cosine_recall": 0.9067796610169492, + "eval_Qnli-dev_dot_accuracy": 0.666015625, + "eval_Qnli-dev_dot_accuracy_threshold": 307.405517578125, + "eval_Qnli-dev_dot_ap": 0.676904834512821, + "eval_Qnli-dev_dot_f1": 0.6718027734976887, + "eval_Qnli-dev_dot_f1_threshold": 182.19287109375, + "eval_Qnli-dev_dot_precision": 0.5278450363196125, + "eval_Qnli-dev_dot_recall": 0.923728813559322, + "eval_Qnli-dev_euclidean_accuracy": 0.70703125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 17.179746627807617, + "eval_Qnli-dev_euclidean_ap": 0.7311696464051982, + "eval_Qnli-dev_euclidean_f1": 0.6860706860706861, + "eval_Qnli-dev_euclidean_f1_threshold": 17.50692367553711, + "eval_Qnli-dev_euclidean_precision": 0.673469387755102, + "eval_Qnli-dev_euclidean_recall": 0.6991525423728814, + "eval_Qnli-dev_manhattan_accuracy": 0.703125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 357.37890625, + "eval_Qnli-dev_manhattan_ap": 0.7327744127614749, + "eval_Qnli-dev_manhattan_f1": 0.680161943319838, + "eval_Qnli-dev_manhattan_f1_threshold": 373.75103759765625, + "eval_Qnli-dev_manhattan_precision": 0.6511627906976745, + "eval_Qnli-dev_manhattan_recall": 0.711864406779661, + "eval_Qnli-dev_max_accuracy": 0.70703125, + "eval_Qnli-dev_max_accuracy_threshold": 357.37890625, + "eval_Qnli-dev_max_ap": 0.7327744127614749, + "eval_Qnli-dev_max_f1": 0.6860706860706861, + "eval_Qnli-dev_max_f1_threshold": 373.75103759765625, + "eval_Qnli-dev_max_precision": 0.673469387755102, + "eval_Qnli-dev_max_recall": 0.923728813559322, + "eval_allNLI-dev_cosine_accuracy": 0.740234375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7851241827011108, + "eval_allNLI-dev_cosine_ap": 0.6340322426598262, + "eval_allNLI-dev_cosine_f1": 0.6371681415929203, + "eval_allNLI-dev_cosine_f1_threshold": 0.5923386812210083, + "eval_allNLI-dev_cosine_precision": 0.5161290322580645, + "eval_allNLI-dev_cosine_recall": 0.8323699421965318, + "eval_allNLI-dev_dot_accuracy": 0.724609375, + "eval_allNLI-dev_dot_accuracy_threshold": 294.0272521972656, + "eval_allNLI-dev_dot_ap": 0.5910094855503549, + "eval_allNLI-dev_dot_f1": 0.5960264900662251, + "eval_allNLI-dev_dot_f1_threshold": 227.6922149658203, + "eval_allNLI-dev_dot_precision": 0.48214285714285715, + "eval_allNLI-dev_dot_recall": 0.7803468208092486, + "eval_allNLI-dev_euclidean_accuracy": 0.748046875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 13.517601013183594, + "eval_allNLI-dev_euclidean_ap": 0.6422252682731053, + "eval_allNLI-dev_euclidean_f1": 0.65, + "eval_allNLI-dev_euclidean_f1_threshold": 17.501617431640625, + "eval_allNLI-dev_euclidean_precision": 0.5355805243445693, + "eval_allNLI-dev_euclidean_recall": 0.8265895953757225, + "eval_allNLI-dev_manhattan_accuracy": 0.74609375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 298.3919677734375, + "eval_allNLI-dev_manhattan_ap": 0.6432220315530937, + "eval_allNLI-dev_manhattan_f1": 0.6618357487922705, + "eval_allNLI-dev_manhattan_f1_threshold": 358.1668701171875, + "eval_allNLI-dev_manhattan_precision": 0.5684647302904564, + "eval_allNLI-dev_manhattan_recall": 0.791907514450867, + "eval_allNLI-dev_max_accuracy": 0.748046875, + "eval_allNLI-dev_max_accuracy_threshold": 298.3919677734375, + "eval_allNLI-dev_max_ap": 0.6432220315530937, + "eval_allNLI-dev_max_f1": 0.6618357487922705, + "eval_allNLI-dev_max_f1_threshold": 358.1668701171875, + "eval_allNLI-dev_max_precision": 0.5684647302904564, + "eval_allNLI-dev_max_recall": 0.8323699421965318, + "eval_sequential_score": 0.7327744127614749, + "eval_sts-test_pearson_cosine": 0.8852985023833193, + "eval_sts-test_pearson_dot": 0.8698187375798812, + "eval_sts-test_pearson_euclidean": 0.9048697523934177, + "eval_sts-test_pearson_manhattan": 0.9054784772409876, + "eval_sts-test_pearson_max": 0.9054784772409876, + "eval_sts-test_spearman_cosine": 0.9079225861122531, + "eval_sts-test_spearman_dot": 0.8661416896957076, + "eval_sts-test_spearman_euclidean": 0.9007419777144345, + "eval_sts-test_spearman_manhattan": 0.9013345291665179, + "eval_sts-test_spearman_max": 0.9079225861122531, + "eval_vitaminc-pairs_loss": 2.7610974311828613, + "eval_vitaminc-pairs_runtime": 3.8706, + "eval_vitaminc-pairs_samples_per_second": 33.07, + "eval_vitaminc-pairs_steps_per_second": 0.258, + "step": 214 + }, + { + "epoch": 0.25029239766081873, + "eval_negation-triplets_loss": 1.0606263875961304, + "eval_negation-triplets_runtime": 0.6073, + "eval_negation-triplets_samples_per_second": 210.757, + "eval_negation-triplets_steps_per_second": 1.647, + "step": 214 + }, + { + "epoch": 0.25029239766081873, + "eval_scitail-pairs-pos_loss": 0.059617433696985245, + "eval_scitail-pairs-pos_runtime": 0.9151, + "eval_scitail-pairs-pos_samples_per_second": 139.868, + "eval_scitail-pairs-pos_steps_per_second": 1.093, + "step": 214 + }, + { + "epoch": 0.25029239766081873, + "eval_scitail-pairs-qa_loss": 2.513320396246854e-07, + "eval_scitail-pairs-qa_runtime": 0.5906, + "eval_scitail-pairs-qa_samples_per_second": 216.739, + "eval_scitail-pairs-qa_steps_per_second": 1.693, + "step": 214 + }, + { + "epoch": 0.25029239766081873, + "eval_xsum-pairs_loss": 0.004431054927408695, + "eval_xsum-pairs_runtime": 2.7873, + "eval_xsum-pairs_samples_per_second": 45.922, + "eval_xsum-pairs_steps_per_second": 0.359, + "step": 214 + }, + { + "epoch": 0.25029239766081873, + "eval_sciq_pairs_loss": 0.013232991099357605, + "eval_sciq_pairs_runtime": 4.1862, + "eval_sciq_pairs_samples_per_second": 30.576, + "eval_sciq_pairs_steps_per_second": 0.239, + "step": 214 + }, + { + "epoch": 0.25029239766081873, + "eval_qasc_pairs_loss": 0.049007147550582886, + "eval_qasc_pairs_runtime": 0.6288, + "eval_qasc_pairs_samples_per_second": 203.554, + "eval_qasc_pairs_steps_per_second": 1.59, + "step": 214 + }, + { + "epoch": 0.25029239766081873, + "eval_openbookqa_pairs_loss": 1.3196656703948975, + "eval_openbookqa_pairs_runtime": 0.5848, + "eval_openbookqa_pairs_samples_per_second": 218.897, + "eval_openbookqa_pairs_steps_per_second": 1.71, + "step": 214 + }, + { + "epoch": 0.25029239766081873, + "eval_msmarco_pairs_loss": 0.13587385416030884, + "eval_msmarco_pairs_runtime": 1.341, + "eval_msmarco_pairs_samples_per_second": 95.453, + "eval_msmarco_pairs_steps_per_second": 0.746, + "step": 214 + }, + { + "epoch": 0.25029239766081873, + "eval_nq_pairs_loss": 0.04134061187505722, + "eval_nq_pairs_runtime": 2.9296, + "eval_nq_pairs_samples_per_second": 43.693, + "eval_nq_pairs_steps_per_second": 0.341, + "step": 214 + }, + { + "epoch": 0.25029239766081873, + "eval_trivia_pairs_loss": 0.211629718542099, + "eval_trivia_pairs_runtime": 3.2033, + "eval_trivia_pairs_samples_per_second": 39.958, + "eval_trivia_pairs_steps_per_second": 0.312, + "step": 214 + }, + { + "epoch": 0.25029239766081873, + "eval_gooaq_pairs_loss": 0.11869887262582779, + "eval_gooaq_pairs_runtime": 0.869, + "eval_gooaq_pairs_samples_per_second": 147.291, + "eval_gooaq_pairs_steps_per_second": 1.151, + "step": 214 + }, + { + "epoch": 0.25029239766081873, + "eval_paws-pos_loss": 0.04590649530291557, + "eval_paws-pos_runtime": 0.6652, + "eval_paws-pos_samples_per_second": 192.433, + "eval_paws-pos_steps_per_second": 1.503, + "step": 214 + }, + { + "epoch": 0.25029239766081873, + "eval_global_dataset_loss": 0.21873071789741516, + "eval_global_dataset_runtime": 24.2472, + "eval_global_dataset_samples_per_second": 27.343, + "eval_global_dataset_steps_per_second": 0.247, + "step": 214 + }, + { + "epoch": 0.2573099415204678, + "grad_norm": 0.2186487466096878, + "learning_rate": 7.016300496102056e-06, + "loss": 0.1545, + "step": 220 + }, + { + "epoch": 0.27017543859649124, + "grad_norm": 0.16798946261405945, + "learning_rate": 7.3671155209071584e-06, + "loss": 0.1201, + "step": 231 + }, + { + "epoch": 0.2830409356725146, + "grad_norm": 0.5161852836608887, + "learning_rate": 7.71793054571226e-06, + "loss": 0.1389, + "step": 242 + }, + { + "epoch": 0.295906432748538, + "grad_norm": 0.6402292251586914, + "learning_rate": 8.068745570517365e-06, + "loss": 0.1657, + "step": 253 + }, + { + "epoch": 0.3087719298245614, + "grad_norm": 0.8064345717430115, + "learning_rate": 8.419560595322467e-06, + "loss": 0.237, + "step": 264 + }, + { + "epoch": 0.3216374269005848, + "grad_norm": 0.8077338337898254, + "learning_rate": 8.770375620127569e-06, + "loss": 0.1094, + "step": 275 + }, + { + "epoch": 0.3345029239766082, + "grad_norm": 0.23748837411403656, + "learning_rate": 9.121190644932673e-06, + "loss": 0.196, + "step": 286 + }, + { + "epoch": 0.3473684210526316, + "grad_norm": 0.26720044016838074, + "learning_rate": 9.472005669737774e-06, + "loss": 0.2164, + "step": 297 + }, + { + "epoch": 0.36023391812865496, + "grad_norm": 0.8082326650619507, + "learning_rate": 9.822820694542878e-06, + "loss": 0.1793, + "step": 308 + }, + { + "epoch": 0.3730994152046784, + "grad_norm": 0.8150445818901062, + "learning_rate": 1.017363571934798e-05, + "loss": 0.2878, + "step": 319 + }, + { + "epoch": 0.38596491228070173, + "grad_norm": 0.2097979038953781, + "learning_rate": 1.0524450744153083e-05, + "loss": 0.1189, + "step": 330 + }, + { + "epoch": 0.39883040935672515, + "grad_norm": 1.0323094129562378, + "learning_rate": 1.0875265768958187e-05, + "loss": 0.1475, + "step": 341 + }, + { + "epoch": 0.41169590643274856, + "grad_norm": 0.41917282342910767, + "learning_rate": 1.1226080793763289e-05, + "loss": 0.1019, + "step": 352 + }, + { + "epoch": 0.4245614035087719, + "grad_norm": 0.17783042788505554, + "learning_rate": 1.1576895818568394e-05, + "loss": 0.1587, + "step": 363 + }, + { + "epoch": 0.43742690058479533, + "grad_norm": 0.8943735957145691, + "learning_rate": 1.1927710843373496e-05, + "loss": 0.2483, + "step": 374 + }, + { + "epoch": 0.4502923976608187, + "grad_norm": 0.7853964567184448, + "learning_rate": 1.2278525868178599e-05, + "loss": 0.1427, + "step": 385 + }, + { + "epoch": 0.4631578947368421, + "grad_norm": 0.4753592908382416, + "learning_rate": 1.2629340892983701e-05, + "loss": 0.1199, + "step": 396 + }, + { + "epoch": 0.4760233918128655, + "grad_norm": 0.47377559542655945, + "learning_rate": 1.2980155917788803e-05, + "loss": 0.2037, + "step": 407 + }, + { + "epoch": 0.4888888888888889, + "grad_norm": 0.6249018311500549, + "learning_rate": 1.3330970942593906e-05, + "loss": 0.1317, + "step": 418 + }, + { + "epoch": 0.5005847953216375, + "eval_Qnli-dev_cosine_accuracy": 0.6953125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6539117097854614, + "eval_Qnli-dev_cosine_ap": 0.7174538904795291, + "eval_Qnli-dev_cosine_f1": 0.6773675762439807, + "eval_Qnli-dev_cosine_f1_threshold": 0.4876832365989685, + "eval_Qnli-dev_cosine_precision": 0.5452196382428941, + "eval_Qnli-dev_cosine_recall": 0.8940677966101694, + "eval_Qnli-dev_dot_accuracy": 0.673828125, + "eval_Qnli-dev_dot_accuracy_threshold": 248.8419647216797, + "eval_Qnli-dev_dot_ap": 0.6692436401997881, + "eval_Qnli-dev_dot_f1": 0.6758832565284177, + "eval_Qnli-dev_dot_f1_threshold": 174.55992126464844, + "eval_Qnli-dev_dot_precision": 0.5301204819277109, + "eval_Qnli-dev_dot_recall": 0.9322033898305084, + "eval_Qnli-dev_euclidean_accuracy": 0.70703125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 16.61339569091797, + "eval_Qnli-dev_euclidean_ap": 0.7278403442362325, + "eval_Qnli-dev_euclidean_f1": 0.6821052631578948, + "eval_Qnli-dev_euclidean_f1_threshold": 17.059036254882812, + "eval_Qnli-dev_euclidean_precision": 0.6778242677824268, + "eval_Qnli-dev_euclidean_recall": 0.6864406779661016, + "eval_Qnli-dev_manhattan_accuracy": 0.697265625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 348.561279296875, + "eval_Qnli-dev_manhattan_ap": 0.7293997108333973, + "eval_Qnli-dev_manhattan_f1": 0.6797642436149313, + "eval_Qnli-dev_manhattan_f1_threshold": 371.4664306640625, + "eval_Qnli-dev_manhattan_precision": 0.6336996336996337, + "eval_Qnli-dev_manhattan_recall": 0.7330508474576272, + "eval_Qnli-dev_max_accuracy": 0.70703125, + "eval_Qnli-dev_max_accuracy_threshold": 348.561279296875, + "eval_Qnli-dev_max_ap": 0.7293997108333973, + "eval_Qnli-dev_max_f1": 0.6821052631578948, + "eval_Qnli-dev_max_f1_threshold": 371.4664306640625, + "eval_Qnli-dev_max_precision": 0.6778242677824268, + "eval_Qnli-dev_max_recall": 0.9322033898305084, + "eval_allNLI-dev_cosine_accuracy": 0.740234375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.789513111114502, + "eval_allNLI-dev_cosine_ap": 0.6326730829325957, + "eval_allNLI-dev_cosine_f1": 0.6369710467706012, + "eval_allNLI-dev_cosine_f1_threshold": 0.6001436710357666, + "eval_allNLI-dev_cosine_precision": 0.5181159420289855, + "eval_allNLI-dev_cosine_recall": 0.8265895953757225, + "eval_allNLI-dev_dot_accuracy": 0.728515625, + "eval_allNLI-dev_dot_accuracy_threshold": 284.5158996582031, + "eval_allNLI-dev_dot_ap": 0.587991265887336, + "eval_allNLI-dev_dot_f1": 0.5996204933586338, + "eval_allNLI-dev_dot_f1_threshold": 187.17926025390625, + "eval_allNLI-dev_dot_precision": 0.4463276836158192, + "eval_allNLI-dev_dot_recall": 0.9132947976878613, + "eval_allNLI-dev_euclidean_accuracy": 0.740234375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 13.375718116760254, + "eval_allNLI-dev_euclidean_ap": 0.6391067622382595, + "eval_allNLI-dev_euclidean_f1": 0.6478260869565217, + "eval_allNLI-dev_euclidean_f1_threshold": 17.581111907958984, + "eval_allNLI-dev_euclidean_precision": 0.519163763066202, + "eval_allNLI-dev_euclidean_recall": 0.861271676300578, + "eval_allNLI-dev_manhattan_accuracy": 0.740234375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 271.5396728515625, + "eval_allNLI-dev_manhattan_ap": 0.6397141445378819, + "eval_allNLI-dev_manhattan_f1": 0.6558139534883721, + "eval_allNLI-dev_manhattan_f1_threshold": 356.73675537109375, + "eval_allNLI-dev_manhattan_precision": 0.5486381322957199, + "eval_allNLI-dev_manhattan_recall": 0.815028901734104, + "eval_allNLI-dev_max_accuracy": 0.740234375, + "eval_allNLI-dev_max_accuracy_threshold": 284.5158996582031, + "eval_allNLI-dev_max_ap": 0.6397141445378819, + "eval_allNLI-dev_max_f1": 0.6558139534883721, + "eval_allNLI-dev_max_f1_threshold": 356.73675537109375, + "eval_allNLI-dev_max_precision": 0.5486381322957199, + "eval_allNLI-dev_max_recall": 0.9132947976878613, + "eval_sequential_score": 0.7293997108333973, + "eval_sts-test_pearson_cosine": 0.8826342923852873, + "eval_sts-test_pearson_dot": 0.8663977738237862, + "eval_sts-test_pearson_euclidean": 0.9007138184358896, + "eval_sts-test_pearson_manhattan": 0.9014088314200603, + "eval_sts-test_pearson_max": 0.9014088314200603, + "eval_sts-test_spearman_cosine": 0.9062742181501622, + "eval_sts-test_spearman_dot": 0.8628408814067393, + "eval_sts-test_spearman_euclidean": 0.897460680975433, + "eval_sts-test_spearman_manhattan": 0.8979340598184311, + "eval_sts-test_spearman_max": 0.9062742181501622, + "eval_vitaminc-pairs_loss": 2.7317824363708496, + "eval_vitaminc-pairs_runtime": 3.8325, + "eval_vitaminc-pairs_samples_per_second": 33.398, + "eval_vitaminc-pairs_steps_per_second": 0.261, + "step": 428 + }, + { + "epoch": 0.5005847953216375, + "eval_negation-triplets_loss": 1.0631705522537231, + "eval_negation-triplets_runtime": 0.6008, + "eval_negation-triplets_samples_per_second": 213.032, + "eval_negation-triplets_steps_per_second": 1.664, + "step": 428 + }, + { + "epoch": 0.5005847953216375, + "eval_scitail-pairs-pos_loss": 0.06051468849182129, + "eval_scitail-pairs-pos_runtime": 0.8927, + "eval_scitail-pairs-pos_samples_per_second": 143.381, + "eval_scitail-pairs-pos_steps_per_second": 1.12, + "step": 428 + }, + { + "epoch": 0.5005847953216375, + "eval_scitail-pairs-qa_loss": 3.731894366865163e-07, + "eval_scitail-pairs-qa_runtime": 0.5829, + "eval_scitail-pairs-qa_samples_per_second": 219.584, + "eval_scitail-pairs-qa_steps_per_second": 1.716, + "step": 428 + }, + { + "epoch": 0.5005847953216375, + "eval_xsum-pairs_loss": 0.004530374892055988, + "eval_xsum-pairs_runtime": 2.7882, + "eval_xsum-pairs_samples_per_second": 45.908, + "eval_xsum-pairs_steps_per_second": 0.359, + "step": 428 + }, + { + "epoch": 0.5005847953216375, + "eval_sciq_pairs_loss": 0.0028694826178252697, + "eval_sciq_pairs_runtime": 4.1341, + "eval_sciq_pairs_samples_per_second": 30.962, + "eval_sciq_pairs_steps_per_second": 0.242, + "step": 428 + }, + { + "epoch": 0.5005847953216375, + "eval_qasc_pairs_loss": 0.0601225420832634, + "eval_qasc_pairs_runtime": 0.6017, + "eval_qasc_pairs_samples_per_second": 212.728, + "eval_qasc_pairs_steps_per_second": 1.662, + "step": 428 + }, + { + "epoch": 0.5005847953216375, + "eval_openbookqa_pairs_loss": 1.2716314792633057, + "eval_openbookqa_pairs_runtime": 0.5746, + "eval_openbookqa_pairs_samples_per_second": 222.757, + "eval_openbookqa_pairs_steps_per_second": 1.74, + "step": 428 + }, + { + "epoch": 0.5005847953216375, + "eval_msmarco_pairs_loss": 0.16218337416648865, + "eval_msmarco_pairs_runtime": 1.3355, + "eval_msmarco_pairs_samples_per_second": 95.842, + "eval_msmarco_pairs_steps_per_second": 0.749, + "step": 428 + }, + { + "epoch": 0.5005847953216375, + "eval_nq_pairs_loss": 0.039985254406929016, + "eval_nq_pairs_runtime": 2.9238, + "eval_nq_pairs_samples_per_second": 43.779, + "eval_nq_pairs_steps_per_second": 0.342, + "step": 428 + }, + { + "epoch": 0.5005847953216375, + "eval_trivia_pairs_loss": 0.20936143398284912, + "eval_trivia_pairs_runtime": 3.1822, + "eval_trivia_pairs_samples_per_second": 40.223, + "eval_trivia_pairs_steps_per_second": 0.314, + "step": 428 + }, + { + "epoch": 0.5005847953216375, + "eval_gooaq_pairs_loss": 0.12705038487911224, + "eval_gooaq_pairs_runtime": 0.8602, + "eval_gooaq_pairs_samples_per_second": 148.795, + "eval_gooaq_pairs_steps_per_second": 1.162, + "step": 428 + }, + { + "epoch": 0.5005847953216375, + "eval_paws-pos_loss": 0.046078916639089584, + "eval_paws-pos_runtime": 0.6578, + "eval_paws-pos_samples_per_second": 194.575, + "eval_paws-pos_steps_per_second": 1.52, + "step": 428 + }, + { + "epoch": 0.5005847953216375, + "eval_global_dataset_loss": 0.21221815049648285, + "eval_global_dataset_runtime": 24.2782, + "eval_global_dataset_samples_per_second": 27.308, + "eval_global_dataset_steps_per_second": 0.247, + "step": 428 + }, + { + "epoch": 0.5017543859649123, + "grad_norm": 0.6901716589927673, + "learning_rate": 1.3681785967399008e-05, + "loss": 0.1293, + "step": 429 + }, + { + "epoch": 0.5146198830409356, + "grad_norm": 1.1465657949447632, + "learning_rate": 1.4032600992204112e-05, + "loss": 0.1902, + "step": 440 + }, + { + "epoch": 0.5274853801169591, + "grad_norm": 1.0795351266860962, + "learning_rate": 1.4383416017009213e-05, + "loss": 0.1429, + "step": 451 + }, + { + "epoch": 0.5403508771929825, + "grad_norm": 0.9797111749649048, + "learning_rate": 1.4734231041814317e-05, + "loss": 0.2446, + "step": 462 + }, + { + "epoch": 0.5532163742690058, + "grad_norm": 0.21088336408138275, + "learning_rate": 1.5085046066619419e-05, + "loss": 0.1623, + "step": 473 + }, + { + "epoch": 0.5660818713450292, + "grad_norm": 0.4258130192756653, + "learning_rate": 1.543586109142452e-05, + "loss": 0.0707, + "step": 484 + }, + { + "epoch": 0.5789473684210527, + "grad_norm": 0.8029544949531555, + "learning_rate": 1.5786676116229624e-05, + "loss": 0.1557, + "step": 495 + }, + { + "epoch": 0.591812865497076, + "grad_norm": 2.734902858734131, + "learning_rate": 1.613749114103473e-05, + "loss": 0.2016, + "step": 506 + }, + { + "epoch": 0.6046783625730994, + "grad_norm": 0.4968857169151306, + "learning_rate": 1.648830616583983e-05, + "loss": 0.1018, + "step": 517 + }, + { + "epoch": 0.6175438596491228, + "grad_norm": 1.5250153541564941, + "learning_rate": 1.6839121190644935e-05, + "loss": 0.1821, + "step": 528 + }, + { + "epoch": 0.6304093567251462, + "grad_norm": 0.7500655055046082, + "learning_rate": 1.7189936215450038e-05, + "loss": 0.1437, + "step": 539 + }, + { + "epoch": 0.6432748538011696, + "grad_norm": 0.19580113887786865, + "learning_rate": 1.7540751240255138e-05, + "loss": 0.1112, + "step": 550 + }, + { + "epoch": 0.656140350877193, + "grad_norm": 0.1376626342535019, + "learning_rate": 1.789156626506024e-05, + "loss": 0.12, + "step": 561 + }, + { + "epoch": 0.6690058479532164, + "grad_norm": 0.49516016244888306, + "learning_rate": 1.8242381289865345e-05, + "loss": 0.0933, + "step": 572 + }, + { + "epoch": 0.6818713450292397, + "grad_norm": 0.20120573043823242, + "learning_rate": 1.859319631467045e-05, + "loss": 0.0939, + "step": 583 + }, + { + "epoch": 0.6947368421052632, + "grad_norm": 0.3648889362812042, + "learning_rate": 1.894401133947555e-05, + "loss": 0.2064, + "step": 594 + }, + { + "epoch": 0.7076023391812866, + "grad_norm": 0.07958167791366577, + "learning_rate": 1.9294826364280652e-05, + "loss": 0.131, + "step": 605 + }, + { + "epoch": 0.7204678362573099, + "grad_norm": 0.7020072340965271, + "learning_rate": 1.9645641389085756e-05, + "loss": 0.161, + "step": 616 + }, + { + "epoch": 0.7333333333333333, + "grad_norm": 1.3057806491851807, + "learning_rate": 1.999645641389086e-05, + "loss": 0.213, + "step": 627 + }, + { + "epoch": 0.7461988304093568, + "grad_norm": 0.7668664455413818, + "learning_rate": 2.034727143869596e-05, + "loss": 0.1853, + "step": 638 + }, + { + "epoch": 0.7508771929824561, + "eval_Qnli-dev_cosine_accuracy": 0.693359375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6550551652908325, + "eval_Qnli-dev_cosine_ap": 0.7155098805507838, + "eval_Qnli-dev_cosine_f1": 0.6794871794871794, + "eval_Qnli-dev_cosine_f1_threshold": 0.4866919219493866, + "eval_Qnli-dev_cosine_precision": 0.5463917525773195, + "eval_Qnli-dev_cosine_recall": 0.8983050847457628, + "eval_Qnli-dev_dot_accuracy": 0.67578125, + "eval_Qnli-dev_dot_accuracy_threshold": 258.037841796875, + "eval_Qnli-dev_dot_ap": 0.672701841309456, + "eval_Qnli-dev_dot_f1": 0.6731078904991948, + "eval_Qnli-dev_dot_f1_threshold": 197.6945343017578, + "eval_Qnli-dev_dot_precision": 0.5428571428571428, + "eval_Qnli-dev_dot_recall": 0.885593220338983, + "eval_Qnli-dev_euclidean_accuracy": 0.697265625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 16.745851516723633, + "eval_Qnli-dev_euclidean_ap": 0.7236455475550527, + "eval_Qnli-dev_euclidean_f1": 0.6750524109014674, + "eval_Qnli-dev_euclidean_f1_threshold": 17.19139862060547, + "eval_Qnli-dev_euclidean_precision": 0.6680497925311203, + "eval_Qnli-dev_euclidean_recall": 0.6822033898305084, + "eval_Qnli-dev_manhattan_accuracy": 0.697265625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 362.6970520019531, + "eval_Qnli-dev_manhattan_ap": 0.7257321921740054, + "eval_Qnli-dev_manhattan_f1": 0.6774847870182557, + "eval_Qnli-dev_manhattan_f1_threshold": 369.7593994140625, + "eval_Qnli-dev_manhattan_precision": 0.6498054474708171, + "eval_Qnli-dev_manhattan_recall": 0.7076271186440678, + "eval_Qnli-dev_max_accuracy": 0.697265625, + "eval_Qnli-dev_max_accuracy_threshold": 362.6970520019531, + "eval_Qnli-dev_max_ap": 0.7257321921740054, + "eval_Qnli-dev_max_f1": 0.6794871794871794, + "eval_Qnli-dev_max_f1_threshold": 369.7593994140625, + "eval_Qnli-dev_max_precision": 0.6680497925311203, + "eval_Qnli-dev_max_recall": 0.8983050847457628, + "eval_allNLI-dev_cosine_accuracy": 0.73828125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7900433540344238, + "eval_allNLI-dev_cosine_ap": 0.6359348988299178, + "eval_allNLI-dev_cosine_f1": 0.639269406392694, + "eval_allNLI-dev_cosine_f1_threshold": 0.6079586744308472, + "eval_allNLI-dev_cosine_precision": 0.5283018867924528, + "eval_allNLI-dev_cosine_recall": 0.8092485549132948, + "eval_allNLI-dev_dot_accuracy": 0.7265625, + "eval_allNLI-dev_dot_accuracy_threshold": 282.57489013671875, + "eval_allNLI-dev_dot_ap": 0.589207838274076, + "eval_allNLI-dev_dot_f1": 0.6059405940594059, + "eval_allNLI-dev_dot_f1_threshold": 203.0603485107422, + "eval_allNLI-dev_dot_precision": 0.4608433734939759, + "eval_allNLI-dev_dot_recall": 0.884393063583815, + "eval_allNLI-dev_euclidean_accuracy": 0.744140625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 13.129840850830078, + "eval_allNLI-dev_euclidean_ap": 0.6431587171633473, + "eval_allNLI-dev_euclidean_f1": 0.6506550218340611, + "eval_allNLI-dev_euclidean_f1_threshold": 17.757301330566406, + "eval_allNLI-dev_euclidean_precision": 0.5228070175438596, + "eval_allNLI-dev_euclidean_recall": 0.861271676300578, + "eval_allNLI-dev_manhattan_accuracy": 0.74609375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 276.65826416015625, + "eval_allNLI-dev_manhattan_ap": 0.6424686977323816, + "eval_allNLI-dev_manhattan_f1": 0.6589861751152074, + "eval_allNLI-dev_manhattan_f1_threshold": 362.68414306640625, + "eval_allNLI-dev_manhattan_precision": 0.5478927203065134, + "eval_allNLI-dev_manhattan_recall": 0.8265895953757225, + "eval_allNLI-dev_max_accuracy": 0.74609375, + "eval_allNLI-dev_max_accuracy_threshold": 282.57489013671875, + "eval_allNLI-dev_max_ap": 0.6431587171633473, + "eval_allNLI-dev_max_f1": 0.6589861751152074, + "eval_allNLI-dev_max_f1_threshold": 362.68414306640625, + "eval_allNLI-dev_max_precision": 0.5478927203065134, + "eval_allNLI-dev_max_recall": 0.884393063583815, + "eval_sequential_score": 0.7257321921740054, + "eval_sts-test_pearson_cosine": 0.8848372605057793, + "eval_sts-test_pearson_dot": 0.8709019996065221, + "eval_sts-test_pearson_euclidean": 0.9038234321498232, + "eval_sts-test_pearson_manhattan": 0.9044737426180963, + "eval_sts-test_pearson_max": 0.9044737426180963, + "eval_sts-test_spearman_cosine": 0.9073185783372533, + "eval_sts-test_spearman_dot": 0.8668432552205406, + "eval_sts-test_spearman_euclidean": 0.899716547310912, + "eval_sts-test_spearman_manhattan": 0.9006361409812439, + "eval_sts-test_spearman_max": 0.9073185783372533, + "eval_vitaminc-pairs_loss": 2.7074272632598877, + "eval_vitaminc-pairs_runtime": 3.7948, + "eval_vitaminc-pairs_samples_per_second": 33.73, + "eval_vitaminc-pairs_steps_per_second": 0.264, + "step": 642 + }, + { + "epoch": 0.7508771929824561, + "eval_negation-triplets_loss": 1.0683540105819702, + "eval_negation-triplets_runtime": 0.5897, + "eval_negation-triplets_samples_per_second": 217.072, + "eval_negation-triplets_steps_per_second": 1.696, + "step": 642 + }, + { + "epoch": 0.7508771929824561, + "eval_scitail-pairs-pos_loss": 0.05928462743759155, + "eval_scitail-pairs-pos_runtime": 0.8353, + "eval_scitail-pairs-pos_samples_per_second": 153.247, + "eval_scitail-pairs-pos_steps_per_second": 1.197, + "step": 642 + }, + { + "epoch": 0.7508771929824561, + "eval_scitail-pairs-qa_loss": 2.8886071845590777e-07, + "eval_scitail-pairs-qa_runtime": 0.58, + "eval_scitail-pairs-qa_samples_per_second": 220.677, + "eval_scitail-pairs-qa_steps_per_second": 1.724, + "step": 642 + }, + { + "epoch": 0.7508771929824561, + "eval_xsum-pairs_loss": 0.005155234597623348, + "eval_xsum-pairs_runtime": 2.8136, + "eval_xsum-pairs_samples_per_second": 45.494, + "eval_xsum-pairs_steps_per_second": 0.355, + "step": 642 + }, + { + "epoch": 0.7508771929824561, + "eval_sciq_pairs_loss": 0.0030471039935946465, + "eval_sciq_pairs_runtime": 4.0962, + "eval_sciq_pairs_samples_per_second": 31.248, + "eval_sciq_pairs_steps_per_second": 0.244, + "step": 642 + }, + { + "epoch": 0.7508771929824561, + "eval_qasc_pairs_loss": 0.05629788711667061, + "eval_qasc_pairs_runtime": 0.5944, + "eval_qasc_pairs_samples_per_second": 215.353, + "eval_qasc_pairs_steps_per_second": 1.682, + "step": 642 + }, + { + "epoch": 0.7508771929824561, + "eval_openbookqa_pairs_loss": 1.2897449731826782, + "eval_openbookqa_pairs_runtime": 0.5597, + "eval_openbookqa_pairs_samples_per_second": 228.687, + "eval_openbookqa_pairs_steps_per_second": 1.787, + "step": 642 + }, + { + "epoch": 0.7508771929824561, + "eval_msmarco_pairs_loss": 0.14785167574882507, + "eval_msmarco_pairs_runtime": 1.3218, + "eval_msmarco_pairs_samples_per_second": 96.836, + "eval_msmarco_pairs_steps_per_second": 0.757, + "step": 642 + }, + { + "epoch": 0.7508771929824561, + "eval_nq_pairs_loss": 0.04229257255792618, + "eval_nq_pairs_runtime": 2.9207, + "eval_nq_pairs_samples_per_second": 43.825, + "eval_nq_pairs_steps_per_second": 0.342, + "step": 642 + }, + { + "epoch": 0.7508771929824561, + "eval_trivia_pairs_loss": 0.2112799882888794, + "eval_trivia_pairs_runtime": 3.1802, + "eval_trivia_pairs_samples_per_second": 40.249, + "eval_trivia_pairs_steps_per_second": 0.314, + "step": 642 + }, + { + "epoch": 0.7508771929824561, + "eval_gooaq_pairs_loss": 0.11469888687133789, + "eval_gooaq_pairs_runtime": 0.8601, + "eval_gooaq_pairs_samples_per_second": 148.825, + "eval_gooaq_pairs_steps_per_second": 1.163, + "step": 642 + }, + { + "epoch": 0.7508771929824561, + "eval_paws-pos_loss": 0.0460214801132679, + "eval_paws-pos_runtime": 0.6612, + "eval_paws-pos_samples_per_second": 193.581, + "eval_paws-pos_steps_per_second": 1.512, + "step": 642 + }, + { + "epoch": 0.7508771929824561, + "eval_global_dataset_loss": 0.20304369926452637, + "eval_global_dataset_runtime": 24.1397, + "eval_global_dataset_samples_per_second": 27.465, + "eval_global_dataset_steps_per_second": 0.249, + "step": 642 + }, + { + "epoch": 0.7590643274853801, + "grad_norm": 0.5546406507492065, + "learning_rate": 2.0698086463501066e-05, + "loss": 0.1919, + "step": 649 + }, + { + "epoch": 0.7719298245614035, + "grad_norm": 0.16179773211479187, + "learning_rate": 2.1048901488306166e-05, + "loss": 0.1395, + "step": 660 + }, + { + "epoch": 0.7847953216374269, + "grad_norm": 0.4947220981121063, + "learning_rate": 2.139971651311127e-05, + "loss": 0.2047, + "step": 671 + }, + { + "epoch": 0.7976608187134503, + "grad_norm": 0.4697045683860779, + "learning_rate": 2.1750531537916374e-05, + "loss": 0.1421, + "step": 682 + }, + { + "epoch": 0.8105263157894737, + "grad_norm": 0.7013148665428162, + "learning_rate": 2.2101346562721477e-05, + "loss": 0.1227, + "step": 693 + }, + { + "epoch": 0.8233918128654971, + "grad_norm": 1.1045671701431274, + "learning_rate": 2.2452161587526577e-05, + "loss": 0.1235, + "step": 704 + }, + { + "epoch": 0.8362573099415205, + "grad_norm": 1.5613625049591064, + "learning_rate": 2.280297661233168e-05, + "loss": 0.2004, + "step": 715 + }, + { + "epoch": 0.8491228070175438, + "grad_norm": 0.23709431290626526, + "learning_rate": 2.3153791637136788e-05, + "loss": 0.1568, + "step": 726 + }, + { + "epoch": 0.8619883040935673, + "grad_norm": 1.1035298109054565, + "learning_rate": 2.3504606661941888e-05, + "loss": 0.1598, + "step": 737 + }, + { + "epoch": 0.8748538011695907, + "grad_norm": 0.7472051978111267, + "learning_rate": 2.385542168674699e-05, + "loss": 0.1328, + "step": 748 + }, + { + "epoch": 0.887719298245614, + "grad_norm": 0.57217937707901, + "learning_rate": 2.420623671155209e-05, + "loss": 0.0999, + "step": 759 + }, + { + "epoch": 0.9005847953216374, + "grad_norm": 0.3447597920894623, + "learning_rate": 2.4557051736357198e-05, + "loss": 0.1058, + "step": 770 + }, + { + "epoch": 0.9134502923976608, + "grad_norm": 1.0243078470230103, + "learning_rate": 2.4907866761162295e-05, + "loss": 0.1673, + "step": 781 + }, + { + "epoch": 0.9263157894736842, + "grad_norm": 1.0685806274414062, + "learning_rate": 2.5258681785967402e-05, + "loss": 0.1905, + "step": 792 + }, + { + "epoch": 0.9391812865497076, + "grad_norm": 1.1175950765609741, + "learning_rate": 2.5609496810772502e-05, + "loss": 0.1463, + "step": 803 + }, + { + "epoch": 0.952046783625731, + "grad_norm": 0.13670021295547485, + "learning_rate": 2.5960311835577605e-05, + "loss": 0.1294, + "step": 814 + }, + { + "epoch": 0.9649122807017544, + "grad_norm": 0.6462442278862, + "learning_rate": 2.6311126860382706e-05, + "loss": 0.1312, + "step": 825 + }, + { + "epoch": 0.9777777777777777, + "grad_norm": 0.1836312711238861, + "learning_rate": 2.6661941885187812e-05, + "loss": 0.1308, + "step": 836 + }, + { + "epoch": 0.9906432748538012, + "grad_norm": 4.2791407395270653e-07, + "learning_rate": 2.7012756909992913e-05, + "loss": 0.1076, + "step": 847 + }, + { + "epoch": 1.001169590643275, + "eval_Qnli-dev_cosine_accuracy": 0.69140625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.657210111618042, + "eval_Qnli-dev_cosine_ap": 0.714041144664227, + "eval_Qnli-dev_cosine_f1": 0.6783439490445861, + "eval_Qnli-dev_cosine_f1_threshold": 0.48659271001815796, + "eval_Qnli-dev_cosine_precision": 0.5433673469387755, + "eval_Qnli-dev_cosine_recall": 0.902542372881356, + "eval_Qnli-dev_dot_accuracy": 0.669921875, + "eval_Qnli-dev_dot_accuracy_threshold": 256.53155517578125, + "eval_Qnli-dev_dot_ap": 0.6648303273464243, + "eval_Qnli-dev_dot_f1": 0.678125, + "eval_Qnli-dev_dot_f1_threshold": 178.63389587402344, + "eval_Qnli-dev_dot_precision": 0.5371287128712872, + "eval_Qnli-dev_dot_recall": 0.9194915254237288, + "eval_Qnli-dev_euclidean_accuracy": 0.6953125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 16.298263549804688, + "eval_Qnli-dev_euclidean_ap": 0.7198419743004867, + "eval_Qnli-dev_euclidean_f1": 0.6771929824561403, + "eval_Qnli-dev_euclidean_f1_threshold": 18.734228134155273, + "eval_Qnli-dev_euclidean_precision": 0.5778443113772455, + "eval_Qnli-dev_euclidean_recall": 0.8177966101694916, + "eval_Qnli-dev_manhattan_accuracy": 0.697265625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 350.3825988769531, + "eval_Qnli-dev_manhattan_ap": 0.7225583351616828, + "eval_Qnli-dev_manhattan_f1": 0.6733466933867736, + "eval_Qnli-dev_manhattan_f1_threshold": 364.5922546386719, + "eval_Qnli-dev_manhattan_precision": 0.6387832699619772, + "eval_Qnli-dev_manhattan_recall": 0.711864406779661, + "eval_Qnli-dev_max_accuracy": 0.697265625, + "eval_Qnli-dev_max_accuracy_threshold": 350.3825988769531, + "eval_Qnli-dev_max_ap": 0.7225583351616828, + "eval_Qnli-dev_max_f1": 0.6783439490445861, + "eval_Qnli-dev_max_f1_threshold": 364.5922546386719, + "eval_Qnli-dev_max_precision": 0.6387832699619772, + "eval_Qnli-dev_max_recall": 0.9194915254237288, + "eval_allNLI-dev_cosine_accuracy": 0.73828125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7839766144752502, + "eval_allNLI-dev_cosine_ap": 0.6377040389504978, + "eval_allNLI-dev_cosine_f1": 0.6371681415929203, + "eval_allNLI-dev_cosine_f1_threshold": 0.5921227931976318, + "eval_allNLI-dev_cosine_precision": 0.5161290322580645, + "eval_allNLI-dev_cosine_recall": 0.8323699421965318, + "eval_allNLI-dev_dot_accuracy": 0.72265625, + "eval_allNLI-dev_dot_accuracy_threshold": 287.0504150390625, + "eval_allNLI-dev_dot_ap": 0.5976338107788919, + "eval_allNLI-dev_dot_f1": 0.615, + "eval_allNLI-dev_dot_f1_threshold": 242.96902465820312, + "eval_allNLI-dev_dot_precision": 0.5418502202643172, + "eval_allNLI-dev_dot_recall": 0.7109826589595376, + "eval_allNLI-dev_euclidean_accuracy": 0.740234375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 13.528583526611328, + "eval_allNLI-dev_euclidean_ap": 0.6364974679625237, + "eval_allNLI-dev_euclidean_f1": 0.6419753086419753, + "eval_allNLI-dev_euclidean_f1_threshold": 16.527751922607422, + "eval_allNLI-dev_euclidean_precision": 0.5603448275862069, + "eval_allNLI-dev_euclidean_recall": 0.7514450867052023, + "eval_allNLI-dev_manhattan_accuracy": 0.73828125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 287.6244812011719, + "eval_allNLI-dev_manhattan_ap": 0.6359733229236403, + "eval_allNLI-dev_manhattan_f1": 0.6461538461538462, + "eval_allNLI-dev_manhattan_f1_threshold": 341.6584167480469, + "eval_allNLI-dev_manhattan_precision": 0.5806451612903226, + "eval_allNLI-dev_manhattan_recall": 0.7283236994219653, + "eval_allNLI-dev_max_accuracy": 0.740234375, + "eval_allNLI-dev_max_accuracy_threshold": 287.6244812011719, + "eval_allNLI-dev_max_ap": 0.6377040389504978, + "eval_allNLI-dev_max_f1": 0.6461538461538462, + "eval_allNLI-dev_max_f1_threshold": 341.6584167480469, + "eval_allNLI-dev_max_precision": 0.5806451612903226, + "eval_allNLI-dev_max_recall": 0.8323699421965318, + "eval_sequential_score": 0.7225583351616828, + "eval_sts-test_pearson_cosine": 0.8846526004146165, + "eval_sts-test_pearson_dot": 0.8694917588757636, + "eval_sts-test_pearson_euclidean": 0.903648547377558, + "eval_sts-test_pearson_manhattan": 0.9046900352481024, + "eval_sts-test_pearson_max": 0.9046900352481024, + "eval_sts-test_spearman_cosine": 0.9069678627017289, + "eval_sts-test_spearman_dot": 0.8661496554202353, + "eval_sts-test_spearman_euclidean": 0.8991392112877015, + "eval_sts-test_spearman_manhattan": 0.9004067012640901, + "eval_sts-test_spearman_max": 0.9069678627017289, + "eval_vitaminc-pairs_loss": 2.658703327178955, + "eval_vitaminc-pairs_runtime": 3.7894, + "eval_vitaminc-pairs_samples_per_second": 33.778, + "eval_vitaminc-pairs_steps_per_second": 0.264, + "step": 856 + }, + { + "epoch": 1.001169590643275, + "eval_negation-triplets_loss": 1.048051357269287, + "eval_negation-triplets_runtime": 0.5948, + "eval_negation-triplets_samples_per_second": 215.195, + "eval_negation-triplets_steps_per_second": 1.681, + "step": 856 + }, + { + "epoch": 1.001169590643275, + "eval_scitail-pairs-pos_loss": 0.05958803743124008, + "eval_scitail-pairs-pos_runtime": 0.8408, + "eval_scitail-pairs-pos_samples_per_second": 152.231, + "eval_scitail-pairs-pos_steps_per_second": 1.189, + "step": 856 + }, + { + "epoch": 1.001169590643275, + "eval_scitail-pairs-qa_loss": 4.770551527144562e-07, + "eval_scitail-pairs-qa_runtime": 0.5629, + "eval_scitail-pairs-qa_samples_per_second": 227.394, + "eval_scitail-pairs-qa_steps_per_second": 1.777, + "step": 856 + }, + { + "epoch": 1.001169590643275, + "eval_xsum-pairs_loss": 0.005623215343803167, + "eval_xsum-pairs_runtime": 2.7656, + "eval_xsum-pairs_samples_per_second": 46.283, + "eval_xsum-pairs_steps_per_second": 0.362, + "step": 856 + }, + { + "epoch": 1.001169590643275, + "eval_sciq_pairs_loss": 0.0031388087663799524, + "eval_sciq_pairs_runtime": 4.099, + "eval_sciq_pairs_samples_per_second": 31.227, + "eval_sciq_pairs_steps_per_second": 0.244, + "step": 856 + }, + { + "epoch": 1.001169590643275, + "eval_qasc_pairs_loss": 0.05423235148191452, + "eval_qasc_pairs_runtime": 0.596, + "eval_qasc_pairs_samples_per_second": 214.767, + "eval_qasc_pairs_steps_per_second": 1.678, + "step": 856 + }, + { + "epoch": 1.001169590643275, + "eval_openbookqa_pairs_loss": 1.2992339134216309, + "eval_openbookqa_pairs_runtime": 0.5529, + "eval_openbookqa_pairs_samples_per_second": 231.507, + "eval_openbookqa_pairs_steps_per_second": 1.809, + "step": 856 + }, + { + "epoch": 1.001169590643275, + "eval_msmarco_pairs_loss": 0.15928690135478973, + "eval_msmarco_pairs_runtime": 1.3201, + "eval_msmarco_pairs_samples_per_second": 96.963, + "eval_msmarco_pairs_steps_per_second": 0.758, + "step": 856 + }, + { + "epoch": 1.001169590643275, + "eval_nq_pairs_loss": 0.04052837938070297, + "eval_nq_pairs_runtime": 2.9252, + "eval_nq_pairs_samples_per_second": 43.758, + "eval_nq_pairs_steps_per_second": 0.342, + "step": 856 + }, + { + "epoch": 1.001169590643275, + "eval_trivia_pairs_loss": 0.21009953320026398, + "eval_trivia_pairs_runtime": 3.1824, + "eval_trivia_pairs_samples_per_second": 40.221, + "eval_trivia_pairs_steps_per_second": 0.314, + "step": 856 + }, + { + "epoch": 1.001169590643275, + "eval_gooaq_pairs_loss": 0.11401581764221191, + "eval_gooaq_pairs_runtime": 0.8553, + "eval_gooaq_pairs_samples_per_second": 149.66, + "eval_gooaq_pairs_steps_per_second": 1.169, + "step": 856 + }, + { + "epoch": 1.001169590643275, + "eval_paws-pos_loss": 0.04620524123311043, + "eval_paws-pos_runtime": 0.6561, + "eval_paws-pos_samples_per_second": 195.099, + "eval_paws-pos_steps_per_second": 1.524, + "step": 856 + }, + { + "epoch": 1.001169590643275, + "eval_global_dataset_loss": 0.20265640318393707, + "eval_global_dataset_runtime": 24.1741, + "eval_global_dataset_samples_per_second": 27.426, + "eval_global_dataset_steps_per_second": 0.248, + "step": 856 + }, + { + "epoch": 1.0035087719298246, + "grad_norm": 1.2066627740859985, + "learning_rate": 2.7363571934798016e-05, + "loss": 0.1085, + "step": 858 + }, + { + "epoch": 1.016374269005848, + "grad_norm": 0.3020640015602112, + "learning_rate": 2.7714386959603123e-05, + "loss": 0.2214, + "step": 869 + }, + { + "epoch": 1.0292397660818713, + "grad_norm": 0.9783104062080383, + "learning_rate": 2.8065201984408223e-05, + "loss": 0.1214, + "step": 880 + }, + { + "epoch": 1.0421052631578946, + "grad_norm": 0.24499095976352692, + "learning_rate": 2.8416017009213327e-05, + "loss": 0.1049, + "step": 891 + }, + { + "epoch": 1.0549707602339182, + "grad_norm": 0.5718978643417358, + "learning_rate": 2.8766832034018427e-05, + "loss": 0.1897, + "step": 902 + }, + { + "epoch": 1.0678362573099416, + "grad_norm": 0.558549165725708, + "learning_rate": 2.9117647058823534e-05, + "loss": 0.1273, + "step": 913 + }, + { + "epoch": 1.080701754385965, + "grad_norm": 0.014839976094663143, + "learning_rate": 2.9468462083628634e-05, + "loss": 0.1474, + "step": 924 + }, + { + "epoch": 1.0935672514619883, + "grad_norm": 0.28388065099716187, + "learning_rate": 2.9819277108433737e-05, + "loss": 0.1313, + "step": 935 + }, + { + "epoch": 1.1064327485380117, + "grad_norm": 0.28269803524017334, + "learning_rate": 3.0170092133238837e-05, + "loss": 0.1769, + "step": 946 + }, + { + "epoch": 1.119298245614035, + "grad_norm": 0.910582959651947, + "learning_rate": 3.052090715804394e-05, + "loss": 0.143, + "step": 957 + }, + { + "epoch": 1.1321637426900586, + "grad_norm": 0.5537557601928711, + "learning_rate": 3.087172218284904e-05, + "loss": 0.1968, + "step": 968 + }, + { + "epoch": 1.145029239766082, + "grad_norm": 0.9547837376594543, + "learning_rate": 3.122253720765415e-05, + "loss": 0.1771, + "step": 979 + }, + { + "epoch": 1.1578947368421053, + "grad_norm": 1.1063166856765747, + "learning_rate": 3.157335223245925e-05, + "loss": 0.1822, + "step": 990 + }, + { + "epoch": 1.1707602339181287, + "grad_norm": 0.3083917796611786, + "learning_rate": 3.1924167257264355e-05, + "loss": 0.2467, + "step": 1001 + }, + { + "epoch": 1.183625730994152, + "grad_norm": 0.8678072690963745, + "learning_rate": 3.227498228206946e-05, + "loss": 0.1419, + "step": 1012 + }, + { + "epoch": 1.1964912280701754, + "grad_norm": 1.0385596752166748, + "learning_rate": 3.262579730687456e-05, + "loss": 0.1782, + "step": 1023 + }, + { + "epoch": 1.2093567251461987, + "grad_norm": 1.0005829334259033, + "learning_rate": 3.297661233167966e-05, + "loss": 0.1297, + "step": 1034 + }, + { + "epoch": 1.2222222222222223, + "grad_norm": 0.7221047282218933, + "learning_rate": 3.332742735648476e-05, + "loss": 0.1972, + "step": 1045 + }, + { + "epoch": 1.2350877192982457, + "grad_norm": 0.20262663066387177, + "learning_rate": 3.367824238128987e-05, + "loss": 0.1491, + "step": 1056 + }, + { + "epoch": 1.247953216374269, + "grad_norm": 0.7044623494148254, + "learning_rate": 3.402905740609497e-05, + "loss": 0.1721, + "step": 1067 + }, + { + "epoch": 1.2514619883040936, + "eval_Qnli-dev_cosine_accuracy": 0.69140625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6546923518180847, + "eval_Qnli-dev_cosine_ap": 0.7153621655342896, + "eval_Qnli-dev_cosine_f1": 0.6783439490445861, + "eval_Qnli-dev_cosine_f1_threshold": 0.4858614206314087, + "eval_Qnli-dev_cosine_precision": 0.5433673469387755, + "eval_Qnli-dev_cosine_recall": 0.902542372881356, + "eval_Qnli-dev_dot_accuracy": 0.671875, + "eval_Qnli-dev_dot_accuracy_threshold": 254.48748779296875, + "eval_Qnli-dev_dot_ap": 0.6630076505293763, + "eval_Qnli-dev_dot_f1": 0.6740157480314961, + "eval_Qnli-dev_dot_f1_threshold": 183.4345703125, + "eval_Qnli-dev_dot_precision": 0.5363408521303258, + "eval_Qnli-dev_dot_recall": 0.9067796610169492, + "eval_Qnli-dev_euclidean_accuracy": 0.701171875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 16.410430908203125, + "eval_Qnli-dev_euclidean_ap": 0.7221448194401613, + "eval_Qnli-dev_euclidean_f1": 0.6811881188118811, + "eval_Qnli-dev_euclidean_f1_threshold": 17.367481231689453, + "eval_Qnli-dev_euclidean_precision": 0.6394052044609665, + "eval_Qnli-dev_euclidean_recall": 0.7288135593220338, + "eval_Qnli-dev_manhattan_accuracy": 0.6953125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 353.0638122558594, + "eval_Qnli-dev_manhattan_ap": 0.7247492174165924, + "eval_Qnli-dev_manhattan_f1": 0.6756238003838773, + "eval_Qnli-dev_manhattan_f1_threshold": 373.6475524902344, + "eval_Qnli-dev_manhattan_precision": 0.6175438596491228, + "eval_Qnli-dev_manhattan_recall": 0.7457627118644068, + "eval_Qnli-dev_max_accuracy": 0.701171875, + "eval_Qnli-dev_max_accuracy_threshold": 353.0638122558594, + "eval_Qnli-dev_max_ap": 0.7247492174165924, + "eval_Qnli-dev_max_f1": 0.6811881188118811, + "eval_Qnli-dev_max_f1_threshold": 373.6475524902344, + "eval_Qnli-dev_max_precision": 0.6394052044609665, + "eval_Qnli-dev_max_recall": 0.9067796610169492, + "eval_allNLI-dev_cosine_accuracy": 0.740234375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7620664834976196, + "eval_allNLI-dev_cosine_ap": 0.6352023549250958, + "eval_allNLI-dev_cosine_f1": 0.6401766004415012, + "eval_allNLI-dev_cosine_f1_threshold": 0.5983242988586426, + "eval_allNLI-dev_cosine_precision": 0.5178571428571429, + "eval_allNLI-dev_cosine_recall": 0.838150289017341, + "eval_allNLI-dev_dot_accuracy": 0.732421875, + "eval_allNLI-dev_dot_accuracy_threshold": 282.25836181640625, + "eval_allNLI-dev_dot_ap": 0.5938709922161639, + "eval_allNLI-dev_dot_f1": 0.612987012987013, + "eval_allNLI-dev_dot_f1_threshold": 245.4638671875, + "eval_allNLI-dev_dot_precision": 0.5566037735849056, + "eval_allNLI-dev_dot_recall": 0.6820809248554913, + "eval_allNLI-dev_euclidean_accuracy": 0.740234375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 12.869304656982422, + "eval_allNLI-dev_euclidean_ap": 0.6367777728048029, + "eval_allNLI-dev_euclidean_f1": 0.6458797327394209, + "eval_allNLI-dev_euclidean_f1_threshold": 17.27260971069336, + "eval_allNLI-dev_euclidean_precision": 0.5253623188405797, + "eval_allNLI-dev_euclidean_recall": 0.838150289017341, + "eval_allNLI-dev_manhattan_accuracy": 0.7421875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 297.8420715332031, + "eval_allNLI-dev_manhattan_ap": 0.6368910837780656, + "eval_allNLI-dev_manhattan_f1": 0.6477024070021882, + "eval_allNLI-dev_manhattan_f1_threshold": 369.8564453125, + "eval_allNLI-dev_manhattan_precision": 0.5211267605633803, + "eval_allNLI-dev_manhattan_recall": 0.8554913294797688, + "eval_allNLI-dev_max_accuracy": 0.7421875, + "eval_allNLI-dev_max_accuracy_threshold": 297.8420715332031, + "eval_allNLI-dev_max_ap": 0.6368910837780656, + "eval_allNLI-dev_max_f1": 0.6477024070021882, + "eval_allNLI-dev_max_f1_threshold": 369.8564453125, + "eval_allNLI-dev_max_precision": 0.5566037735849056, + "eval_allNLI-dev_max_recall": 0.8554913294797688, + "eval_sequential_score": 0.7247492174165924, + "eval_sts-test_pearson_cosine": 0.884621677997079, + "eval_sts-test_pearson_dot": 0.8662022984817861, + "eval_sts-test_pearson_euclidean": 0.9042360469838505, + "eval_sts-test_pearson_manhattan": 0.9051133756387458, + "eval_sts-test_pearson_max": 0.9051133756387458, + "eval_sts-test_spearman_cosine": 0.9071986449567242, + "eval_sts-test_spearman_dot": 0.8602733672869306, + "eval_sts-test_spearman_euclidean": 0.8999521179508763, + "eval_sts-test_spearman_manhattan": 0.901083564092635, + "eval_sts-test_spearman_max": 0.9071986449567242, + "eval_vitaminc-pairs_loss": 2.655998468399048, + "eval_vitaminc-pairs_runtime": 3.8396, + "eval_vitaminc-pairs_samples_per_second": 33.337, + "eval_vitaminc-pairs_steps_per_second": 0.26, + "step": 1070 + }, + { + "epoch": 1.2514619883040936, + "eval_negation-triplets_loss": 1.0599219799041748, + "eval_negation-triplets_runtime": 0.6284, + "eval_negation-triplets_samples_per_second": 203.693, + "eval_negation-triplets_steps_per_second": 1.591, + "step": 1070 + }, + { + "epoch": 1.2514619883040936, + "eval_scitail-pairs-pos_loss": 0.061028748750686646, + "eval_scitail-pairs-pos_runtime": 0.9206, + "eval_scitail-pairs-pos_samples_per_second": 139.044, + "eval_scitail-pairs-pos_steps_per_second": 1.086, + "step": 1070 + }, + { + "epoch": 1.2514619883040936, + "eval_scitail-pairs-qa_loss": 4.0994524397319765e-07, + "eval_scitail-pairs-qa_runtime": 0.5941, + "eval_scitail-pairs-qa_samples_per_second": 215.464, + "eval_scitail-pairs-qa_steps_per_second": 1.683, + "step": 1070 + }, + { + "epoch": 1.2514619883040936, + "eval_xsum-pairs_loss": 0.005265752784907818, + "eval_xsum-pairs_runtime": 2.8061, + "eval_xsum-pairs_samples_per_second": 45.615, + "eval_xsum-pairs_steps_per_second": 0.356, + "step": 1070 + }, + { + "epoch": 1.2514619883040936, + "eval_sciq_pairs_loss": 0.0029235396068543196, + "eval_sciq_pairs_runtime": 4.2052, + "eval_sciq_pairs_samples_per_second": 30.438, + "eval_sciq_pairs_steps_per_second": 0.238, + "step": 1070 + }, + { + "epoch": 1.2514619883040936, + "eval_qasc_pairs_loss": 0.05088697001338005, + "eval_qasc_pairs_runtime": 0.6097, + "eval_qasc_pairs_samples_per_second": 209.926, + "eval_qasc_pairs_steps_per_second": 1.64, + "step": 1070 + }, + { + "epoch": 1.2514619883040936, + "eval_openbookqa_pairs_loss": 1.281742811203003, + "eval_openbookqa_pairs_runtime": 0.5692, + "eval_openbookqa_pairs_samples_per_second": 224.86, + "eval_openbookqa_pairs_steps_per_second": 1.757, + "step": 1070 + }, + { + "epoch": 1.2514619883040936, + "eval_msmarco_pairs_loss": 0.1445852667093277, + "eval_msmarco_pairs_runtime": 1.3418, + "eval_msmarco_pairs_samples_per_second": 95.392, + "eval_msmarco_pairs_steps_per_second": 0.745, + "step": 1070 + }, + { + "epoch": 1.2514619883040936, + "eval_nq_pairs_loss": 0.041951049119234085, + "eval_nq_pairs_runtime": 2.9501, + "eval_nq_pairs_samples_per_second": 43.389, + "eval_nq_pairs_steps_per_second": 0.339, + "step": 1070 + }, + { + "epoch": 1.2514619883040936, + "eval_trivia_pairs_loss": 0.2087937891483307, + "eval_trivia_pairs_runtime": 3.2077, + "eval_trivia_pairs_samples_per_second": 39.904, + "eval_trivia_pairs_steps_per_second": 0.312, + "step": 1070 + }, + { + "epoch": 1.2514619883040936, + "eval_gooaq_pairs_loss": 0.11122985929250717, + "eval_gooaq_pairs_runtime": 0.8735, + "eval_gooaq_pairs_samples_per_second": 146.542, + "eval_gooaq_pairs_steps_per_second": 1.145, + "step": 1070 + }, + { + "epoch": 1.2514619883040936, + "eval_paws-pos_loss": 0.04601467400789261, + "eval_paws-pos_runtime": 0.6727, + "eval_paws-pos_samples_per_second": 190.278, + "eval_paws-pos_steps_per_second": 1.487, + "step": 1070 + }, + { + "epoch": 1.2514619883040936, + "eval_global_dataset_loss": 0.21354959905147552, + "eval_global_dataset_runtime": 24.3059, + "eval_global_dataset_samples_per_second": 27.277, + "eval_global_dataset_steps_per_second": 0.247, + "step": 1070 + }, + { + "epoch": 1.2608187134502924, + "grad_norm": 0.5193822383880615, + "learning_rate": 3.4379872430900076e-05, + "loss": 0.1279, + "step": 1078 + }, + { + "epoch": 1.2736842105263158, + "grad_norm": 1.694514513015747, + "learning_rate": 3.4730687455705176e-05, + "loss": 0.106, + "step": 1089 + }, + { + "epoch": 1.286549707602339, + "grad_norm": 0.8599225282669067, + "learning_rate": 3.5081502480510276e-05, + "loss": 0.1597, + "step": 1100 + }, + { + "epoch": 1.2994152046783625, + "grad_norm": 0.7980732917785645, + "learning_rate": 3.5432317505315376e-05, + "loss": 0.192, + "step": 1111 + }, + { + "epoch": 1.312280701754386, + "grad_norm": 0.33019399642944336, + "learning_rate": 3.578313253012048e-05, + "loss": 0.165, + "step": 1122 + }, + { + "epoch": 1.3251461988304094, + "grad_norm": 0.006413115654140711, + "learning_rate": 3.6133947554925583e-05, + "loss": 0.1472, + "step": 1133 + }, + { + "epoch": 1.3380116959064328, + "grad_norm": 0.38985970616340637, + "learning_rate": 3.648476257973069e-05, + "loss": 0.1528, + "step": 1144 + }, + { + "epoch": 1.3508771929824561, + "grad_norm": 1.1716166734695435, + "learning_rate": 3.68355776045358e-05, + "loss": 0.202, + "step": 1155 + }, + { + "epoch": 1.3637426900584795, + "grad_norm": 0.6012507081031799, + "learning_rate": 3.71863926293409e-05, + "loss": 0.1974, + "step": 1166 + }, + { + "epoch": 1.376608187134503, + "grad_norm": 0.15670593082904816, + "learning_rate": 3.7537207654146e-05, + "loss": 0.2229, + "step": 1177 + }, + { + "epoch": 1.3894736842105262, + "grad_norm": 0.5632084608078003, + "learning_rate": 3.78880226789511e-05, + "loss": 0.1104, + "step": 1188 + }, + { + "epoch": 1.4023391812865498, + "grad_norm": 0.7236007452011108, + "learning_rate": 3.8238837703756205e-05, + "loss": 0.1544, + "step": 1199 + }, + { + "epoch": 1.4152046783625731, + "grad_norm": 0.22364391386508942, + "learning_rate": 3.8589652728561305e-05, + "loss": 0.0875, + "step": 1210 + }, + { + "epoch": 1.4280701754385965, + "grad_norm": 0.23043423891067505, + "learning_rate": 3.894046775336641e-05, + "loss": 0.1607, + "step": 1221 + }, + { + "epoch": 1.4409356725146198, + "grad_norm": 0.7280920147895813, + "learning_rate": 3.929128277817151e-05, + "loss": 0.2026, + "step": 1232 + }, + { + "epoch": 1.4538011695906432, + "grad_norm": 0.1936722993850708, + "learning_rate": 3.964209780297662e-05, + "loss": 0.185, + "step": 1243 + }, + { + "epoch": 1.4666666666666668, + "grad_norm": 0.4526028633117676, + "learning_rate": 3.999291282778172e-05, + "loss": 0.1114, + "step": 1254 + }, + { + "epoch": 1.47953216374269, + "grad_norm": 0.2208670973777771, + "learning_rate": 4.034372785258682e-05, + "loss": 0.2033, + "step": 1265 + }, + { + "epoch": 1.4923976608187135, + "grad_norm": 1.5043717622756958, + "learning_rate": 4.069454287739192e-05, + "loss": 0.1216, + "step": 1276 + }, + { + "epoch": 1.5017543859649123, + "eval_Qnli-dev_cosine_accuracy": 0.69140625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6479891538619995, + "eval_Qnli-dev_cosine_ap": 0.7145834476822306, + "eval_Qnli-dev_cosine_f1": 0.6772655007949125, + "eval_Qnli-dev_cosine_f1_threshold": 0.4831158518791199, + "eval_Qnli-dev_cosine_precision": 0.5419847328244275, + "eval_Qnli-dev_cosine_recall": 0.902542372881356, + "eval_Qnli-dev_dot_accuracy": 0.669921875, + "eval_Qnli-dev_dot_accuracy_threshold": 253.99070739746094, + "eval_Qnli-dev_dot_ap": 0.6600609959091455, + "eval_Qnli-dev_dot_f1": 0.6749226006191951, + "eval_Qnli-dev_dot_f1_threshold": 177.03799438476562, + "eval_Qnli-dev_dot_precision": 0.5317073170731708, + "eval_Qnli-dev_dot_recall": 0.923728813559322, + "eval_Qnli-dev_euclidean_accuracy": 0.703125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 16.42089080810547, + "eval_Qnli-dev_euclidean_ap": 0.7233470144987824, + "eval_Qnli-dev_euclidean_f1": 0.6786427145708582, + "eval_Qnli-dev_euclidean_f1_threshold": 17.447650909423828, + "eval_Qnli-dev_euclidean_precision": 0.6415094339622641, + "eval_Qnli-dev_euclidean_recall": 0.7203389830508474, + "eval_Qnli-dev_manhattan_accuracy": 0.69921875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 348.24969482421875, + "eval_Qnli-dev_manhattan_ap": 0.7248042664576803, + "eval_Qnli-dev_manhattan_f1": 0.6735966735966736, + "eval_Qnli-dev_manhattan_f1_threshold": 363.0526123046875, + "eval_Qnli-dev_manhattan_precision": 0.6612244897959184, + "eval_Qnli-dev_manhattan_recall": 0.6864406779661016, + "eval_Qnli-dev_max_accuracy": 0.703125, + "eval_Qnli-dev_max_accuracy_threshold": 348.24969482421875, + "eval_Qnli-dev_max_ap": 0.7248042664576803, + "eval_Qnli-dev_max_f1": 0.6786427145708582, + "eval_Qnli-dev_max_f1_threshold": 363.0526123046875, + "eval_Qnli-dev_max_precision": 0.6612244897959184, + "eval_Qnli-dev_max_recall": 0.923728813559322, + "eval_allNLI-dev_cosine_accuracy": 0.740234375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7793502807617188, + "eval_allNLI-dev_cosine_ap": 0.6356059008560961, + "eval_allNLI-dev_cosine_f1": 0.6412556053811659, + "eval_allNLI-dev_cosine_f1_threshold": 0.5976245999336243, + "eval_allNLI-dev_cosine_precision": 0.5238095238095238, + "eval_allNLI-dev_cosine_recall": 0.8265895953757225, + "eval_allNLI-dev_dot_accuracy": 0.728515625, + "eval_allNLI-dev_dot_accuracy_threshold": 289.59637451171875, + "eval_allNLI-dev_dot_ap": 0.6011407360027551, + "eval_allNLI-dev_dot_f1": 0.6177215189873418, + "eval_allNLI-dev_dot_f1_threshold": 246.49258422851562, + "eval_allNLI-dev_dot_precision": 0.5495495495495496, + "eval_allNLI-dev_dot_recall": 0.7052023121387283, + "eval_allNLI-dev_euclidean_accuracy": 0.744140625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 13.19027328491211, + "eval_allNLI-dev_euclidean_ap": 0.6358049693529386, + "eval_allNLI-dev_euclidean_f1": 0.6430155210643015, + "eval_allNLI-dev_euclidean_f1_threshold": 17.634546279907227, + "eval_allNLI-dev_euclidean_precision": 0.5215827338129496, + "eval_allNLI-dev_euclidean_recall": 0.838150289017341, + "eval_allNLI-dev_manhattan_accuracy": 0.73828125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 281.0400085449219, + "eval_allNLI-dev_manhattan_ap": 0.6342668702497214, + "eval_allNLI-dev_manhattan_f1": 0.6512820512820513, + "eval_allNLI-dev_manhattan_f1_threshold": 342.90478515625, + "eval_allNLI-dev_manhattan_precision": 0.5852534562211982, + "eval_allNLI-dev_manhattan_recall": 0.7341040462427746, + "eval_allNLI-dev_max_accuracy": 0.744140625, + "eval_allNLI-dev_max_accuracy_threshold": 289.59637451171875, + "eval_allNLI-dev_max_ap": 0.6358049693529386, + "eval_allNLI-dev_max_f1": 0.6512820512820513, + "eval_allNLI-dev_max_f1_threshold": 342.90478515625, + "eval_allNLI-dev_max_precision": 0.5852534562211982, + "eval_allNLI-dev_max_recall": 0.838150289017341, + "eval_sequential_score": 0.7248042664576803, + "eval_sts-test_pearson_cosine": 0.8852081089755639, + "eval_sts-test_pearson_dot": 0.8700618346105049, + "eval_sts-test_pearson_euclidean": 0.904503930603612, + "eval_sts-test_pearson_manhattan": 0.905215331612789, + "eval_sts-test_pearson_max": 0.905215331612789, + "eval_sts-test_spearman_cosine": 0.9076046731513284, + "eval_sts-test_spearman_dot": 0.8651591757518594, + "eval_sts-test_spearman_euclidean": 0.9000229144463976, + "eval_sts-test_spearman_manhattan": 0.9012659254828049, + "eval_sts-test_spearman_max": 0.9076046731513284, + "eval_vitaminc-pairs_loss": 2.694180727005005, + "eval_vitaminc-pairs_runtime": 3.8321, + "eval_vitaminc-pairs_samples_per_second": 33.402, + "eval_vitaminc-pairs_steps_per_second": 0.261, + "step": 1284 + }, + { + "epoch": 1.5017543859649123, + "eval_negation-triplets_loss": 1.049867033958435, + "eval_negation-triplets_runtime": 0.6019, + "eval_negation-triplets_samples_per_second": 212.664, + "eval_negation-triplets_steps_per_second": 1.661, + "step": 1284 + }, + { + "epoch": 1.5017543859649123, + "eval_scitail-pairs-pos_loss": 0.05977838858962059, + "eval_scitail-pairs-pos_runtime": 0.8746, + "eval_scitail-pairs-pos_samples_per_second": 146.356, + "eval_scitail-pairs-pos_steps_per_second": 1.143, + "step": 1284 + }, + { + "epoch": 1.5017543859649123, + "eval_scitail-pairs-qa_loss": 3.577365532692056e-07, + "eval_scitail-pairs-qa_runtime": 0.577, + "eval_scitail-pairs-qa_samples_per_second": 221.823, + "eval_scitail-pairs-qa_steps_per_second": 1.733, + "step": 1284 + }, + { + "epoch": 1.5017543859649123, + "eval_xsum-pairs_loss": 0.005321314092725515, + "eval_xsum-pairs_runtime": 2.7947, + "eval_xsum-pairs_samples_per_second": 45.801, + "eval_xsum-pairs_steps_per_second": 0.358, + "step": 1284 + }, + { + "epoch": 1.5017543859649123, + "eval_sciq_pairs_loss": 0.003324751742184162, + "eval_sciq_pairs_runtime": 4.2166, + "eval_sciq_pairs_samples_per_second": 30.356, + "eval_sciq_pairs_steps_per_second": 0.237, + "step": 1284 + }, + { + "epoch": 1.5017543859649123, + "eval_qasc_pairs_loss": 0.0436801016330719, + "eval_qasc_pairs_runtime": 0.6277, + "eval_qasc_pairs_samples_per_second": 203.908, + "eval_qasc_pairs_steps_per_second": 1.593, + "step": 1284 + }, + { + "epoch": 1.5017543859649123, + "eval_openbookqa_pairs_loss": 1.3073405027389526, + "eval_openbookqa_pairs_runtime": 0.5694, + "eval_openbookqa_pairs_samples_per_second": 224.81, + "eval_openbookqa_pairs_steps_per_second": 1.756, + "step": 1284 + }, + { + "epoch": 1.5017543859649123, + "eval_msmarco_pairs_loss": 0.15479393303394318, + "eval_msmarco_pairs_runtime": 1.3314, + "eval_msmarco_pairs_samples_per_second": 96.137, + "eval_msmarco_pairs_steps_per_second": 0.751, + "step": 1284 + }, + { + "epoch": 1.5017543859649123, + "eval_nq_pairs_loss": 0.03904910385608673, + "eval_nq_pairs_runtime": 2.9759, + "eval_nq_pairs_samples_per_second": 43.012, + "eval_nq_pairs_steps_per_second": 0.336, + "step": 1284 + }, + { + "epoch": 1.5017543859649123, + "eval_trivia_pairs_loss": 0.2038673609495163, + "eval_trivia_pairs_runtime": 3.1963, + "eval_trivia_pairs_samples_per_second": 40.047, + "eval_trivia_pairs_steps_per_second": 0.313, + "step": 1284 + }, + { + "epoch": 1.5017543859649123, + "eval_gooaq_pairs_loss": 0.10658062994480133, + "eval_gooaq_pairs_runtime": 0.8625, + "eval_gooaq_pairs_samples_per_second": 148.407, + "eval_gooaq_pairs_steps_per_second": 1.159, + "step": 1284 + }, + { + "epoch": 1.5017543859649123, + "eval_paws-pos_loss": 0.04552536457777023, + "eval_paws-pos_runtime": 0.6659, + "eval_paws-pos_samples_per_second": 192.226, + "eval_paws-pos_steps_per_second": 1.502, + "step": 1284 + }, + { + "epoch": 1.5017543859649123, + "eval_global_dataset_loss": 0.2039254754781723, + "eval_global_dataset_runtime": 24.2218, + "eval_global_dataset_samples_per_second": 27.372, + "eval_global_dataset_steps_per_second": 0.248, + "step": 1284 + }, + { + "epoch": 1.5052631578947369, + "grad_norm": 0.985459566116333, + "learning_rate": 4.1045357902197026e-05, + "loss": 0.1108, + "step": 1287 + }, + { + "epoch": 1.5181286549707602, + "grad_norm": 0.1348501294851303, + "learning_rate": 4.139617292700213e-05, + "loss": 0.188, + "step": 1298 + }, + { + "epoch": 1.5309941520467836, + "grad_norm": 0.7772109508514404, + "learning_rate": 4.174698795180723e-05, + "loss": 0.1731, + "step": 1309 + }, + { + "epoch": 1.543859649122807, + "grad_norm": 0.47818320989608765, + "learning_rate": 4.209780297661233e-05, + "loss": 0.2191, + "step": 1320 + }, + { + "epoch": 1.5567251461988305, + "grad_norm": 0.09036953002214432, + "learning_rate": 4.244861800141743e-05, + "loss": 0.146, + "step": 1331 + }, + { + "epoch": 1.5695906432748536, + "grad_norm": 1.416448950767517, + "learning_rate": 4.279943302622254e-05, + "loss": 0.1045, + "step": 1342 + }, + { + "epoch": 1.5824561403508772, + "grad_norm": 0.6509301662445068, + "learning_rate": 4.315024805102764e-05, + "loss": 0.1901, + "step": 1353 + }, + { + "epoch": 1.5953216374269006, + "grad_norm": 0.9354438781738281, + "learning_rate": 4.350106307583275e-05, + "loss": 0.1898, + "step": 1364 + }, + { + "epoch": 1.608187134502924, + "grad_norm": 0.3101876378059387, + "learning_rate": 4.385187810063785e-05, + "loss": 0.0942, + "step": 1375 + }, + { + "epoch": 1.6210526315789475, + "grad_norm": 1.1938691139221191, + "learning_rate": 4.4202693125442954e-05, + "loss": 0.1809, + "step": 1386 + }, + { + "epoch": 1.6339181286549707, + "grad_norm": 0.30765020847320557, + "learning_rate": 4.4553508150248054e-05, + "loss": 0.1083, + "step": 1397 + }, + { + "epoch": 1.6467836257309942, + "grad_norm": 0.24205227196216583, + "learning_rate": 4.4904323175053154e-05, + "loss": 0.1277, + "step": 1408 + }, + { + "epoch": 1.6596491228070176, + "grad_norm": 0.3352263569831848, + "learning_rate": 4.499930693588227e-05, + "loss": 0.1039, + "step": 1419 + }, + { + "epoch": 1.672514619883041, + "grad_norm": 0.07631414383649826, + "learning_rate": 4.49960908016302e-05, + "loss": 0.0933, + "step": 1430 + }, + { + "epoch": 1.6853801169590643, + "grad_norm": 0.7656713128089905, + "learning_rate": 4.4990254602804186e-05, + "loss": 0.11, + "step": 1441 + }, + { + "epoch": 1.6982456140350877, + "grad_norm": 0.9842977523803711, + "learning_rate": 4.4981799189100865e-05, + "loss": 0.2423, + "step": 1452 + }, + { + "epoch": 1.7111111111111112, + "grad_norm": 0.7596272826194763, + "learning_rate": 4.497072579155035e-05, + "loss": 0.1085, + "step": 1463 + }, + { + "epoch": 1.7239766081871344, + "grad_norm": 0.9825586080551147, + "learning_rate": 4.495703602233704e-05, + "loss": 0.1678, + "step": 1474 + }, + { + "epoch": 1.736842105263158, + "grad_norm": 0.6598060131072998, + "learning_rate": 4.494073187456486e-05, + "loss": 0.1799, + "step": 1485 + }, + { + "epoch": 1.7497076023391813, + "grad_norm": 1.0556715726852417, + "learning_rate": 4.492181572196711e-05, + "loss": 0.1811, + "step": 1496 + }, + { + "epoch": 1.752046783625731, + "eval_Qnli-dev_cosine_accuracy": 0.693359375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6543508768081665, + "eval_Qnli-dev_cosine_ap": 0.7124557023921706, + "eval_Qnli-dev_cosine_f1": 0.6733668341708542, + "eval_Qnli-dev_cosine_f1_threshold": 0.5232125520706177, + "eval_Qnli-dev_cosine_precision": 0.556786703601108, + "eval_Qnli-dev_cosine_recall": 0.8516949152542372, + "eval_Qnli-dev_dot_accuracy": 0.666015625, + "eval_Qnli-dev_dot_accuracy_threshold": 256.9267272949219, + "eval_Qnli-dev_dot_ap": 0.6584233742794683, + "eval_Qnli-dev_dot_f1": 0.6719242902208202, + "eval_Qnli-dev_dot_f1_threshold": 186.3544921875, + "eval_Qnli-dev_dot_precision": 0.535175879396985, + "eval_Qnli-dev_dot_recall": 0.902542372881356, + "eval_Qnli-dev_euclidean_accuracy": 0.6953125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 16.447265625, + "eval_Qnli-dev_euclidean_ap": 0.7183967614546444, + "eval_Qnli-dev_euclidean_f1": 0.6785714285714286, + "eval_Qnli-dev_euclidean_f1_threshold": 17.252986907958984, + "eval_Qnli-dev_euclidean_precision": 0.6380597014925373, + "eval_Qnli-dev_euclidean_recall": 0.7245762711864406, + "eval_Qnli-dev_manhattan_accuracy": 0.6953125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 357.7731018066406, + "eval_Qnli-dev_manhattan_ap": 0.7226095638737486, + "eval_Qnli-dev_manhattan_f1": 0.6757281553398058, + "eval_Qnli-dev_manhattan_f1_threshold": 370.83966064453125, + "eval_Qnli-dev_manhattan_precision": 0.6236559139784946, + "eval_Qnli-dev_manhattan_recall": 0.7372881355932204, + "eval_Qnli-dev_max_accuracy": 0.6953125, + "eval_Qnli-dev_max_accuracy_threshold": 357.7731018066406, + "eval_Qnli-dev_max_ap": 0.7226095638737486, + "eval_Qnli-dev_max_f1": 0.6785714285714286, + "eval_Qnli-dev_max_f1_threshold": 370.83966064453125, + "eval_Qnli-dev_max_precision": 0.6380597014925373, + "eval_Qnli-dev_max_recall": 0.902542372881356, + "eval_allNLI-dev_cosine_accuracy": 0.736328125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7933480739593506, + "eval_allNLI-dev_cosine_ap": 0.6338667537371122, + "eval_allNLI-dev_cosine_f1": 0.6428571428571429, + "eval_allNLI-dev_cosine_f1_threshold": 0.6018309593200684, + "eval_allNLI-dev_cosine_precision": 0.5236363636363637, + "eval_allNLI-dev_cosine_recall": 0.8323699421965318, + "eval_allNLI-dev_dot_accuracy": 0.73046875, + "eval_allNLI-dev_dot_accuracy_threshold": 290.1109313964844, + "eval_allNLI-dev_dot_ap": 0.5977502254208901, + "eval_allNLI-dev_dot_f1": 0.6099009900990099, + "eval_allNLI-dev_dot_f1_threshold": 200.0629425048828, + "eval_allNLI-dev_dot_precision": 0.463855421686747, + "eval_allNLI-dev_dot_recall": 0.8901734104046243, + "eval_allNLI-dev_euclidean_accuracy": 0.744140625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 12.976269721984863, + "eval_allNLI-dev_euclidean_ap": 0.6359674505468691, + "eval_allNLI-dev_euclidean_f1": 0.6419213973799127, + "eval_allNLI-dev_euclidean_f1_threshold": 17.44091033935547, + "eval_allNLI-dev_euclidean_precision": 0.5157894736842106, + "eval_allNLI-dev_euclidean_recall": 0.8497109826589595, + "eval_allNLI-dev_manhattan_accuracy": 0.7421875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 273.52783203125, + "eval_allNLI-dev_manhattan_ap": 0.6336272156038607, + "eval_allNLI-dev_manhattan_f1": 0.6501128668171559, + "eval_allNLI-dev_manhattan_f1_threshold": 362.5467529296875, + "eval_allNLI-dev_manhattan_precision": 0.5333333333333333, + "eval_allNLI-dev_manhattan_recall": 0.8323699421965318, + "eval_allNLI-dev_max_accuracy": 0.744140625, + "eval_allNLI-dev_max_accuracy_threshold": 290.1109313964844, + "eval_allNLI-dev_max_ap": 0.6359674505468691, + "eval_allNLI-dev_max_f1": 0.6501128668171559, + "eval_allNLI-dev_max_f1_threshold": 362.5467529296875, + "eval_allNLI-dev_max_precision": 0.5333333333333333, + "eval_allNLI-dev_max_recall": 0.8901734104046243, + "eval_sequential_score": 0.7226095638737486, + "eval_sts-test_pearson_cosine": 0.884780047363058, + "eval_sts-test_pearson_dot": 0.8686782237189175, + "eval_sts-test_pearson_euclidean": 0.904848497388399, + "eval_sts-test_pearson_manhattan": 0.9057450598837923, + "eval_sts-test_pearson_max": 0.9057450598837923, + "eval_sts-test_spearman_cosine": 0.9069954294787459, + "eval_sts-test_spearman_dot": 0.8651062350096334, + "eval_sts-test_spearman_euclidean": 0.9006308603323997, + "eval_sts-test_spearman_manhattan": 0.9018185140584661, + "eval_sts-test_spearman_max": 0.9069954294787459, + "eval_vitaminc-pairs_loss": 2.6750705242156982, + "eval_vitaminc-pairs_runtime": 3.8295, + "eval_vitaminc-pairs_samples_per_second": 33.425, + "eval_vitaminc-pairs_steps_per_second": 0.261, + "step": 1498 + }, + { + "epoch": 1.752046783625731, + "eval_negation-triplets_loss": 1.0360219478607178, + "eval_negation-triplets_runtime": 0.6077, + "eval_negation-triplets_samples_per_second": 210.626, + "eval_negation-triplets_steps_per_second": 1.646, + "step": 1498 + }, + { + "epoch": 1.752046783625731, + "eval_scitail-pairs-pos_loss": 0.05649742856621742, + "eval_scitail-pairs-pos_runtime": 0.8925, + "eval_scitail-pairs-pos_samples_per_second": 143.418, + "eval_scitail-pairs-pos_steps_per_second": 1.12, + "step": 1498 + }, + { + "epoch": 1.752046783625731, + "eval_scitail-pairs-qa_loss": 4.5177861807133013e-07, + "eval_scitail-pairs-qa_runtime": 0.5887, + "eval_scitail-pairs-qa_samples_per_second": 217.414, + "eval_scitail-pairs-qa_steps_per_second": 1.699, + "step": 1498 + }, + { + "epoch": 1.752046783625731, + "eval_xsum-pairs_loss": 0.006031599827110767, + "eval_xsum-pairs_runtime": 2.7875, + "eval_xsum-pairs_samples_per_second": 45.92, + "eval_xsum-pairs_steps_per_second": 0.359, + "step": 1498 + }, + { + "epoch": 1.752046783625731, + "eval_sciq_pairs_loss": 0.0044244336895644665, + "eval_sciq_pairs_runtime": 4.1478, + "eval_sciq_pairs_samples_per_second": 30.859, + "eval_sciq_pairs_steps_per_second": 0.241, + "step": 1498 + }, + { + "epoch": 1.752046783625731, + "eval_qasc_pairs_loss": 0.05687851831316948, + "eval_qasc_pairs_runtime": 0.6029, + "eval_qasc_pairs_samples_per_second": 212.298, + "eval_qasc_pairs_steps_per_second": 1.659, + "step": 1498 + }, + { + "epoch": 1.752046783625731, + "eval_openbookqa_pairs_loss": 1.2448060512542725, + "eval_openbookqa_pairs_runtime": 0.5722, + "eval_openbookqa_pairs_samples_per_second": 223.708, + "eval_openbookqa_pairs_steps_per_second": 1.748, + "step": 1498 + }, + { + "epoch": 1.752046783625731, + "eval_msmarco_pairs_loss": 0.14275667071342468, + "eval_msmarco_pairs_runtime": 1.3302, + "eval_msmarco_pairs_samples_per_second": 96.223, + "eval_msmarco_pairs_steps_per_second": 0.752, + "step": 1498 + }, + { + "epoch": 1.752046783625731, + "eval_nq_pairs_loss": 0.04014577344059944, + "eval_nq_pairs_runtime": 2.9351, + "eval_nq_pairs_samples_per_second": 43.61, + "eval_nq_pairs_steps_per_second": 0.341, + "step": 1498 + }, + { + "epoch": 1.752046783625731, + "eval_trivia_pairs_loss": 0.19849494099617004, + "eval_trivia_pairs_runtime": 3.1918, + "eval_trivia_pairs_samples_per_second": 40.103, + "eval_trivia_pairs_steps_per_second": 0.313, + "step": 1498 + }, + { + "epoch": 1.752046783625731, + "eval_gooaq_pairs_loss": 0.10808175802230835, + "eval_gooaq_pairs_runtime": 0.8763, + "eval_gooaq_pairs_samples_per_second": 146.065, + "eval_gooaq_pairs_steps_per_second": 1.141, + "step": 1498 + }, + { + "epoch": 1.752046783625731, + "eval_paws-pos_loss": 0.045870762318372726, + "eval_paws-pos_runtime": 0.6763, + "eval_paws-pos_samples_per_second": 189.269, + "eval_paws-pos_steps_per_second": 1.479, + "step": 1498 + }, + { + "epoch": 1.752046783625731, + "eval_global_dataset_loss": 0.20272643864154816, + "eval_global_dataset_runtime": 24.2206, + "eval_global_dataset_samples_per_second": 27.373, + "eval_global_dataset_steps_per_second": 0.248, + "step": 1498 + }, + { + "epoch": 1.7625730994152047, + "grad_norm": 1.1749067306518555, + "learning_rate": 4.490029031856087e-05, + "loss": 0.1746, + "step": 1507 + }, + { + "epoch": 1.775438596491228, + "grad_norm": 0.8320943713188171, + "learning_rate": 4.4876158798246026e-05, + "loss": 0.1603, + "step": 1518 + }, + { + "epoch": 1.7883040935672514, + "grad_norm": 1.3382437229156494, + "learning_rate": 4.484942467434899e-05, + "loss": 0.1784, + "step": 1529 + }, + { + "epoch": 1.801169590643275, + "grad_norm": 0.6635130643844604, + "learning_rate": 4.4820091839111255e-05, + "loss": 0.1041, + "step": 1540 + }, + { + "epoch": 1.8140350877192981, + "grad_norm": 0.534389853477478, + "learning_rate": 4.478816456312265e-05, + "loss": 0.1067, + "step": 1551 + }, + { + "epoch": 1.8269005847953217, + "grad_norm": 0.6235479712486267, + "learning_rate": 4.475364749469961e-05, + "loss": 0.1293, + "step": 1562 + }, + { + "epoch": 1.839766081871345, + "grad_norm": 1.2809871435165405, + "learning_rate": 4.471654565920844e-05, + "loss": 0.18, + "step": 1573 + }, + { + "epoch": 1.8526315789473684, + "grad_norm": 0.2697462737560272, + "learning_rate": 4.467686445833364e-05, + "loss": 0.1481, + "step": 1584 + }, + { + "epoch": 1.8654970760233918, + "grad_norm": 0.4068855345249176, + "learning_rate": 4.4634609669291485e-05, + "loss": 0.1573, + "step": 1595 + }, + { + "epoch": 1.8783625730994151, + "grad_norm": 0.42257580161094666, + "learning_rate": 4.458978744398891e-05, + "loss": 0.1434, + "step": 1606 + }, + { + "epoch": 1.8912280701754387, + "grad_norm": 0.928801417350769, + "learning_rate": 4.454240430812786e-05, + "loss": 0.0975, + "step": 1617 + }, + { + "epoch": 1.9040935672514618, + "grad_norm": 0.5400815606117249, + "learning_rate": 4.449246716025518e-05, + "loss": 0.1133, + "step": 1628 + }, + { + "epoch": 1.9169590643274854, + "grad_norm": 2.2877585887908936, + "learning_rate": 4.443998327075831e-05, + "loss": 0.1661, + "step": 1639 + }, + { + "epoch": 1.9298245614035088, + "grad_norm": 0.12138649821281433, + "learning_rate": 4.438496028080669e-05, + "loss": 0.1518, + "step": 1650 + }, + { + "epoch": 1.9426900584795321, + "grad_norm": 0.9370628595352173, + "learning_rate": 4.432740620123937e-05, + "loss": 0.1365, + "step": 1661 + }, + { + "epoch": 1.9555555555555557, + "grad_norm": 0.5390945076942444, + "learning_rate": 4.426732941139866e-05, + "loss": 0.1226, + "step": 1672 + }, + { + "epoch": 1.9684210526315788, + "grad_norm": 0.5244420766830444, + "learning_rate": 4.4204738657910166e-05, + "loss": 0.1347, + "step": 1683 + }, + { + "epoch": 1.9812865497076024, + "grad_norm": 0.09539759159088135, + "learning_rate": 4.413964305340939e-05, + "loss": 0.1121, + "step": 1694 + }, + { + "epoch": 1.9941520467836256, + "grad_norm": 0.35612645745277405, + "learning_rate": 4.407205207521499e-05, + "loss": 0.1224, + "step": 1705 + }, + { + "epoch": 2.00233918128655, + "eval_Qnli-dev_cosine_accuracy": 0.69140625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.656867504119873, + "eval_Qnli-dev_cosine_ap": 0.712494851178384, + "eval_Qnli-dev_cosine_f1": 0.6740506329113923, + "eval_Qnli-dev_cosine_f1_threshold": 0.4821203351020813, + "eval_Qnli-dev_cosine_precision": 0.5378787878787878, + "eval_Qnli-dev_cosine_recall": 0.902542372881356, + "eval_Qnli-dev_dot_accuracy": 0.66015625, + "eval_Qnli-dev_dot_accuracy_threshold": 297.9504699707031, + "eval_Qnli-dev_dot_ap": 0.6569442778223444, + "eval_Qnli-dev_dot_f1": 0.6739469578783152, + "eval_Qnli-dev_dot_f1_threshold": 180.67715454101562, + "eval_Qnli-dev_dot_precision": 0.5333333333333333, + "eval_Qnli-dev_dot_recall": 0.9152542372881356, + "eval_Qnli-dev_euclidean_accuracy": 0.69921875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 16.230770111083984, + "eval_Qnli-dev_euclidean_ap": 0.7199985561066051, + "eval_Qnli-dev_euclidean_f1": 0.6795366795366796, + "eval_Qnli-dev_euclidean_f1_threshold": 17.748735427856445, + "eval_Qnli-dev_euclidean_precision": 0.624113475177305, + "eval_Qnli-dev_euclidean_recall": 0.7457627118644068, + "eval_Qnli-dev_manhattan_accuracy": 0.69140625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 322.521728515625, + "eval_Qnli-dev_manhattan_ap": 0.7224632454419206, + "eval_Qnli-dev_manhattan_f1": 0.6817325800376648, + "eval_Qnli-dev_manhattan_f1_threshold": 380.47283935546875, + "eval_Qnli-dev_manhattan_precision": 0.6135593220338983, + "eval_Qnli-dev_manhattan_recall": 0.7669491525423728, + "eval_Qnli-dev_max_accuracy": 0.69921875, + "eval_Qnli-dev_max_accuracy_threshold": 322.521728515625, + "eval_Qnli-dev_max_ap": 0.7224632454419206, + "eval_Qnli-dev_max_f1": 0.6817325800376648, + "eval_Qnli-dev_max_f1_threshold": 380.47283935546875, + "eval_Qnli-dev_max_precision": 0.624113475177305, + "eval_Qnli-dev_max_recall": 0.9152542372881356, + "eval_allNLI-dev_cosine_accuracy": 0.7421875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7574848532676697, + "eval_allNLI-dev_cosine_ap": 0.6348133852635633, + "eval_allNLI-dev_cosine_f1": 0.6420824295010846, + "eval_allNLI-dev_cosine_f1_threshold": 0.5789805054664612, + "eval_allNLI-dev_cosine_precision": 0.5138888888888888, + "eval_allNLI-dev_cosine_recall": 0.8554913294797688, + "eval_allNLI-dev_dot_accuracy": 0.732421875, + "eval_allNLI-dev_dot_accuracy_threshold": 285.750244140625, + "eval_allNLI-dev_dot_ap": 0.59931731879636, + "eval_allNLI-dev_dot_f1": 0.6150121065375304, + "eval_allNLI-dev_dot_f1_threshold": 243.38694763183594, + "eval_allNLI-dev_dot_precision": 0.5291666666666667, + "eval_allNLI-dev_dot_recall": 0.7341040462427746, + "eval_allNLI-dev_euclidean_accuracy": 0.744140625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 13.472793579101562, + "eval_allNLI-dev_euclidean_ap": 0.6356913317153288, + "eval_allNLI-dev_euclidean_f1": 0.6475770925110133, + "eval_allNLI-dev_euclidean_f1_threshold": 17.89451789855957, + "eval_allNLI-dev_euclidean_precision": 0.5231316725978647, + "eval_allNLI-dev_euclidean_recall": 0.8497109826589595, + "eval_allNLI-dev_manhattan_accuracy": 0.73828125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 275.39105224609375, + "eval_allNLI-dev_manhattan_ap": 0.6342177592020837, + "eval_allNLI-dev_manhattan_f1": 0.6511627906976745, + "eval_allNLI-dev_manhattan_f1_threshold": 347.1697998046875, + "eval_allNLI-dev_manhattan_precision": 0.5887850467289719, + "eval_allNLI-dev_manhattan_recall": 0.7283236994219653, + "eval_allNLI-dev_max_accuracy": 0.744140625, + "eval_allNLI-dev_max_accuracy_threshold": 285.750244140625, + "eval_allNLI-dev_max_ap": 0.6356913317153288, + "eval_allNLI-dev_max_f1": 0.6511627906976745, + "eval_allNLI-dev_max_f1_threshold": 347.1697998046875, + "eval_allNLI-dev_max_precision": 0.5887850467289719, + "eval_allNLI-dev_max_recall": 0.8554913294797688, + "eval_sequential_score": 0.7224632454419206, + "eval_sts-test_pearson_cosine": 0.885504470943527, + "eval_sts-test_pearson_dot": 0.8680100258122485, + "eval_sts-test_pearson_euclidean": 0.9053088527197759, + "eval_sts-test_pearson_manhattan": 0.9062637284673284, + "eval_sts-test_pearson_max": 0.9062637284673284, + "eval_sts-test_spearman_cosine": 0.907385436721772, + "eval_sts-test_spearman_dot": 0.8633328767744765, + "eval_sts-test_spearman_euclidean": 0.9008774845339277, + "eval_sts-test_spearman_manhattan": 0.9025189160505028, + "eval_sts-test_spearman_max": 0.907385436721772, + "eval_vitaminc-pairs_loss": 2.711371898651123, + "eval_vitaminc-pairs_runtime": 3.8178, + "eval_vitaminc-pairs_samples_per_second": 33.527, + "eval_vitaminc-pairs_steps_per_second": 0.262, + "step": 1712 + }, + { + "epoch": 2.00233918128655, + "eval_negation-triplets_loss": 1.0452457666397095, + "eval_negation-triplets_runtime": 0.606, + "eval_negation-triplets_samples_per_second": 211.229, + "eval_negation-triplets_steps_per_second": 1.65, + "step": 1712 + }, + { + "epoch": 2.00233918128655, + "eval_scitail-pairs-pos_loss": 0.05906897410750389, + "eval_scitail-pairs-pos_runtime": 0.8912, + "eval_scitail-pairs-pos_samples_per_second": 143.627, + "eval_scitail-pairs-pos_steps_per_second": 1.122, + "step": 1712 + }, + { + "epoch": 2.00233918128655, + "eval_scitail-pairs-qa_loss": 3.425042223170749e-07, + "eval_scitail-pairs-qa_runtime": 0.586, + "eval_scitail-pairs-qa_samples_per_second": 218.414, + "eval_scitail-pairs-qa_steps_per_second": 1.706, + "step": 1712 + }, + { + "epoch": 2.00233918128655, + "eval_xsum-pairs_loss": 0.006712059490382671, + "eval_xsum-pairs_runtime": 2.7794, + "eval_xsum-pairs_samples_per_second": 46.053, + "eval_xsum-pairs_steps_per_second": 0.36, + "step": 1712 + }, + { + "epoch": 2.00233918128655, + "eval_sciq_pairs_loss": 0.0031704979483038187, + "eval_sciq_pairs_runtime": 4.1826, + "eval_sciq_pairs_samples_per_second": 30.603, + "eval_sciq_pairs_steps_per_second": 0.239, + "step": 1712 + }, + { + "epoch": 2.00233918128655, + "eval_qasc_pairs_loss": 0.04258028417825699, + "eval_qasc_pairs_runtime": 0.6039, + "eval_qasc_pairs_samples_per_second": 211.938, + "eval_qasc_pairs_steps_per_second": 1.656, + "step": 1712 + }, + { + "epoch": 2.00233918128655, + "eval_openbookqa_pairs_loss": 1.329345941543579, + "eval_openbookqa_pairs_runtime": 0.5602, + "eval_openbookqa_pairs_samples_per_second": 228.507, + "eval_openbookqa_pairs_steps_per_second": 1.785, + "step": 1712 + }, + { + "epoch": 2.00233918128655, + "eval_msmarco_pairs_loss": 0.15084555745124817, + "eval_msmarco_pairs_runtime": 1.327, + "eval_msmarco_pairs_samples_per_second": 96.46, + "eval_msmarco_pairs_steps_per_second": 0.754, + "step": 1712 + }, + { + "epoch": 2.00233918128655, + "eval_nq_pairs_loss": 0.042938798666000366, + "eval_nq_pairs_runtime": 2.9551, + "eval_nq_pairs_samples_per_second": 43.314, + "eval_nq_pairs_steps_per_second": 0.338, + "step": 1712 + }, + { + "epoch": 2.00233918128655, + "eval_trivia_pairs_loss": 0.20491687953472137, + "eval_trivia_pairs_runtime": 3.219, + "eval_trivia_pairs_samples_per_second": 39.764, + "eval_trivia_pairs_steps_per_second": 0.311, + "step": 1712 + }, + { + "epoch": 2.00233918128655, + "eval_gooaq_pairs_loss": 0.10198851674795151, + "eval_gooaq_pairs_runtime": 0.8632, + "eval_gooaq_pairs_samples_per_second": 148.277, + "eval_gooaq_pairs_steps_per_second": 1.158, + "step": 1712 + }, + { + "epoch": 2.00233918128655, + "eval_paws-pos_loss": 0.04579318314790726, + "eval_paws-pos_runtime": 0.6652, + "eval_paws-pos_samples_per_second": 192.431, + "eval_paws-pos_steps_per_second": 1.503, + "step": 1712 + }, + { + "epoch": 2.00233918128655, + "eval_global_dataset_loss": 0.18868622183799744, + "eval_global_dataset_runtime": 24.2368, + "eval_global_dataset_samples_per_second": 27.355, + "eval_global_dataset_steps_per_second": 0.248, + "step": 1712 + }, + { + "epoch": 2.007017543859649, + "grad_norm": 0.6866632699966431, + "learning_rate": 4.400197556394901e-05, + "loss": 0.1359, + "step": 1716 + }, + { + "epoch": 2.0198830409356727, + "grad_norm": 1.721706509590149, + "learning_rate": 4.392942372210411e-05, + "loss": 0.1889, + "step": 1727 + }, + { + "epoch": 2.032748538011696, + "grad_norm": 1.0161302089691162, + "learning_rate": 4.3854407112558235e-05, + "loss": 0.0972, + "step": 1738 + }, + { + "epoch": 2.0456140350877194, + "grad_norm": 0.4521200358867645, + "learning_rate": 4.377693665703674e-05, + "loss": 0.1209, + "step": 1749 + }, + { + "epoch": 2.0584795321637426, + "grad_norm": 0.4354567229747772, + "learning_rate": 4.369702363452229e-05, + "loss": 0.1707, + "step": 1760 + }, + { + "epoch": 2.071345029239766, + "grad_norm": 0.5484113097190857, + "learning_rate": 4.3614679679612715e-05, + "loss": 0.1405, + "step": 1771 + }, + { + "epoch": 2.0842105263157893, + "grad_norm": 0.8810815811157227, + "learning_rate": 4.352991678082717e-05, + "loss": 0.1099, + "step": 1782 + }, + { + "epoch": 2.097076023391813, + "grad_norm": 0.322817862033844, + "learning_rate": 4.344274727886069e-05, + "loss": 0.1611, + "step": 1793 + }, + { + "epoch": 2.1099415204678365, + "grad_norm": 0.8280489444732666, + "learning_rate": 4.3353183864787466e-05, + "loss": 0.1464, + "step": 1804 + }, + { + "epoch": 2.1228070175438596, + "grad_norm": 1.0790857076644897, + "learning_rate": 4.326123957821325e-05, + "loss": 0.1525, + "step": 1815 + }, + { + "epoch": 2.135672514619883, + "grad_norm": 0.5455640554428101, + "learning_rate": 4.316692780537676e-05, + "loss": 0.1947, + "step": 1826 + }, + { + "epoch": 2.1485380116959063, + "grad_norm": 0.2670153081417084, + "learning_rate": 4.307026227720089e-05, + "loss": 0.1191, + "step": 1837 + }, + { + "epoch": 2.16140350877193, + "grad_norm": 1.0682767629623413, + "learning_rate": 4.2971257067293544e-05, + "loss": 0.2513, + "step": 1848 + }, + { + "epoch": 2.174269005847953, + "grad_norm": 0.35559388995170593, + "learning_rate": 4.286992658989868e-05, + "loss": 0.1819, + "step": 1859 + }, + { + "epoch": 2.1871345029239766, + "grad_norm": 1.2426172494888306, + "learning_rate": 4.276628559779772e-05, + "loss": 0.1874, + "step": 1870 + }, + { + "epoch": 2.2, + "grad_norm": 1.0591782331466675, + "learning_rate": 4.2660349180161695e-05, + "loss": 0.1416, + "step": 1881 + }, + { + "epoch": 2.2128654970760233, + "grad_norm": 0.22729463875293732, + "learning_rate": 4.2552132760354375e-05, + "loss": 0.1146, + "step": 1892 + }, + { + "epoch": 2.225730994152047, + "grad_norm": 0.36285898089408875, + "learning_rate": 4.244165209368678e-05, + "loss": 0.2032, + "step": 1903 + }, + { + "epoch": 2.23859649122807, + "grad_norm": 2.697350263595581, + "learning_rate": 4.2328923265123376e-05, + "loss": 0.1711, + "step": 1914 + }, + { + "epoch": 2.2514619883040936, + "grad_norm": 1.9733246564865112, + "learning_rate": 4.221396268694021e-05, + "loss": 0.1282, + "step": 1925 + }, + { + "epoch": 2.2526315789473683, + "eval_Qnli-dev_cosine_accuracy": 0.69140625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6526055335998535, + "eval_Qnli-dev_cosine_ap": 0.7147593742135198, + "eval_Qnli-dev_cosine_f1": 0.6751188589540412, + "eval_Qnli-dev_cosine_f1_threshold": 0.4857839047908783, + "eval_Qnli-dev_cosine_precision": 0.5392405063291139, + "eval_Qnli-dev_cosine_recall": 0.902542372881356, + "eval_Qnli-dev_dot_accuracy": 0.671875, + "eval_Qnli-dev_dot_accuracy_threshold": 245.6788330078125, + "eval_Qnli-dev_dot_ap": 0.6587738399719532, + "eval_Qnli-dev_dot_f1": 0.6770670826833073, + "eval_Qnli-dev_dot_f1_threshold": 176.555419921875, + "eval_Qnli-dev_dot_precision": 0.5358024691358024, + "eval_Qnli-dev_dot_recall": 0.9194915254237288, + "eval_Qnli-dev_euclidean_accuracy": 0.69921875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 16.296737670898438, + "eval_Qnli-dev_euclidean_ap": 0.7231348856348505, + "eval_Qnli-dev_euclidean_f1": 0.6788321167883212, + "eval_Qnli-dev_euclidean_f1_threshold": 17.902896881103516, + "eval_Qnli-dev_euclidean_precision": 0.5961538461538461, + "eval_Qnli-dev_euclidean_recall": 0.788135593220339, + "eval_Qnli-dev_manhattan_accuracy": 0.69921875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 342.3966979980469, + "eval_Qnli-dev_manhattan_ap": 0.7240429427054647, + "eval_Qnli-dev_manhattan_f1": 0.6791744840525328, + "eval_Qnli-dev_manhattan_f1_threshold": 372.05224609375, + "eval_Qnli-dev_manhattan_precision": 0.6094276094276094, + "eval_Qnli-dev_manhattan_recall": 0.7669491525423728, + "eval_Qnli-dev_max_accuracy": 0.69921875, + "eval_Qnli-dev_max_accuracy_threshold": 342.3966979980469, + "eval_Qnli-dev_max_ap": 0.7240429427054647, + "eval_Qnli-dev_max_f1": 0.6791744840525328, + "eval_Qnli-dev_max_f1_threshold": 372.05224609375, + "eval_Qnli-dev_max_precision": 0.6094276094276094, + "eval_Qnli-dev_max_recall": 0.9194915254237288, + "eval_allNLI-dev_cosine_accuracy": 0.740234375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7803784608840942, + "eval_allNLI-dev_cosine_ap": 0.6307077516423368, + "eval_allNLI-dev_cosine_f1": 0.6382022471910113, + "eval_allNLI-dev_cosine_f1_threshold": 0.609204888343811, + "eval_allNLI-dev_cosine_precision": 0.5220588235294118, + "eval_allNLI-dev_cosine_recall": 0.8208092485549133, + "eval_allNLI-dev_dot_accuracy": 0.73046875, + "eval_allNLI-dev_dot_accuracy_threshold": 282.7746887207031, + "eval_allNLI-dev_dot_ap": 0.5937116347608066, + "eval_allNLI-dev_dot_f1": 0.6070038910505836, + "eval_allNLI-dev_dot_f1_threshold": 197.9422149658203, + "eval_allNLI-dev_dot_precision": 0.4574780058651026, + "eval_allNLI-dev_dot_recall": 0.9017341040462428, + "eval_allNLI-dev_euclidean_accuracy": 0.744140625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 13.138710021972656, + "eval_allNLI-dev_euclidean_ap": 0.635229512907449, + "eval_allNLI-dev_euclidean_f1": 0.6458797327394209, + "eval_allNLI-dev_euclidean_f1_threshold": 17.1611328125, + "eval_allNLI-dev_euclidean_precision": 0.5253623188405797, + "eval_allNLI-dev_euclidean_recall": 0.838150289017341, + "eval_allNLI-dev_manhattan_accuracy": 0.7421875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 278.34954833984375, + "eval_allNLI-dev_manhattan_ap": 0.6345395417826952, + "eval_allNLI-dev_manhattan_f1": 0.6532438478747203, + "eval_allNLI-dev_manhattan_f1_threshold": 361.567626953125, + "eval_allNLI-dev_manhattan_precision": 0.5328467153284672, + "eval_allNLI-dev_manhattan_recall": 0.8439306358381503, + "eval_allNLI-dev_max_accuracy": 0.744140625, + "eval_allNLI-dev_max_accuracy_threshold": 282.7746887207031, + "eval_allNLI-dev_max_ap": 0.635229512907449, + "eval_allNLI-dev_max_f1": 0.6532438478747203, + "eval_allNLI-dev_max_f1_threshold": 361.567626953125, + "eval_allNLI-dev_max_precision": 0.5328467153284672, + "eval_allNLI-dev_max_recall": 0.9017341040462428, + "eval_sequential_score": 0.7240429427054647, + "eval_sts-test_pearson_cosine": 0.8838830663702884, + "eval_sts-test_pearson_dot": 0.8605683837270216, + "eval_sts-test_pearson_euclidean": 0.9044604172987989, + "eval_sts-test_pearson_manhattan": 0.9050995663550978, + "eval_sts-test_pearson_max": 0.9050995663550978, + "eval_sts-test_spearman_cosine": 0.9069073589963277, + "eval_sts-test_spearman_dot": 0.8518454964828726, + "eval_sts-test_spearman_euclidean": 0.9001444141210752, + "eval_sts-test_spearman_manhattan": 0.9010118725718861, + "eval_sts-test_spearman_max": 0.9069073589963277, + "eval_vitaminc-pairs_loss": 2.658858299255371, + "eval_vitaminc-pairs_runtime": 3.7881, + "eval_vitaminc-pairs_samples_per_second": 33.79, + "eval_vitaminc-pairs_steps_per_second": 0.264, + "step": 1926 + }, + { + "epoch": 2.2526315789473683, + "eval_negation-triplets_loss": 1.051798939704895, + "eval_negation-triplets_runtime": 0.5864, + "eval_negation-triplets_samples_per_second": 218.293, + "eval_negation-triplets_steps_per_second": 1.705, + "step": 1926 + }, + { + "epoch": 2.2526315789473683, + "eval_scitail-pairs-pos_loss": 0.058272454887628555, + "eval_scitail-pairs-pos_runtime": 0.8071, + "eval_scitail-pairs-pos_samples_per_second": 158.587, + "eval_scitail-pairs-pos_steps_per_second": 1.239, + "step": 1926 + }, + { + "epoch": 2.2526315789473683, + "eval_scitail-pairs-qa_loss": 5.024415372645308e-07, + "eval_scitail-pairs-qa_runtime": 0.5659, + "eval_scitail-pairs-qa_samples_per_second": 226.172, + "eval_scitail-pairs-qa_steps_per_second": 1.767, + "step": 1926 + }, + { + "epoch": 2.2526315789473683, + "eval_xsum-pairs_loss": 0.0069998325780034065, + "eval_xsum-pairs_runtime": 2.7666, + "eval_xsum-pairs_samples_per_second": 46.267, + "eval_xsum-pairs_steps_per_second": 0.361, + "step": 1926 + }, + { + "epoch": 2.2526315789473683, + "eval_sciq_pairs_loss": 0.003587064566090703, + "eval_sciq_pairs_runtime": 4.0629, + "eval_sciq_pairs_samples_per_second": 31.505, + "eval_sciq_pairs_steps_per_second": 0.246, + "step": 1926 + }, + { + "epoch": 2.2526315789473683, + "eval_qasc_pairs_loss": 0.05682336166501045, + "eval_qasc_pairs_runtime": 0.5866, + "eval_qasc_pairs_samples_per_second": 218.204, + "eval_qasc_pairs_steps_per_second": 1.705, + "step": 1926 + }, + { + "epoch": 2.2526315789473683, + "eval_openbookqa_pairs_loss": 1.2767432928085327, + "eval_openbookqa_pairs_runtime": 0.5523, + "eval_openbookqa_pairs_samples_per_second": 231.759, + "eval_openbookqa_pairs_steps_per_second": 1.811, + "step": 1926 + }, + { + "epoch": 2.2526315789473683, + "eval_msmarco_pairs_loss": 0.15254586935043335, + "eval_msmarco_pairs_runtime": 1.3181, + "eval_msmarco_pairs_samples_per_second": 97.106, + "eval_msmarco_pairs_steps_per_second": 0.759, + "step": 1926 + }, + { + "epoch": 2.2526315789473683, + "eval_nq_pairs_loss": 0.04068170115351677, + "eval_nq_pairs_runtime": 2.9018, + "eval_nq_pairs_samples_per_second": 44.11, + "eval_nq_pairs_steps_per_second": 0.345, + "step": 1926 + }, + { + "epoch": 2.2526315789473683, + "eval_trivia_pairs_loss": 0.20555011928081512, + "eval_trivia_pairs_runtime": 3.1824, + "eval_trivia_pairs_samples_per_second": 40.222, + "eval_trivia_pairs_steps_per_second": 0.314, + "step": 1926 + }, + { + "epoch": 2.2526315789473683, + "eval_gooaq_pairs_loss": 0.11310738325119019, + "eval_gooaq_pairs_runtime": 0.8496, + "eval_gooaq_pairs_samples_per_second": 150.654, + "eval_gooaq_pairs_steps_per_second": 1.177, + "step": 1926 + }, + { + "epoch": 2.2526315789473683, + "eval_paws-pos_loss": 0.046021316200494766, + "eval_paws-pos_runtime": 0.6545, + "eval_paws-pos_samples_per_second": 195.571, + "eval_paws-pos_steps_per_second": 1.528, + "step": 1926 + }, + { + "epoch": 2.2526315789473683, + "eval_global_dataset_loss": 0.2109522521495819, + "eval_global_dataset_runtime": 24.1239, + "eval_global_dataset_samples_per_second": 27.483, + "eval_global_dataset_steps_per_second": 0.249, + "step": 1926 + }, + { + "epoch": 2.264327485380117, + "grad_norm": 0.06653619557619095, + "learning_rate": 4.2096787096335474e-05, + "loss": 0.1317, + "step": 1936 + }, + { + "epoch": 2.2771929824561403, + "grad_norm": 0.3638446033000946, + "learning_rate": 4.197741355299269e-05, + "loss": 0.1155, + "step": 1947 + }, + { + "epoch": 2.290058479532164, + "grad_norm": 0.36138415336608887, + "learning_rate": 4.185585943659701e-05, + "loss": 0.2087, + "step": 1958 + }, + { + "epoch": 2.302923976608187, + "grad_norm": 0.39415091276168823, + "learning_rate": 4.1732142444304866e-05, + "loss": 0.1564, + "step": 1969 + }, + { + "epoch": 2.3157894736842106, + "grad_norm": 0.4760674834251404, + "learning_rate": 4.160628058816745e-05, + "loss": 0.1427, + "step": 1980 + }, + { + "epoch": 2.3286549707602338, + "grad_norm": 2.399590253829956, + "learning_rate": 4.147829219250831e-05, + "loss": 0.142, + "step": 1991 + }, + { + "epoch": 2.3415204678362573, + "grad_norm": 0.19875767827033997, + "learning_rate": 4.134819589125551e-05, + "loss": 0.2259, + "step": 2002 + }, + { + "epoch": 2.3543859649122805, + "grad_norm": 0.49662378430366516, + "learning_rate": 4.121601062522869e-05, + "loss": 0.1366, + "step": 2013 + }, + { + "epoch": 2.367251461988304, + "grad_norm": 0.9105015397071838, + "learning_rate": 4.1081755639381465e-05, + "loss": 0.2698, + "step": 2024 + }, + { + "epoch": 2.3801169590643276, + "grad_norm": 0.2802727520465851, + "learning_rate": 4.094545047999952e-05, + "loss": 0.1317, + "step": 2035 + }, + { + "epoch": 2.3929824561403508, + "grad_norm": 0.28353074193000793, + "learning_rate": 4.08071149918549e-05, + "loss": 0.1137, + "step": 2046 + }, + { + "epoch": 2.4058479532163743, + "grad_norm": 0.8746175765991211, + "learning_rate": 4.066676931531674e-05, + "loss": 0.1261, + "step": 2057 + }, + { + "epoch": 2.4187134502923975, + "grad_norm": 0.7557067275047302, + "learning_rate": 4.0524433883419034e-05, + "loss": 0.1068, + "step": 2068 + }, + { + "epoch": 2.431578947368421, + "grad_norm": 1.8618159294128418, + "learning_rate": 4.038012941888579e-05, + "loss": 0.1546, + "step": 2079 + }, + { + "epoch": 2.4444444444444446, + "grad_norm": 0.5580687522888184, + "learning_rate": 4.023387693111395e-05, + "loss": 0.1739, + "step": 2090 + }, + { + "epoch": 2.4573099415204678, + "grad_norm": 1.2131623029708862, + "learning_rate": 4.008569771311469e-05, + "loss": 0.1445, + "step": 2101 + }, + { + "epoch": 2.4701754385964914, + "grad_norm": 1.3317950963974, + "learning_rate": 3.993561333841325e-05, + "loss": 0.1521, + "step": 2112 + }, + { + "epoch": 2.4830409356725145, + "grad_norm": 1.0864313840866089, + "learning_rate": 3.978364565790814e-05, + "loss": 0.1299, + "step": 2123 + }, + { + "epoch": 2.495906432748538, + "grad_norm": 1.382237434387207, + "learning_rate": 3.962981679668976e-05, + "loss": 0.1205, + "step": 2134 + }, + { + "epoch": 2.502923976608187, + "eval_Qnli-dev_cosine_accuracy": 0.6953125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6459293961524963, + "eval_Qnli-dev_cosine_ap": 0.7130285997310533, + "eval_Qnli-dev_cosine_f1": 0.6733668341708542, + "eval_Qnli-dev_cosine_f1_threshold": 0.5255661606788635, + "eval_Qnli-dev_cosine_precision": 0.556786703601108, + "eval_Qnli-dev_cosine_recall": 0.8516949152542372, + "eval_Qnli-dev_dot_accuracy": 0.666015625, + "eval_Qnli-dev_dot_accuracy_threshold": 248.84127807617188, + "eval_Qnli-dev_dot_ap": 0.6544554872798674, + "eval_Qnli-dev_dot_f1": 0.6728682170542635, + "eval_Qnli-dev_dot_f1_threshold": 177.0298614501953, + "eval_Qnli-dev_dot_precision": 0.530562347188264, + "eval_Qnli-dev_dot_recall": 0.9194915254237288, + "eval_Qnli-dev_euclidean_accuracy": 0.6953125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 16.394847869873047, + "eval_Qnli-dev_euclidean_ap": 0.7197613869313146, + "eval_Qnli-dev_euclidean_f1": 0.6799999999999999, + "eval_Qnli-dev_euclidean_f1_threshold": 17.270109176635742, + "eval_Qnli-dev_euclidean_precision": 0.6439393939393939, + "eval_Qnli-dev_euclidean_recall": 0.7203389830508474, + "eval_Qnli-dev_manhattan_accuracy": 0.6953125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 346.01763916015625, + "eval_Qnli-dev_manhattan_ap": 0.7232893442260485, + "eval_Qnli-dev_manhattan_f1": 0.678714859437751, + "eval_Qnli-dev_manhattan_f1_threshold": 363.3157653808594, + "eval_Qnli-dev_manhattan_precision": 0.6450381679389313, + "eval_Qnli-dev_manhattan_recall": 0.7161016949152542, + "eval_Qnli-dev_max_accuracy": 0.6953125, + "eval_Qnli-dev_max_accuracy_threshold": 346.01763916015625, + "eval_Qnli-dev_max_ap": 0.7232893442260485, + "eval_Qnli-dev_max_f1": 0.6799999999999999, + "eval_Qnli-dev_max_f1_threshold": 363.3157653808594, + "eval_Qnli-dev_max_precision": 0.6450381679389313, + "eval_Qnli-dev_max_recall": 0.9194915254237288, + "eval_allNLI-dev_cosine_accuracy": 0.73828125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7844070792198181, + "eval_allNLI-dev_cosine_ap": 0.6334966395637692, + "eval_allNLI-dev_cosine_f1": 0.6426966292134833, + "eval_allNLI-dev_cosine_f1_threshold": 0.6026103496551514, + "eval_allNLI-dev_cosine_precision": 0.5257352941176471, + "eval_allNLI-dev_cosine_recall": 0.8265895953757225, + "eval_allNLI-dev_dot_accuracy": 0.732421875, + "eval_allNLI-dev_dot_accuracy_threshold": 283.69488525390625, + "eval_allNLI-dev_dot_ap": 0.5980221780780246, + "eval_allNLI-dev_dot_f1": 0.6095617529880478, + "eval_allNLI-dev_dot_f1_threshold": 201.48480224609375, + "eval_allNLI-dev_dot_precision": 0.46504559270516715, + "eval_allNLI-dev_dot_recall": 0.884393063583815, + "eval_allNLI-dev_euclidean_accuracy": 0.740234375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 12.730911254882812, + "eval_allNLI-dev_euclidean_ap": 0.6353313789932706, + "eval_allNLI-dev_euclidean_f1": 0.6496519721577727, + "eval_allNLI-dev_euclidean_f1_threshold": 16.85952377319336, + "eval_allNLI-dev_euclidean_precision": 0.5426356589147286, + "eval_allNLI-dev_euclidean_recall": 0.8092485549132948, + "eval_allNLI-dev_manhattan_accuracy": 0.740234375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 273.82623291015625, + "eval_allNLI-dev_manhattan_ap": 0.6346873828150861, + "eval_allNLI-dev_manhattan_f1": 0.6563876651982379, + "eval_allNLI-dev_manhattan_f1_threshold": 367.0364990234375, + "eval_allNLI-dev_manhattan_precision": 0.5302491103202847, + "eval_allNLI-dev_manhattan_recall": 0.861271676300578, + "eval_allNLI-dev_max_accuracy": 0.740234375, + "eval_allNLI-dev_max_accuracy_threshold": 283.69488525390625, + "eval_allNLI-dev_max_ap": 0.6353313789932706, + "eval_allNLI-dev_max_f1": 0.6563876651982379, + "eval_allNLI-dev_max_f1_threshold": 367.0364990234375, + "eval_allNLI-dev_max_precision": 0.5426356589147286, + "eval_allNLI-dev_max_recall": 0.884393063583815, + "eval_sequential_score": 0.7232893442260485, + "eval_sts-test_pearson_cosine": 0.8849752478250572, + "eval_sts-test_pearson_dot": 0.8644269802811118, + "eval_sts-test_pearson_euclidean": 0.9057105226753817, + "eval_sts-test_pearson_manhattan": 0.9066465477509797, + "eval_sts-test_pearson_max": 0.9066465477509797, + "eval_sts-test_spearman_cosine": 0.9074029344649761, + "eval_sts-test_spearman_dot": 0.8581655381241343, + "eval_sts-test_spearman_euclidean": 0.9012328095493753, + "eval_sts-test_spearman_manhattan": 0.9026904476354162, + "eval_sts-test_spearman_max": 0.9074029344649761, + "eval_vitaminc-pairs_loss": 2.633016586303711, + "eval_vitaminc-pairs_runtime": 3.7763, + "eval_vitaminc-pairs_samples_per_second": 33.896, + "eval_vitaminc-pairs_steps_per_second": 0.265, + "step": 2140 + }, + { + "epoch": 2.502923976608187, + "eval_negation-triplets_loss": 1.0334161520004272, + "eval_negation-triplets_runtime": 0.5895, + "eval_negation-triplets_samples_per_second": 217.138, + "eval_negation-triplets_steps_per_second": 1.696, + "step": 2140 + }, + { + "epoch": 2.502923976608187, + "eval_scitail-pairs-pos_loss": 0.059830233454704285, + "eval_scitail-pairs-pos_runtime": 0.823, + "eval_scitail-pairs-pos_samples_per_second": 155.531, + "eval_scitail-pairs-pos_steps_per_second": 1.215, + "step": 2140 + }, + { + "epoch": 2.502923976608187, + "eval_scitail-pairs-qa_loss": 4.852225288232148e-07, + "eval_scitail-pairs-qa_runtime": 0.5641, + "eval_scitail-pairs-qa_samples_per_second": 226.926, + "eval_scitail-pairs-qa_steps_per_second": 1.773, + "step": 2140 + }, + { + "epoch": 2.502923976608187, + "eval_xsum-pairs_loss": 0.005186527036130428, + "eval_xsum-pairs_runtime": 2.759, + "eval_xsum-pairs_samples_per_second": 46.394, + "eval_xsum-pairs_steps_per_second": 0.362, + "step": 2140 + }, + { + "epoch": 2.502923976608187, + "eval_sciq_pairs_loss": 0.0033733854070305824, + "eval_sciq_pairs_runtime": 4.0923, + "eval_sciq_pairs_samples_per_second": 31.278, + "eval_sciq_pairs_steps_per_second": 0.244, + "step": 2140 + }, + { + "epoch": 2.502923976608187, + "eval_qasc_pairs_loss": 0.043544329702854156, + "eval_qasc_pairs_runtime": 0.6082, + "eval_qasc_pairs_samples_per_second": 210.466, + "eval_qasc_pairs_steps_per_second": 1.644, + "step": 2140 + }, + { + "epoch": 2.502923976608187, + "eval_openbookqa_pairs_loss": 1.2967089414596558, + "eval_openbookqa_pairs_runtime": 0.5811, + "eval_openbookqa_pairs_samples_per_second": 220.275, + "eval_openbookqa_pairs_steps_per_second": 1.721, + "step": 2140 + }, + { + "epoch": 2.502923976608187, + "eval_msmarco_pairs_loss": 0.15381145477294922, + "eval_msmarco_pairs_runtime": 1.3305, + "eval_msmarco_pairs_samples_per_second": 96.204, + "eval_msmarco_pairs_steps_per_second": 0.752, + "step": 2140 + }, + { + "epoch": 2.502923976608187, + "eval_nq_pairs_loss": 0.041369207203388214, + "eval_nq_pairs_runtime": 2.8977, + "eval_nq_pairs_samples_per_second": 44.173, + "eval_nq_pairs_steps_per_second": 0.345, + "step": 2140 + }, + { + "epoch": 2.502923976608187, + "eval_trivia_pairs_loss": 0.20083561539649963, + "eval_trivia_pairs_runtime": 3.1634, + "eval_trivia_pairs_samples_per_second": 40.462, + "eval_trivia_pairs_steps_per_second": 0.316, + "step": 2140 + }, + { + "epoch": 2.502923976608187, + "eval_gooaq_pairs_loss": 0.10723757743835449, + "eval_gooaq_pairs_runtime": 0.8585, + "eval_gooaq_pairs_samples_per_second": 149.096, + "eval_gooaq_pairs_steps_per_second": 1.165, + "step": 2140 + }, + { + "epoch": 2.502923976608187, + "eval_paws-pos_loss": 0.04568905383348465, + "eval_paws-pos_runtime": 0.6533, + "eval_paws-pos_samples_per_second": 195.937, + "eval_paws-pos_steps_per_second": 1.531, + "step": 2140 + }, + { + "epoch": 2.502923976608187, + "eval_global_dataset_loss": 0.17478355765342712, + "eval_global_dataset_runtime": 24.0466, + "eval_global_dataset_samples_per_second": 27.572, + "eval_global_dataset_steps_per_second": 0.25, + "step": 2140 + }, + { + "epoch": 2.5087719298245617, + "grad_norm": 0.695496678352356, + "learning_rate": 3.9474149150819254e-05, + "loss": 0.1055, + "step": 2145 + }, + { + "epoch": 2.521637426900585, + "grad_norm": 0.9295246005058289, + "learning_rate": 3.931666538406779e-05, + "loss": 0.1802, + "step": 2156 + }, + { + "epoch": 2.534502923976608, + "grad_norm": 0.26146623492240906, + "learning_rate": 3.915738842461701e-05, + "loss": 0.1823, + "step": 2167 + }, + { + "epoch": 2.5473684210526315, + "grad_norm": 0.3814794421195984, + "learning_rate": 3.899634146172081e-05, + "loss": 0.197, + "step": 2178 + }, + { + "epoch": 2.560233918128655, + "grad_norm": 0.25856730341911316, + "learning_rate": 3.883354794232923e-05, + "loss": 0.1102, + "step": 2189 + }, + { + "epoch": 2.573099415204678, + "grad_norm": 0.80962735414505, + "learning_rate": 3.866903156767484e-05, + "loss": 0.1109, + "step": 2200 + }, + { + "epoch": 2.585964912280702, + "grad_norm": 0.25340330600738525, + "learning_rate": 3.8502816289821976e-05, + "loss": 0.1522, + "step": 2211 + }, + { + "epoch": 2.598830409356725, + "grad_norm": 0.3234657943248749, + "learning_rate": 3.8334926308179625e-05, + "loss": 0.1801, + "step": 2222 + }, + { + "epoch": 2.6116959064327485, + "grad_norm": 1.4256774187088013, + "learning_rate": 3.816538606597815e-05, + "loss": 0.1452, + "step": 2233 + }, + { + "epoch": 2.624561403508772, + "grad_norm": 1.5761231184005737, + "learning_rate": 3.799422024671061e-05, + "loss": 0.1672, + "step": 2244 + }, + { + "epoch": 2.6374269005847952, + "grad_norm": 0.43666452169418335, + "learning_rate": 3.782145377053907e-05, + "loss": 0.0764, + "step": 2255 + }, + { + "epoch": 2.650292397660819, + "grad_norm": 0.31038856506347656, + "learning_rate": 3.7647111790666466e-05, + "loss": 0.1002, + "step": 2266 + }, + { + "epoch": 2.663157894736842, + "grad_norm": 0.17478525638580322, + "learning_rate": 3.7471219689674493e-05, + "loss": 0.1208, + "step": 2277 + }, + { + "epoch": 2.6760233918128655, + "grad_norm": 0.19572170078754425, + "learning_rate": 3.729380307582819e-05, + "loss": 0.0877, + "step": 2288 + }, + { + "epoch": 2.688888888888889, + "grad_norm": 1.2740304470062256, + "learning_rate": 3.711488777934755e-05, + "loss": 0.1054, + "step": 2299 + }, + { + "epoch": 2.7017543859649122, + "grad_norm": 1.924720287322998, + "learning_rate": 3.693449984864695e-05, + "loss": 0.2137, + "step": 2310 + }, + { + "epoch": 2.714619883040936, + "grad_norm": 1.0254656076431274, + "learning_rate": 3.6752665546542685e-05, + "loss": 0.1162, + "step": 2321 + }, + { + "epoch": 2.727485380116959, + "grad_norm": 0.05343569815158844, + "learning_rate": 3.656941134642935e-05, + "loss": 0.1293, + "step": 2332 + }, + { + "epoch": 2.7403508771929825, + "grad_norm": 0.3289341926574707, + "learning_rate": 3.638476392842562e-05, + "loss": 0.1654, + "step": 2343 + }, + { + "epoch": 2.753216374269006, + "grad_norm": 0.42492780089378357, + "learning_rate": 3.619875017548977e-05, + "loss": 0.196, + "step": 2354 + }, + { + "epoch": 2.753216374269006, + "eval_Qnli-dev_cosine_accuracy": 0.689453125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6470696926116943, + "eval_Qnli-dev_cosine_ap": 0.7138444872758793, + "eval_Qnli-dev_cosine_f1": 0.6761710794297352, + "eval_Qnli-dev_cosine_f1_threshold": 0.6348719596862793, + "eval_Qnli-dev_cosine_precision": 0.6509803921568628, + "eval_Qnli-dev_cosine_recall": 0.7033898305084746, + "eval_Qnli-dev_dot_accuracy": 0.671875, + "eval_Qnli-dev_dot_accuracy_threshold": 250.6613006591797, + "eval_Qnli-dev_dot_ap": 0.6606473298104015, + "eval_Qnli-dev_dot_f1": 0.679304897314376, + "eval_Qnli-dev_dot_f1_threshold": 188.62791442871094, + "eval_Qnli-dev_dot_precision": 0.5415617128463476, + "eval_Qnli-dev_dot_recall": 0.9110169491525424, + "eval_Qnli-dev_euclidean_accuracy": 0.689453125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 16.503406524658203, + "eval_Qnli-dev_euclidean_ap": 0.718268865818281, + "eval_Qnli-dev_euclidean_f1": 0.6761565836298933, + "eval_Qnli-dev_euclidean_f1_threshold": 18.686172485351562, + "eval_Qnli-dev_euclidean_precision": 0.5828220858895705, + "eval_Qnli-dev_euclidean_recall": 0.8050847457627118, + "eval_Qnli-dev_manhattan_accuracy": 0.689453125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 353.69952392578125, + "eval_Qnli-dev_manhattan_ap": 0.7222565203523592, + "eval_Qnli-dev_manhattan_f1": 0.6748251748251748, + "eval_Qnli-dev_manhattan_f1_threshold": 403.3136901855469, + "eval_Qnli-dev_manhattan_precision": 0.5744047619047619, + "eval_Qnli-dev_manhattan_recall": 0.8177966101694916, + "eval_Qnli-dev_max_accuracy": 0.689453125, + "eval_Qnli-dev_max_accuracy_threshold": 353.69952392578125, + "eval_Qnli-dev_max_ap": 0.7222565203523592, + "eval_Qnli-dev_max_f1": 0.679304897314376, + "eval_Qnli-dev_max_f1_threshold": 403.3136901855469, + "eval_Qnli-dev_max_precision": 0.6509803921568628, + "eval_Qnli-dev_max_recall": 0.9110169491525424, + "eval_allNLI-dev_cosine_accuracy": 0.740234375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7886154651641846, + "eval_allNLI-dev_cosine_ap": 0.6335542403765874, + "eval_allNLI-dev_cosine_f1": 0.6401766004415012, + "eval_allNLI-dev_cosine_f1_threshold": 0.5957518219947815, + "eval_allNLI-dev_cosine_precision": 0.5178571428571429, + "eval_allNLI-dev_cosine_recall": 0.838150289017341, + "eval_allNLI-dev_dot_accuracy": 0.736328125, + "eval_allNLI-dev_dot_accuracy_threshold": 291.9097595214844, + "eval_allNLI-dev_dot_ap": 0.6032477980963038, + "eval_allNLI-dev_dot_f1": 0.610236220472441, + "eval_allNLI-dev_dot_f1_threshold": 202.70657348632812, + "eval_allNLI-dev_dot_precision": 0.4626865671641791, + "eval_allNLI-dev_dot_recall": 0.8959537572254336, + "eval_allNLI-dev_euclidean_accuracy": 0.744140625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 14.178364753723145, + "eval_allNLI-dev_euclidean_ap": 0.6364564166264124, + "eval_allNLI-dev_euclidean_f1": 0.6464208242950108, + "eval_allNLI-dev_euclidean_f1_threshold": 17.9101505279541, + "eval_allNLI-dev_euclidean_precision": 0.5173611111111112, + "eval_allNLI-dev_euclidean_recall": 0.861271676300578, + "eval_allNLI-dev_manhattan_accuracy": 0.7421875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 300.3572692871094, + "eval_allNLI-dev_manhattan_ap": 0.6360368960362834, + "eval_allNLI-dev_manhattan_f1": 0.6505494505494506, + "eval_allNLI-dev_manhattan_f1_threshold": 373.840087890625, + "eval_allNLI-dev_manhattan_precision": 0.524822695035461, + "eval_allNLI-dev_manhattan_recall": 0.8554913294797688, + "eval_allNLI-dev_max_accuracy": 0.744140625, + "eval_allNLI-dev_max_accuracy_threshold": 300.3572692871094, + "eval_allNLI-dev_max_ap": 0.6364564166264124, + "eval_allNLI-dev_max_f1": 0.6505494505494506, + "eval_allNLI-dev_max_f1_threshold": 373.840087890625, + "eval_allNLI-dev_max_precision": 0.524822695035461, + "eval_allNLI-dev_max_recall": 0.8959537572254336, + "eval_sequential_score": 0.7222565203523592, + "eval_sts-test_pearson_cosine": 0.8858322159196899, + "eval_sts-test_pearson_dot": 0.863140192479192, + "eval_sts-test_pearson_euclidean": 0.9069987843019685, + "eval_sts-test_pearson_manhattan": 0.90776987866274, + "eval_sts-test_pearson_max": 0.90776987866274, + "eval_sts-test_spearman_cosine": 0.9070992524051741, + "eval_sts-test_spearman_dot": 0.8579407972894266, + "eval_sts-test_spearman_euclidean": 0.9020572620379892, + "eval_sts-test_spearman_manhattan": 0.9038032325011747, + "eval_sts-test_spearman_max": 0.9070992524051741, + "eval_vitaminc-pairs_loss": 2.6483359336853027, + "eval_vitaminc-pairs_runtime": 3.7652, + "eval_vitaminc-pairs_samples_per_second": 33.995, + "eval_vitaminc-pairs_steps_per_second": 0.266, + "step": 2354 + }, + { + "epoch": 2.753216374269006, + "eval_negation-triplets_loss": 1.0219027996063232, + "eval_negation-triplets_runtime": 0.5974, + "eval_negation-triplets_samples_per_second": 214.252, + "eval_negation-triplets_steps_per_second": 1.674, + "step": 2354 + }, + { + "epoch": 2.753216374269006, + "eval_scitail-pairs-pos_loss": 0.055567264556884766, + "eval_scitail-pairs-pos_runtime": 0.8194, + "eval_scitail-pairs-pos_samples_per_second": 156.221, + "eval_scitail-pairs-pos_steps_per_second": 1.22, + "step": 2354 + }, + { + "epoch": 2.753216374269006, + "eval_scitail-pairs-qa_loss": 5.879808782083273e-07, + "eval_scitail-pairs-qa_runtime": 0.5709, + "eval_scitail-pairs-qa_samples_per_second": 224.208, + "eval_scitail-pairs-qa_steps_per_second": 1.752, + "step": 2354 + }, + { + "epoch": 2.753216374269006, + "eval_xsum-pairs_loss": 0.005587125197052956, + "eval_xsum-pairs_runtime": 2.7626, + "eval_xsum-pairs_samples_per_second": 46.334, + "eval_xsum-pairs_steps_per_second": 0.362, + "step": 2354 + }, + { + "epoch": 2.753216374269006, + "eval_sciq_pairs_loss": 0.003840280929580331, + "eval_sciq_pairs_runtime": 4.0341, + "eval_sciq_pairs_samples_per_second": 31.73, + "eval_sciq_pairs_steps_per_second": 0.248, + "step": 2354 + }, + { + "epoch": 2.753216374269006, + "eval_qasc_pairs_loss": 0.04792066663503647, + "eval_qasc_pairs_runtime": 0.5935, + "eval_qasc_pairs_samples_per_second": 215.652, + "eval_qasc_pairs_steps_per_second": 1.685, + "step": 2354 + }, + { + "epoch": 2.753216374269006, + "eval_openbookqa_pairs_loss": 1.2425512075424194, + "eval_openbookqa_pairs_runtime": 0.5546, + "eval_openbookqa_pairs_samples_per_second": 230.793, + "eval_openbookqa_pairs_steps_per_second": 1.803, + "step": 2354 + }, + { + "epoch": 2.753216374269006, + "eval_msmarco_pairs_loss": 0.14102253317832947, + "eval_msmarco_pairs_runtime": 1.315, + "eval_msmarco_pairs_samples_per_second": 97.337, + "eval_msmarco_pairs_steps_per_second": 0.76, + "step": 2354 + }, + { + "epoch": 2.753216374269006, + "eval_nq_pairs_loss": 0.04392912983894348, + "eval_nq_pairs_runtime": 2.9288, + "eval_nq_pairs_samples_per_second": 43.703, + "eval_nq_pairs_steps_per_second": 0.341, + "step": 2354 + }, + { + "epoch": 2.753216374269006, + "eval_trivia_pairs_loss": 0.20342783629894257, + "eval_trivia_pairs_runtime": 3.1575, + "eval_trivia_pairs_samples_per_second": 40.538, + "eval_trivia_pairs_steps_per_second": 0.317, + "step": 2354 + }, + { + "epoch": 2.753216374269006, + "eval_gooaq_pairs_loss": 0.10841929912567139, + "eval_gooaq_pairs_runtime": 0.849, + "eval_gooaq_pairs_samples_per_second": 150.772, + "eval_gooaq_pairs_steps_per_second": 1.178, + "step": 2354 + }, + { + "epoch": 2.753216374269006, + "eval_paws-pos_loss": 0.045823585242033005, + "eval_paws-pos_runtime": 0.6513, + "eval_paws-pos_samples_per_second": 196.526, + "eval_paws-pos_steps_per_second": 1.535, + "step": 2354 + }, + { + "epoch": 2.753216374269006, + "eval_global_dataset_loss": 0.17920929193496704, + "eval_global_dataset_runtime": 24.0295, + "eval_global_dataset_samples_per_second": 27.591, + "eval_global_dataset_steps_per_second": 0.25, + "step": 2354 + } + ], + "logging_steps": 11, + "max_steps": 4275, + "num_input_tokens_seen": 0, + "num_train_epochs": 5, + "save_steps": 214, + "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 +}