diff --git "a/checkpoint-728/trainer_state.json" "b/checkpoint-728/trainer_state.json" new file mode 100644--- /dev/null +++ "b/checkpoint-728/trainer_state.json" @@ -0,0 +1,7131 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 1.403598971722365, + "eval_steps": 26, + "global_step": 728, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.005784061696658098, + "grad_norm": NaN, + "learning_rate": 0.0, + "loss": 7.8503, + "step": 3 + }, + { + "epoch": 0.011568123393316195, + "grad_norm": 127.71849060058594, + "learning_rate": 5.791505791505792e-08, + "loss": 8.4022, + "step": 6 + }, + { + "epoch": 0.017352185089974295, + "grad_norm": 244.984619140625, + "learning_rate": 1.7374517374517373e-07, + "loss": 11.1776, + "step": 9 + }, + { + "epoch": 0.02313624678663239, + "grad_norm": 425.6427917480469, + "learning_rate": 3.4749034749034746e-07, + "loss": 9.7845, + "step": 12 + }, + { + "epoch": 0.02892030848329049, + "grad_norm": 238.8634490966797, + "learning_rate": 5.212355212355212e-07, + "loss": 8.9224, + "step": 15 + }, + { + "epoch": 0.03470437017994859, + "grad_norm": 173.4210968017578, + "learning_rate": 6.949806949806949e-07, + "loss": 11.1202, + "step": 18 + }, + { + "epoch": 0.04048843187660668, + "grad_norm": 77.89308166503906, + "learning_rate": 8.687258687258688e-07, + "loss": 7.413, + "step": 21 + }, + { + "epoch": 0.04627249357326478, + "grad_norm": Infinity, + "learning_rate": 9.845559845559844e-07, + "loss": 7.7803, + "step": 24 + }, + { + "epoch": 0.05012853470437018, + "eval_Qnli-dev_cosine_accuracy": 0.609375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.9602330923080444, + "eval_Qnli-dev_cosine_ap": 0.587737015984062, + "eval_Qnli-dev_cosine_f1": 0.6358695652173914, + "eval_Qnli-dev_cosine_f1_threshold": 0.7741116285324097, + "eval_Qnli-dev_cosine_precision": 0.468, + "eval_Qnli-dev_cosine_recall": 0.9915254237288136, + "eval_Qnli-dev_dot_accuracy": 0.56640625, + "eval_Qnli-dev_dot_accuracy_threshold": 750.123779296875, + "eval_Qnli-dev_dot_ap": 0.5198069471819926, + "eval_Qnli-dev_dot_f1": 0.6318607764390897, + "eval_Qnli-dev_dot_f1_threshold": 387.37371826171875, + "eval_Qnli-dev_dot_precision": 0.461839530332681, + "eval_Qnli-dev_dot_recall": 1.0, + "eval_Qnli-dev_euclidean_accuracy": 0.6171875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 7.253280162811279, + "eval_Qnli-dev_euclidean_ap": 0.5993562103857255, + "eval_Qnli-dev_euclidean_f1": 0.6376021798365122, + "eval_Qnli-dev_euclidean_f1_threshold": 17.42431640625, + "eval_Qnli-dev_euclidean_precision": 0.46987951807228917, + "eval_Qnli-dev_euclidean_recall": 0.9915254237288136, + "eval_Qnli-dev_manhattan_accuracy": 0.625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 212.0604705810547, + "eval_Qnli-dev_manhattan_ap": 0.609409682371795, + "eval_Qnli-dev_manhattan_f1": 0.6358695652173914, + "eval_Qnli-dev_manhattan_f1_threshold": 461.66192626953125, + "eval_Qnli-dev_manhattan_precision": 0.468, + "eval_Qnli-dev_manhattan_recall": 0.9915254237288136, + "eval_Qnli-dev_max_accuracy": 0.625, + "eval_Qnli-dev_max_accuracy_threshold": 750.123779296875, + "eval_Qnli-dev_max_ap": 0.609409682371795, + "eval_Qnli-dev_max_f1": 0.6376021798365122, + "eval_Qnli-dev_max_f1_threshold": 461.66192626953125, + "eval_Qnli-dev_max_precision": 0.46987951807228917, + "eval_Qnli-dev_max_recall": 1.0, + "eval_allNLI-dev_cosine_accuracy": 0.666015625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9931652545928955, + "eval_allNLI-dev_cosine_ap": 0.3330277035108763, + "eval_allNLI-dev_cosine_f1": 0.5088235294117648, + "eval_allNLI-dev_cosine_f1_threshold": 0.40660202503204346, + "eval_allNLI-dev_cosine_precision": 0.34122287968441817, + "eval_allNLI-dev_cosine_recall": 1.0, + "eval_allNLI-dev_dot_accuracy": 0.662109375, + "eval_allNLI-dev_dot_accuracy_threshold": 810.7718505859375, + "eval_allNLI-dev_dot_ap": 0.2980492563498839, + "eval_allNLI-dev_dot_f1": 0.5088235294117648, + "eval_allNLI-dev_dot_f1_threshold": 294.67071533203125, + "eval_allNLI-dev_dot_precision": 0.34122287968441817, + "eval_allNLI-dev_dot_recall": 1.0, + "eval_allNLI-dev_euclidean_accuracy": 0.666015625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 3.3851852416992188, + "eval_allNLI-dev_euclidean_ap": 0.33790858355605125, + "eval_allNLI-dev_euclidean_f1": 0.5095729013254787, + "eval_allNLI-dev_euclidean_f1_threshold": 28.78042984008789, + "eval_allNLI-dev_euclidean_precision": 0.34189723320158105, + "eval_allNLI-dev_euclidean_recall": 1.0, + "eval_allNLI-dev_manhattan_accuracy": 0.666015625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 98.2369613647461, + "eval_allNLI-dev_manhattan_ap": 0.3452833632811151, + "eval_allNLI-dev_manhattan_f1": 0.5091463414634146, + "eval_allNLI-dev_manhattan_f1_threshold": 626.5145874023438, + "eval_allNLI-dev_manhattan_precision": 0.34575569358178054, + "eval_allNLI-dev_manhattan_recall": 0.9653179190751445, + "eval_allNLI-dev_max_accuracy": 0.666015625, + "eval_allNLI-dev_max_accuracy_threshold": 810.7718505859375, + "eval_allNLI-dev_max_ap": 0.3452833632811151, + "eval_allNLI-dev_max_f1": 0.5095729013254787, + "eval_allNLI-dev_max_f1_threshold": 626.5145874023438, + "eval_allNLI-dev_max_precision": 0.34575569358178054, + "eval_allNLI-dev_max_recall": 1.0, + "eval_sequential_score": 0.609409682371795, + "eval_sts-test_pearson_cosine": 0.24255956652335486, + "eval_sts-test_pearson_dot": 0.18022025813169615, + "eval_sts-test_pearson_euclidean": 0.28657386934462287, + "eval_sts-test_pearson_manhattan": 0.32587411772454034, + "eval_sts-test_pearson_max": 0.32587411772454034, + "eval_sts-test_spearman_cosine": 0.32225123770630715, + "eval_sts-test_spearman_dot": 0.17184726031108927, + "eval_sts-test_spearman_euclidean": 0.3279675400801376, + "eval_sts-test_spearman_manhattan": 0.3635547276663291, + "eval_sts-test_spearman_max": 0.3635547276663291, + "eval_vitaminc-pairs_loss": 4.122440338134766, + "eval_vitaminc-pairs_runtime": 4.7337, + "eval_vitaminc-pairs_samples_per_second": 27.04, + "eval_vitaminc-pairs_steps_per_second": 0.211, + "step": 26 + }, + { + "epoch": 0.05012853470437018, + "eval_negation-triplets_loss": 3.917590379714966, + "eval_negation-triplets_runtime": 3.3679, + "eval_negation-triplets_samples_per_second": 38.006, + "eval_negation-triplets_steps_per_second": 0.297, + "step": 26 + }, + { + "epoch": 0.05012853470437018, + "eval_scitail-pairs-pos_loss": 2.5062460899353027, + "eval_scitail-pairs-pos_runtime": 2.7274, + "eval_scitail-pairs-pos_samples_per_second": 46.931, + "eval_scitail-pairs-pos_steps_per_second": 0.367, + "step": 26 + }, + { + "epoch": 0.05012853470437018, + "eval_scitail-pairs-qa_loss": 2.978612184524536, + "eval_scitail-pairs-qa_runtime": 2.2452, + "eval_scitail-pairs-qa_samples_per_second": 57.011, + "eval_scitail-pairs-qa_steps_per_second": 0.445, + "step": 26 + }, + { + "epoch": 0.05012853470437018, + "eval_xsum-pairs_loss": 5.394700050354004, + "eval_xsum-pairs_runtime": 3.0886, + "eval_xsum-pairs_samples_per_second": 41.442, + "eval_xsum-pairs_steps_per_second": 0.324, + "step": 26 + }, + { + "epoch": 0.05012853470437018, + "eval_sciq_pairs_loss": 0.8977569937705994, + "eval_sciq_pairs_runtime": 4.7516, + "eval_sciq_pairs_samples_per_second": 26.938, + "eval_sciq_pairs_steps_per_second": 0.21, + "step": 26 + }, + { + "epoch": 0.05012853470437018, + "eval_qasc_pairs_loss": 6.6808977127075195, + "eval_qasc_pairs_runtime": 2.1087, + "eval_qasc_pairs_samples_per_second": 60.701, + "eval_qasc_pairs_steps_per_second": 0.474, + "step": 26 + }, + { + "epoch": 0.05012853470437018, + "eval_openbookqa_pairs_loss": 6.402599334716797, + "eval_openbookqa_pairs_runtime": 2.2222, + "eval_openbookqa_pairs_samples_per_second": 57.601, + "eval_openbookqa_pairs_steps_per_second": 0.45, + "step": 26 + }, + { + "epoch": 0.05012853470437018, + "eval_msmarco_pairs_loss": 15.782674789428711, + "eval_msmarco_pairs_runtime": 2.307, + "eval_msmarco_pairs_samples_per_second": 55.483, + "eval_msmarco_pairs_steps_per_second": 0.433, + "step": 26 + }, + { + "epoch": 0.05012853470437018, + "eval_nq_pairs_loss": 10.447425842285156, + "eval_nq_pairs_runtime": 3.6239, + "eval_nq_pairs_samples_per_second": 35.321, + "eval_nq_pairs_steps_per_second": 0.276, + "step": 26 + }, + { + "epoch": 0.05012853470437018, + "eval_trivia_pairs_loss": 6.107333183288574, + "eval_trivia_pairs_runtime": 3.1299, + "eval_trivia_pairs_samples_per_second": 40.895, + "eval_trivia_pairs_steps_per_second": 0.319, + "step": 26 + }, + { + "epoch": 0.05012853470437018, + "eval_gooaq_pairs_loss": 8.095273971557617, + "eval_gooaq_pairs_runtime": 2.1098, + "eval_gooaq_pairs_samples_per_second": 60.669, + "eval_gooaq_pairs_steps_per_second": 0.474, + "step": 26 + }, + { + "epoch": 0.05012853470437018, + "eval_paws-pos_loss": 1.6077361106872559, + "eval_paws-pos_runtime": 2.2412, + "eval_paws-pos_samples_per_second": 57.112, + "eval_paws-pos_steps_per_second": 0.446, + "step": 26 + }, + { + "epoch": 0.05012853470437018, + "eval_global_dataset_loss": 6.348417282104492, + "eval_global_dataset_runtime": 8.3683, + "eval_global_dataset_samples_per_second": 38.837, + "eval_global_dataset_steps_per_second": 0.358, + "step": 26 + }, + { + "epoch": 0.05205655526992288, + "grad_norm": 172.1571807861328, + "learning_rate": 1.1583011583011583e-06, + "loss": 7.9729, + "step": 27 + }, + { + "epoch": 0.05784061696658098, + "grad_norm": 106.302001953125, + "learning_rate": 1.3320463320463321e-06, + "loss": 6.0587, + "step": 30 + }, + { + "epoch": 0.06362467866323908, + "grad_norm": 72.31057739257812, + "learning_rate": 1.5057915057915057e-06, + "loss": 5.6742, + "step": 33 + }, + { + "epoch": 0.06940874035989718, + "grad_norm": 258.2577209472656, + "learning_rate": 1.6795366795366796e-06, + "loss": 6.5406, + "step": 36 + }, + { + "epoch": 0.07519280205655526, + "grad_norm": 156.0019073486328, + "learning_rate": 1.7953667953667953e-06, + "loss": 5.4429, + "step": 39 + }, + { + "epoch": 0.08097686375321336, + "grad_norm": 227.6062469482422, + "learning_rate": 1.969111969111969e-06, + "loss": 6.7855, + "step": 42 + }, + { + "epoch": 0.08676092544987146, + "grad_norm": 221.12347412109375, + "learning_rate": 2.1428571428571427e-06, + "loss": 5.3403, + "step": 45 + }, + { + "epoch": 0.09254498714652956, + "grad_norm": 363.6319274902344, + "learning_rate": 2.3166023166023166e-06, + "loss": 4.2282, + "step": 48 + }, + { + "epoch": 0.09832904884318766, + "grad_norm": 119.22310638427734, + "learning_rate": 2.4903474903474904e-06, + "loss": 4.7411, + "step": 51 + }, + { + "epoch": 0.10025706940874037, + "eval_Qnli-dev_cosine_accuracy": 0.58984375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.9407284259796143, + "eval_Qnli-dev_cosine_ap": 0.5957496162735323, + "eval_Qnli-dev_cosine_f1": 0.6314363143631436, + "eval_Qnli-dev_cosine_f1_threshold": 0.7475543022155762, + "eval_Qnli-dev_cosine_precision": 0.4641434262948207, + "eval_Qnli-dev_cosine_recall": 0.9872881355932204, + "eval_Qnli-dev_dot_accuracy": 0.591796875, + "eval_Qnli-dev_dot_accuracy_threshold": 787.171630859375, + "eval_Qnli-dev_dot_ap": 0.556540262374452, + "eval_Qnli-dev_dot_f1": 0.6328767123287672, + "eval_Qnli-dev_dot_f1_threshold": 576.0565795898438, + "eval_Qnli-dev_dot_precision": 0.4676113360323887, + "eval_Qnli-dev_dot_recall": 0.9788135593220338, + "eval_Qnli-dev_euclidean_accuracy": 0.587890625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 9.92269515991211, + "eval_Qnli-dev_euclidean_ap": 0.5913529071853763, + "eval_Qnli-dev_euclidean_f1": 0.6322930800542741, + "eval_Qnli-dev_euclidean_f1_threshold": 19.546676635742188, + "eval_Qnli-dev_euclidean_precision": 0.46506986027944114, + "eval_Qnli-dev_euclidean_recall": 0.9872881355932204, + "eval_Qnli-dev_manhattan_accuracy": 0.591796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 314.9248352050781, + "eval_Qnli-dev_manhattan_ap": 0.5944730102616272, + "eval_Qnli-dev_manhattan_f1": 0.632952691680261, + "eval_Qnli-dev_manhattan_f1_threshold": 424.9676818847656, + "eval_Qnli-dev_manhattan_precision": 0.5145888594164456, + "eval_Qnli-dev_manhattan_recall": 0.8220338983050848, + "eval_Qnli-dev_max_accuracy": 0.591796875, + "eval_Qnli-dev_max_accuracy_threshold": 787.171630859375, + "eval_Qnli-dev_max_ap": 0.5957496162735323, + "eval_Qnli-dev_max_f1": 0.632952691680261, + "eval_Qnli-dev_max_f1_threshold": 576.0565795898438, + "eval_Qnli-dev_max_precision": 0.5145888594164456, + "eval_Qnli-dev_max_recall": 0.9872881355932204, + "eval_allNLI-dev_cosine_accuracy": 0.666015625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9878466129302979, + "eval_allNLI-dev_cosine_ap": 0.3898337405414008, + "eval_allNLI-dev_cosine_f1": 0.5065885797950219, + "eval_allNLI-dev_cosine_f1_threshold": 0.6534674167633057, + "eval_allNLI-dev_cosine_precision": 0.3392156862745098, + "eval_allNLI-dev_cosine_recall": 1.0, + "eval_allNLI-dev_dot_accuracy": 0.666015625, + "eval_allNLI-dev_dot_accuracy_threshold": 925.5989990234375, + "eval_allNLI-dev_dot_ap": 0.3391870470132937, + "eval_allNLI-dev_dot_f1": 0.5046153846153847, + "eval_allNLI-dev_dot_f1_threshold": 608.7012939453125, + "eval_allNLI-dev_dot_precision": 0.3438155136268344, + "eval_allNLI-dev_dot_recall": 0.9479768786127167, + "eval_allNLI-dev_euclidean_accuracy": 0.666015625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 4.708977699279785, + "eval_allNLI-dev_euclidean_ap": 0.39556533875401867, + "eval_allNLI-dev_euclidean_f1": 0.5080763582966226, + "eval_allNLI-dev_euclidean_f1_threshold": 23.160057067871094, + "eval_allNLI-dev_euclidean_precision": 0.3405511811023622, + "eval_allNLI-dev_euclidean_recall": 1.0, + "eval_allNLI-dev_manhattan_accuracy": 0.666015625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 142.7260284423828, + "eval_allNLI-dev_manhattan_ap": 0.4050661514301679, + "eval_allNLI-dev_manhattan_f1": 0.5118483412322274, + "eval_allNLI-dev_manhattan_f1_threshold": 504.7154846191406, + "eval_allNLI-dev_manhattan_precision": 0.3521739130434783, + "eval_allNLI-dev_manhattan_recall": 0.9364161849710982, + "eval_allNLI-dev_max_accuracy": 0.666015625, + "eval_allNLI-dev_max_accuracy_threshold": 925.5989990234375, + "eval_allNLI-dev_max_ap": 0.4050661514301679, + "eval_allNLI-dev_max_f1": 0.5118483412322274, + "eval_allNLI-dev_max_f1_threshold": 608.7012939453125, + "eval_allNLI-dev_max_precision": 0.3521739130434783, + "eval_allNLI-dev_max_recall": 1.0, + "eval_sequential_score": 0.5957496162735323, + "eval_sts-test_pearson_cosine": 0.5067228548513025, + "eval_sts-test_pearson_dot": 0.28867418819353924, + "eval_sts-test_pearson_euclidean": 0.5315956249457565, + "eval_sts-test_pearson_manhattan": 0.5620907203006935, + "eval_sts-test_pearson_max": 0.5620907203006935, + "eval_sts-test_spearman_cosine": 0.5257409961657202, + "eval_sts-test_spearman_dot": 0.280153994622986, + "eval_sts-test_spearman_euclidean": 0.5306475244646732, + "eval_sts-test_spearman_manhattan": 0.5605804796765843, + "eval_sts-test_spearman_max": 0.5605804796765843, + "eval_vitaminc-pairs_loss": 3.929797887802124, + "eval_vitaminc-pairs_runtime": 4.6018, + "eval_vitaminc-pairs_samples_per_second": 27.815, + "eval_vitaminc-pairs_steps_per_second": 0.217, + "step": 52 + }, + { + "epoch": 0.10025706940874037, + "eval_negation-triplets_loss": 3.267047166824341, + "eval_negation-triplets_runtime": 3.2813, + "eval_negation-triplets_samples_per_second": 39.009, + "eval_negation-triplets_steps_per_second": 0.305, + "step": 52 + }, + { + "epoch": 0.10025706940874037, + "eval_scitail-pairs-pos_loss": 1.134178876876831, + "eval_scitail-pairs-pos_runtime": 2.6261, + "eval_scitail-pairs-pos_samples_per_second": 48.742, + "eval_scitail-pairs-pos_steps_per_second": 0.381, + "step": 52 + }, + { + "epoch": 0.10025706940874037, + "eval_scitail-pairs-qa_loss": 2.0085601806640625, + "eval_scitail-pairs-qa_runtime": 2.2161, + "eval_scitail-pairs-qa_samples_per_second": 57.759, + "eval_scitail-pairs-qa_steps_per_second": 0.451, + "step": 52 + }, + { + "epoch": 0.10025706940874037, + "eval_xsum-pairs_loss": 4.462350845336914, + "eval_xsum-pairs_runtime": 3.1126, + "eval_xsum-pairs_samples_per_second": 41.124, + "eval_xsum-pairs_steps_per_second": 0.321, + "step": 52 + }, + { + "epoch": 0.10025706940874037, + "eval_sciq_pairs_loss": 0.71834796667099, + "eval_sciq_pairs_runtime": 4.7505, + "eval_sciq_pairs_samples_per_second": 26.944, + "eval_sciq_pairs_steps_per_second": 0.211, + "step": 52 + }, + { + "epoch": 0.10025706940874037, + "eval_qasc_pairs_loss": 4.889453887939453, + "eval_qasc_pairs_runtime": 2.1101, + "eval_qasc_pairs_samples_per_second": 60.662, + "eval_qasc_pairs_steps_per_second": 0.474, + "step": 52 + }, + { + "epoch": 0.10025706940874037, + "eval_openbookqa_pairs_loss": 3.4273369312286377, + "eval_openbookqa_pairs_runtime": 2.2311, + "eval_openbookqa_pairs_samples_per_second": 57.371, + "eval_openbookqa_pairs_steps_per_second": 0.448, + "step": 52 + }, + { + "epoch": 0.10025706940874037, + "eval_msmarco_pairs_loss": 7.532540321350098, + "eval_msmarco_pairs_runtime": 2.2959, + "eval_msmarco_pairs_samples_per_second": 55.751, + "eval_msmarco_pairs_steps_per_second": 0.436, + "step": 52 + }, + { + "epoch": 0.10025706940874037, + "eval_nq_pairs_loss": 6.29541015625, + "eval_nq_pairs_runtime": 3.661, + "eval_nq_pairs_samples_per_second": 34.963, + "eval_nq_pairs_steps_per_second": 0.273, + "step": 52 + }, + { + "epoch": 0.10025706940874037, + "eval_trivia_pairs_loss": 5.789977073669434, + "eval_trivia_pairs_runtime": 3.0826, + "eval_trivia_pairs_samples_per_second": 41.523, + "eval_trivia_pairs_steps_per_second": 0.324, + "step": 52 + }, + { + "epoch": 0.10025706940874037, + "eval_gooaq_pairs_loss": 5.785163879394531, + "eval_gooaq_pairs_runtime": 2.1382, + "eval_gooaq_pairs_samples_per_second": 59.863, + "eval_gooaq_pairs_steps_per_second": 0.468, + "step": 52 + }, + { + "epoch": 0.10025706940874037, + "eval_paws-pos_loss": 0.47830095887184143, + "eval_paws-pos_runtime": 2.3113, + "eval_paws-pos_samples_per_second": 55.379, + "eval_paws-pos_steps_per_second": 0.433, + "step": 52 + }, + { + "epoch": 0.10025706940874037, + "eval_global_dataset_loss": 3.8913991451263428, + "eval_global_dataset_runtime": 8.3623, + "eval_global_dataset_samples_per_second": 38.865, + "eval_global_dataset_steps_per_second": 0.359, + "step": 52 + }, + { + "epoch": 0.10411311053984576, + "grad_norm": 55.83478546142578, + "learning_rate": 2.6640926640926642e-06, + "loss": 3.9082, + "step": 54 + }, + { + "epoch": 0.10989717223650386, + "grad_norm": 57.4188117980957, + "learning_rate": 2.7799227799227797e-06, + "loss": 4.3922, + "step": 57 + }, + { + "epoch": 0.11568123393316196, + "grad_norm": 116.89701843261719, + "learning_rate": 2.9536679536679535e-06, + "loss": 3.2655, + "step": 60 + }, + { + "epoch": 0.12146529562982006, + "grad_norm": 53.49580001831055, + "learning_rate": 3.1274131274131274e-06, + "loss": 3.1043, + "step": 63 + }, + { + "epoch": 0.12724935732647816, + "grad_norm": 51.044410705566406, + "learning_rate": 3.3011583011583012e-06, + "loss": 2.2074, + "step": 66 + }, + { + "epoch": 0.13303341902313626, + "grad_norm": 139.9288787841797, + "learning_rate": 3.474903474903475e-06, + "loss": 1.4414, + "step": 69 + }, + { + "epoch": 0.13881748071979436, + "grad_norm": 37.221168518066406, + "learning_rate": 3.5907335907335905e-06, + "loss": 1.5937, + "step": 72 + }, + { + "epoch": 0.14460154241645246, + "grad_norm": 25.855661392211914, + "learning_rate": 3.7644787644787644e-06, + "loss": 1.0306, + "step": 75 + }, + { + "epoch": 0.15038560411311053, + "grad_norm": 29.79730796813965, + "learning_rate": 3.938223938223938e-06, + "loss": 1.0784, + "step": 78 + }, + { + "epoch": 0.15038560411311053, + "eval_Qnli-dev_cosine_accuracy": 0.638671875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.856202244758606, + "eval_Qnli-dev_cosine_ap": 0.6292803059732273, + "eval_Qnli-dev_cosine_f1": 0.6408839779005525, + "eval_Qnli-dev_cosine_f1_threshold": 0.6206527352333069, + "eval_Qnli-dev_cosine_precision": 0.47540983606557374, + "eval_Qnli-dev_cosine_recall": 0.9830508474576272, + "eval_Qnli-dev_dot_accuracy": 0.626953125, + "eval_Qnli-dev_dot_accuracy_threshold": 887.830810546875, + "eval_Qnli-dev_dot_ap": 0.592075047542057, + "eval_Qnli-dev_dot_f1": 0.640990371389271, + "eval_Qnli-dev_dot_f1_threshold": 601.355224609375, + "eval_Qnli-dev_dot_precision": 0.4745417515274949, + "eval_Qnli-dev_dot_recall": 0.9872881355932204, + "eval_Qnli-dev_euclidean_accuracy": 0.638671875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 17.094371795654297, + "eval_Qnli-dev_euclidean_ap": 0.6289192660680781, + "eval_Qnli-dev_euclidean_f1": 0.6423357664233577, + "eval_Qnli-dev_euclidean_f1_threshold": 24.894683837890625, + "eval_Qnli-dev_euclidean_precision": 0.48997772828507796, + "eval_Qnli-dev_euclidean_recall": 0.9322033898305084, + "eval_Qnli-dev_manhattan_accuracy": 0.640625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 526.0111083984375, + "eval_Qnli-dev_manhattan_ap": 0.6256143574081243, + "eval_Qnli-dev_manhattan_f1": 0.6417704011065007, + "eval_Qnli-dev_manhattan_f1_threshold": 853.242431640625, + "eval_Qnli-dev_manhattan_precision": 0.47638603696098564, + "eval_Qnli-dev_manhattan_recall": 0.9830508474576272, + "eval_Qnli-dev_max_accuracy": 0.640625, + "eval_Qnli-dev_max_accuracy_threshold": 887.830810546875, + "eval_Qnli-dev_max_ap": 0.6292803059732273, + "eval_Qnli-dev_max_f1": 0.6423357664233577, + "eval_Qnli-dev_max_f1_threshold": 853.242431640625, + "eval_Qnli-dev_max_precision": 0.48997772828507796, + "eval_Qnli-dev_max_recall": 0.9872881355932204, + "eval_allNLI-dev_cosine_accuracy": 0.6796875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.9426460862159729, + "eval_allNLI-dev_cosine_ap": 0.473508692504868, + "eval_allNLI-dev_cosine_f1": 0.5328947368421053, + "eval_allNLI-dev_cosine_f1_threshold": 0.6884603500366211, + "eval_allNLI-dev_cosine_precision": 0.3724137931034483, + "eval_allNLI-dev_cosine_recall": 0.9364161849710982, + "eval_allNLI-dev_dot_accuracy": 0.673828125, + "eval_allNLI-dev_dot_accuracy_threshold": 980.9593505859375, + "eval_allNLI-dev_dot_ap": 0.44188020452374555, + "eval_allNLI-dev_dot_f1": 0.53156146179402, + "eval_allNLI-dev_dot_f1_threshold": 683.7774658203125, + "eval_allNLI-dev_dot_precision": 0.372960372960373, + "eval_allNLI-dev_dot_recall": 0.9248554913294798, + "eval_allNLI-dev_euclidean_accuracy": 0.6796875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 10.647405624389648, + "eval_allNLI-dev_euclidean_ap": 0.47440611686548606, + "eval_allNLI-dev_euclidean_f1": 0.5292207792207793, + "eval_allNLI-dev_euclidean_f1_threshold": 25.08370590209961, + "eval_allNLI-dev_euclidean_precision": 0.36794582392776526, + "eval_allNLI-dev_euclidean_recall": 0.9421965317919075, + "eval_allNLI-dev_manhattan_accuracy": 0.6796875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 318.408203125, + "eval_allNLI-dev_manhattan_ap": 0.4747137405292771, + "eval_allNLI-dev_manhattan_f1": 0.5305343511450381, + "eval_allNLI-dev_manhattan_f1_threshold": 697.8466796875, + "eval_allNLI-dev_manhattan_precision": 0.396011396011396, + "eval_allNLI-dev_manhattan_recall": 0.8034682080924855, + "eval_allNLI-dev_max_accuracy": 0.6796875, + "eval_allNLI-dev_max_accuracy_threshold": 980.9593505859375, + "eval_allNLI-dev_max_ap": 0.4747137405292771, + "eval_allNLI-dev_max_f1": 0.5328947368421053, + "eval_allNLI-dev_max_f1_threshold": 697.8466796875, + "eval_allNLI-dev_max_precision": 0.396011396011396, + "eval_allNLI-dev_max_recall": 0.9421965317919075, + "eval_sequential_score": 0.6292803059732273, + "eval_sts-test_pearson_cosine": 0.8216305666941104, + "eval_sts-test_pearson_dot": 0.7893058557928705, + "eval_sts-test_pearson_euclidean": 0.841701309677291, + "eval_sts-test_pearson_manhattan": 0.8438559225680149, + "eval_sts-test_pearson_max": 0.8438559225680149, + "eval_sts-test_spearman_cosine": 0.8291168678373173, + "eval_sts-test_spearman_dot": 0.7782024179758329, + "eval_sts-test_spearman_euclidean": 0.829029110613729, + "eval_sts-test_spearman_manhattan": 0.8308224279114665, + "eval_sts-test_spearman_max": 0.8308224279114665, + "eval_vitaminc-pairs_loss": 3.763850212097168, + "eval_vitaminc-pairs_runtime": 4.6142, + "eval_vitaminc-pairs_samples_per_second": 27.741, + "eval_vitaminc-pairs_steps_per_second": 0.217, + "step": 78 + }, + { + "epoch": 0.15038560411311053, + "eval_negation-triplets_loss": 1.4772971868515015, + "eval_negation-triplets_runtime": 3.4145, + "eval_negation-triplets_samples_per_second": 37.488, + "eval_negation-triplets_steps_per_second": 0.293, + "step": 78 + }, + { + "epoch": 0.15038560411311053, + "eval_scitail-pairs-pos_loss": 0.11086193472146988, + "eval_scitail-pairs-pos_runtime": 2.8091, + "eval_scitail-pairs-pos_samples_per_second": 45.565, + "eval_scitail-pairs-pos_steps_per_second": 0.356, + "step": 78 + }, + { + "epoch": 0.15038560411311053, + "eval_scitail-pairs-qa_loss": 0.12822097539901733, + "eval_scitail-pairs-qa_runtime": 2.312, + "eval_scitail-pairs-qa_samples_per_second": 55.364, + "eval_scitail-pairs-qa_steps_per_second": 0.433, + "step": 78 + }, + { + "epoch": 0.15038560411311053, + "eval_xsum-pairs_loss": 0.43598121404647827, + "eval_xsum-pairs_runtime": 3.1471, + "eval_xsum-pairs_samples_per_second": 40.673, + "eval_xsum-pairs_steps_per_second": 0.318, + "step": 78 + }, + { + "epoch": 0.15038560411311053, + "eval_sciq_pairs_loss": 0.20338551700115204, + "eval_sciq_pairs_runtime": 4.8592, + "eval_sciq_pairs_samples_per_second": 26.342, + "eval_sciq_pairs_steps_per_second": 0.206, + "step": 78 + }, + { + "epoch": 0.15038560411311053, + "eval_qasc_pairs_loss": 1.3059492111206055, + "eval_qasc_pairs_runtime": 2.1191, + "eval_qasc_pairs_samples_per_second": 60.403, + "eval_qasc_pairs_steps_per_second": 0.472, + "step": 78 + }, + { + "epoch": 0.15038560411311053, + "eval_openbookqa_pairs_loss": 1.1711227893829346, + "eval_openbookqa_pairs_runtime": 2.2105, + "eval_openbookqa_pairs_samples_per_second": 57.907, + "eval_openbookqa_pairs_steps_per_second": 0.452, + "step": 78 + }, + { + "epoch": 0.15038560411311053, + "eval_msmarco_pairs_loss": 1.3856205940246582, + "eval_msmarco_pairs_runtime": 2.3065, + "eval_msmarco_pairs_samples_per_second": 55.494, + "eval_msmarco_pairs_steps_per_second": 0.434, + "step": 78 + }, + { + "epoch": 0.15038560411311053, + "eval_nq_pairs_loss": 1.2713654041290283, + "eval_nq_pairs_runtime": 3.6729, + "eval_nq_pairs_samples_per_second": 34.85, + "eval_nq_pairs_steps_per_second": 0.272, + "step": 78 + }, + { + "epoch": 0.15038560411311053, + "eval_trivia_pairs_loss": 1.082784652709961, + "eval_trivia_pairs_runtime": 3.145, + "eval_trivia_pairs_samples_per_second": 40.7, + "eval_trivia_pairs_steps_per_second": 0.318, + "step": 78 + }, + { + "epoch": 0.15038560411311053, + "eval_gooaq_pairs_loss": 0.703183650970459, + "eval_gooaq_pairs_runtime": 2.1975, + "eval_gooaq_pairs_samples_per_second": 58.248, + "eval_gooaq_pairs_steps_per_second": 0.455, + "step": 78 + }, + { + "epoch": 0.15038560411311053, + "eval_paws-pos_loss": 0.04489962384104729, + "eval_paws-pos_runtime": 2.3636, + "eval_paws-pos_samples_per_second": 54.155, + "eval_paws-pos_steps_per_second": 0.423, + "step": 78 + }, + { + "epoch": 0.15038560411311053, + "eval_global_dataset_loss": 0.6713590621948242, + "eval_global_dataset_runtime": 8.4648, + "eval_global_dataset_samples_per_second": 38.394, + "eval_global_dataset_steps_per_second": 0.354, + "step": 78 + }, + { + "epoch": 0.15616966580976863, + "grad_norm": 22.163389205932617, + "learning_rate": 4.1119691119691125e-06, + "loss": 0.9674, + "step": 81 + }, + { + "epoch": 0.16195372750642673, + "grad_norm": 25.121131896972656, + "learning_rate": 4.2857142857142855e-06, + "loss": 0.9335, + "step": 84 + }, + { + "epoch": 0.16773778920308482, + "grad_norm": 44.308528900146484, + "learning_rate": 4.45945945945946e-06, + "loss": 0.8806, + "step": 87 + }, + { + "epoch": 0.17352185089974292, + "grad_norm": 9.972362518310547, + "learning_rate": 4.633204633204633e-06, + "loss": 0.631, + "step": 90 + }, + { + "epoch": 0.17930591259640102, + "grad_norm": 13.609597206115723, + "learning_rate": 4.806949806949808e-06, + "loss": 0.3384, + "step": 93 + }, + { + "epoch": 0.18508997429305912, + "grad_norm": 6.892751693725586, + "learning_rate": 4.980694980694981e-06, + "loss": 0.404, + "step": 96 + }, + { + "epoch": 0.19087403598971722, + "grad_norm": 7.414526462554932, + "learning_rate": 5.154440154440155e-06, + "loss": 0.6488, + "step": 99 + }, + { + "epoch": 0.19665809768637532, + "grad_norm": 23.75099754333496, + "learning_rate": 5.3281853281853285e-06, + "loss": 0.4728, + "step": 102 + }, + { + "epoch": 0.20051413881748073, + "eval_Qnli-dev_cosine_accuracy": 0.693359375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.8276629447937012, + "eval_Qnli-dev_cosine_ap": 0.7127503402574797, + "eval_Qnli-dev_cosine_f1": 0.6730038022813688, + "eval_Qnli-dev_cosine_f1_threshold": 0.7761298418045044, + "eval_Qnli-dev_cosine_precision": 0.6103448275862069, + "eval_Qnli-dev_cosine_recall": 0.75, + "eval_Qnli-dev_dot_accuracy": 0.681640625, + "eval_Qnli-dev_dot_accuracy_threshold": 857.5540771484375, + "eval_Qnli-dev_dot_ap": 0.6617256136301952, + "eval_Qnli-dev_dot_f1": 0.6678141135972461, + "eval_Qnli-dev_dot_f1_threshold": 756.1309204101562, + "eval_Qnli-dev_dot_precision": 0.5623188405797102, + "eval_Qnli-dev_dot_recall": 0.8220338983050848, + "eval_Qnli-dev_euclidean_accuracy": 0.69140625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 18.45541763305664, + "eval_Qnli-dev_euclidean_ap": 0.7154480594709788, + "eval_Qnli-dev_euclidean_f1": 0.6725043782837128, + "eval_Qnli-dev_euclidean_f1_threshold": 22.723068237304688, + "eval_Qnli-dev_euclidean_precision": 0.573134328358209, + "eval_Qnli-dev_euclidean_recall": 0.8135593220338984, + "eval_Qnli-dev_manhattan_accuracy": 0.69140625, + "eval_Qnli-dev_manhattan_accuracy_threshold": 585.3442993164062, + "eval_Qnli-dev_manhattan_ap": 0.713289230749026, + "eval_Qnli-dev_manhattan_f1": 0.6733333333333333, + "eval_Qnli-dev_manhattan_f1_threshold": 733.2077026367188, + "eval_Qnli-dev_manhattan_precision": 0.554945054945055, + "eval_Qnli-dev_manhattan_recall": 0.8559322033898306, + "eval_Qnli-dev_max_accuracy": 0.693359375, + "eval_Qnli-dev_max_accuracy_threshold": 857.5540771484375, + "eval_Qnli-dev_max_ap": 0.7154480594709788, + "eval_Qnli-dev_max_f1": 0.6733333333333333, + "eval_Qnli-dev_max_f1_threshold": 756.1309204101562, + "eval_Qnli-dev_max_precision": 0.6103448275862069, + "eval_Qnli-dev_max_recall": 0.8559322033898306, + "eval_allNLI-dev_cosine_accuracy": 0.693359375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8820043802261353, + "eval_allNLI-dev_cosine_ap": 0.5380653098565689, + "eval_allNLI-dev_cosine_f1": 0.5689655172413792, + "eval_allNLI-dev_cosine_f1_threshold": 0.7308224439620972, + "eval_allNLI-dev_cosine_precision": 0.4536082474226804, + "eval_allNLI-dev_cosine_recall": 0.7630057803468208, + "eval_allNLI-dev_dot_accuracy": 0.6875, + "eval_allNLI-dev_dot_accuracy_threshold": 904.0134887695312, + "eval_allNLI-dev_dot_ap": 0.5175334662627952, + "eval_allNLI-dev_dot_f1": 0.5675675675675675, + "eval_allNLI-dev_dot_f1_threshold": 762.652587890625, + "eval_allNLI-dev_dot_precision": 0.46494464944649444, + "eval_allNLI-dev_dot_recall": 0.7283236994219653, + "eval_allNLI-dev_euclidean_accuracy": 0.6953125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 15.50227165222168, + "eval_allNLI-dev_euclidean_ap": 0.5410854453600531, + "eval_allNLI-dev_euclidean_f1": 0.5645514223194749, + "eval_allNLI-dev_euclidean_f1_threshold": 23.451454162597656, + "eval_allNLI-dev_euclidean_precision": 0.45422535211267606, + "eval_allNLI-dev_euclidean_recall": 0.7456647398843931, + "eval_allNLI-dev_manhattan_accuracy": 0.6953125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 518.75146484375, + "eval_allNLI-dev_manhattan_ap": 0.5430149701961973, + "eval_allNLI-dev_manhattan_f1": 0.5639913232104121, + "eval_allNLI-dev_manhattan_f1_threshold": 733.8897705078125, + "eval_allNLI-dev_manhattan_precision": 0.4513888888888889, + "eval_allNLI-dev_manhattan_recall": 0.7514450867052023, + "eval_allNLI-dev_max_accuracy": 0.6953125, + "eval_allNLI-dev_max_accuracy_threshold": 904.0134887695312, + "eval_allNLI-dev_max_ap": 0.5430149701961973, + "eval_allNLI-dev_max_f1": 0.5689655172413792, + "eval_allNLI-dev_max_f1_threshold": 762.652587890625, + "eval_allNLI-dev_max_precision": 0.46494464944649444, + "eval_allNLI-dev_max_recall": 0.7630057803468208, + "eval_sequential_score": 0.7154480594709788, + "eval_sts-test_pearson_cosine": 0.8841456597076958, + "eval_sts-test_pearson_dot": 0.8597263949983345, + "eval_sts-test_pearson_euclidean": 0.8977156894521804, + "eval_sts-test_pearson_manhattan": 0.8973099980705126, + "eval_sts-test_pearson_max": 0.8977156894521804, + "eval_sts-test_spearman_cosine": 0.8862653474156692, + "eval_sts-test_spearman_dot": 0.8451249310521576, + "eval_sts-test_spearman_euclidean": 0.8866065310325173, + "eval_sts-test_spearman_manhattan": 0.8860841943095571, + "eval_sts-test_spearman_max": 0.8866065310325173, + "eval_vitaminc-pairs_loss": 3.314167022705078, + "eval_vitaminc-pairs_runtime": 4.4886, + "eval_vitaminc-pairs_samples_per_second": 28.517, + "eval_vitaminc-pairs_steps_per_second": 0.223, + "step": 104 + }, + { + "epoch": 0.20051413881748073, + "eval_negation-triplets_loss": 1.3400739431381226, + "eval_negation-triplets_runtime": 3.3068, + "eval_negation-triplets_samples_per_second": 38.708, + "eval_negation-triplets_steps_per_second": 0.302, + "step": 104 + }, + { + "epoch": 0.20051413881748073, + "eval_scitail-pairs-pos_loss": 0.07263079285621643, + "eval_scitail-pairs-pos_runtime": 2.6213, + "eval_scitail-pairs-pos_samples_per_second": 48.83, + "eval_scitail-pairs-pos_steps_per_second": 0.381, + "step": 104 + }, + { + "epoch": 0.20051413881748073, + "eval_scitail-pairs-qa_loss": 0.048478204756975174, + "eval_scitail-pairs-qa_runtime": 2.202, + "eval_scitail-pairs-qa_samples_per_second": 58.13, + "eval_scitail-pairs-qa_steps_per_second": 0.454, + "step": 104 + }, + { + "epoch": 0.20051413881748073, + "eval_xsum-pairs_loss": 0.14091075956821442, + "eval_xsum-pairs_runtime": 3.0891, + "eval_xsum-pairs_samples_per_second": 41.436, + "eval_xsum-pairs_steps_per_second": 0.324, + "step": 104 + }, + { + "epoch": 0.20051413881748073, + "eval_sciq_pairs_loss": 0.1505264937877655, + "eval_sciq_pairs_runtime": 4.762, + "eval_sciq_pairs_samples_per_second": 26.879, + "eval_sciq_pairs_steps_per_second": 0.21, + "step": 104 + }, + { + "epoch": 0.20051413881748073, + "eval_qasc_pairs_loss": 0.4244489073753357, + "eval_qasc_pairs_runtime": 2.0783, + "eval_qasc_pairs_samples_per_second": 61.589, + "eval_qasc_pairs_steps_per_second": 0.481, + "step": 104 + }, + { + "epoch": 0.20051413881748073, + "eval_openbookqa_pairs_loss": 0.5923758745193481, + "eval_openbookqa_pairs_runtime": 2.1938, + "eval_openbookqa_pairs_samples_per_second": 58.345, + "eval_openbookqa_pairs_steps_per_second": 0.456, + "step": 104 + }, + { + "epoch": 0.20051413881748073, + "eval_msmarco_pairs_loss": 0.5882836580276489, + "eval_msmarco_pairs_runtime": 2.2483, + "eval_msmarco_pairs_samples_per_second": 56.933, + "eval_msmarco_pairs_steps_per_second": 0.445, + "step": 104 + }, + { + "epoch": 0.20051413881748073, + "eval_nq_pairs_loss": 0.5424538254737854, + "eval_nq_pairs_runtime": 3.6424, + "eval_nq_pairs_samples_per_second": 35.142, + "eval_nq_pairs_steps_per_second": 0.275, + "step": 104 + }, + { + "epoch": 0.20051413881748073, + "eval_trivia_pairs_loss": 0.6437886953353882, + "eval_trivia_pairs_runtime": 3.0809, + "eval_trivia_pairs_samples_per_second": 41.546, + "eval_trivia_pairs_steps_per_second": 0.325, + "step": 104 + }, + { + "epoch": 0.20051413881748073, + "eval_gooaq_pairs_loss": 0.22378715872764587, + "eval_gooaq_pairs_runtime": 2.101, + "eval_gooaq_pairs_samples_per_second": 60.924, + "eval_gooaq_pairs_steps_per_second": 0.476, + "step": 104 + }, + { + "epoch": 0.20051413881748073, + "eval_paws-pos_loss": 0.03581737354397774, + "eval_paws-pos_runtime": 2.2447, + "eval_paws-pos_samples_per_second": 57.024, + "eval_paws-pos_steps_per_second": 0.445, + "step": 104 + }, + { + "epoch": 0.20051413881748073, + "eval_global_dataset_loss": 0.3989756405353546, + "eval_global_dataset_runtime": 8.402, + "eval_global_dataset_samples_per_second": 38.681, + "eval_global_dataset_steps_per_second": 0.357, + "step": 104 + }, + { + "epoch": 0.20244215938303342, + "grad_norm": 96.23986053466797, + "learning_rate": 5.5019305019305015e-06, + "loss": 0.5094, + "step": 105 + }, + { + "epoch": 0.20822622107969152, + "grad_norm": 33.88426208496094, + "learning_rate": 5.675675675675676e-06, + "loss": 0.8002, + "step": 108 + }, + { + "epoch": 0.21401028277634962, + "grad_norm": 7.319549560546875, + "learning_rate": 5.849420849420849e-06, + "loss": 0.3886, + "step": 111 + }, + { + "epoch": 0.21979434447300772, + "grad_norm": 8.65904712677002, + "learning_rate": 6.023166023166023e-06, + "loss": 0.6937, + "step": 114 + }, + { + "epoch": 0.22557840616966582, + "grad_norm": 376.61663818359375, + "learning_rate": 6.196911196911197e-06, + "loss": 0.2909, + "step": 117 + }, + { + "epoch": 0.23136246786632392, + "grad_norm": 10.809043884277344, + "learning_rate": 6.370656370656371e-06, + "loss": 0.3885, + "step": 120 + }, + { + "epoch": 0.23714652956298202, + "grad_norm": 8.564398765563965, + "learning_rate": 6.5444015444015445e-06, + "loss": 0.29, + "step": 123 + }, + { + "epoch": 0.24293059125964012, + "grad_norm": 1.5215222835540771, + "learning_rate": 6.718146718146718e-06, + "loss": 0.3485, + "step": 126 + }, + { + "epoch": 0.2487146529562982, + "grad_norm": 181.20809936523438, + "learning_rate": 6.891891891891892e-06, + "loss": 0.3931, + "step": 129 + }, + { + "epoch": 0.2506426735218509, + "eval_Qnli-dev_cosine_accuracy": 0.67578125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.7986456155776978, + "eval_Qnli-dev_cosine_ap": 0.6985736393833883, + "eval_Qnli-dev_cosine_f1": 0.6777777777777777, + "eval_Qnli-dev_cosine_f1_threshold": 0.726032018661499, + "eval_Qnli-dev_cosine_precision": 0.6019736842105263, + "eval_Qnli-dev_cosine_recall": 0.7754237288135594, + "eval_Qnli-dev_dot_accuracy": 0.66796875, + "eval_Qnli-dev_dot_accuracy_threshold": 840.5162353515625, + "eval_Qnli-dev_dot_ap": 0.6656432640480556, + "eval_Qnli-dev_dot_f1": 0.662020905923345, + "eval_Qnli-dev_dot_f1_threshold": 715.1839599609375, + "eval_Qnli-dev_dot_precision": 0.5621301775147929, + "eval_Qnli-dev_dot_recall": 0.8050847457627118, + "eval_Qnli-dev_euclidean_accuracy": 0.677734375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 20.397323608398438, + "eval_Qnli-dev_euclidean_ap": 0.6992667316315335, + "eval_Qnli-dev_euclidean_f1": 0.67296786389414, + "eval_Qnli-dev_euclidean_f1_threshold": 23.44580841064453, + "eval_Qnli-dev_euclidean_precision": 0.6075085324232082, + "eval_Qnli-dev_euclidean_recall": 0.7542372881355932, + "eval_Qnli-dev_manhattan_accuracy": 0.671875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 628.4102172851562, + "eval_Qnli-dev_manhattan_ap": 0.6982133105013253, + "eval_Qnli-dev_manhattan_f1": 0.673992673992674, + "eval_Qnli-dev_manhattan_f1_threshold": 746.4862060546875, + "eval_Qnli-dev_manhattan_precision": 0.5935483870967742, + "eval_Qnli-dev_manhattan_recall": 0.7796610169491526, + "eval_Qnli-dev_max_accuracy": 0.677734375, + "eval_Qnli-dev_max_accuracy_threshold": 840.5162353515625, + "eval_Qnli-dev_max_ap": 0.6992667316315335, + "eval_Qnli-dev_max_f1": 0.6777777777777777, + "eval_Qnli-dev_max_f1_threshold": 746.4862060546875, + "eval_Qnli-dev_max_precision": 0.6075085324232082, + "eval_Qnli-dev_max_recall": 0.8050847457627118, + "eval_allNLI-dev_cosine_accuracy": 0.7109375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8466876745223999, + "eval_allNLI-dev_cosine_ap": 0.5614885136859779, + "eval_allNLI-dev_cosine_f1": 0.5809128630705395, + "eval_allNLI-dev_cosine_f1_threshold": 0.6695160269737244, + "eval_allNLI-dev_cosine_precision": 0.45307443365695793, + "eval_allNLI-dev_cosine_recall": 0.8092485549132948, + "eval_allNLI-dev_dot_accuracy": 0.69921875, + "eval_allNLI-dev_dot_accuracy_threshold": 901.9229736328125, + "eval_allNLI-dev_dot_ap": 0.5637771052921187, + "eval_allNLI-dev_dot_f1": 0.5812807881773399, + "eval_allNLI-dev_dot_f1_threshold": 752.6895141601562, + "eval_allNLI-dev_dot_precision": 0.5064377682403434, + "eval_allNLI-dev_dot_recall": 0.6820809248554913, + "eval_allNLI-dev_euclidean_accuracy": 0.708984375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 17.844379425048828, + "eval_allNLI-dev_euclidean_ap": 0.5616160818796474, + "eval_allNLI-dev_euclidean_f1": 0.5838509316770187, + "eval_allNLI-dev_euclidean_f1_threshold": 26.068946838378906, + "eval_allNLI-dev_euclidean_precision": 0.45483870967741935, + "eval_allNLI-dev_euclidean_recall": 0.815028901734104, + "eval_allNLI-dev_manhattan_accuracy": 0.7109375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 553.34228515625, + "eval_allNLI-dev_manhattan_ap": 0.5624349287984178, + "eval_allNLI-dev_manhattan_f1": 0.5778688524590163, + "eval_allNLI-dev_manhattan_f1_threshold": 816.2254028320312, + "eval_allNLI-dev_manhattan_precision": 0.44761904761904764, + "eval_allNLI-dev_manhattan_recall": 0.815028901734104, + "eval_allNLI-dev_max_accuracy": 0.7109375, + "eval_allNLI-dev_max_accuracy_threshold": 901.9229736328125, + "eval_allNLI-dev_max_ap": 0.5637771052921187, + "eval_allNLI-dev_max_f1": 0.5838509316770187, + "eval_allNLI-dev_max_f1_threshold": 816.2254028320312, + "eval_allNLI-dev_max_precision": 0.5064377682403434, + "eval_allNLI-dev_max_recall": 0.815028901734104, + "eval_sequential_score": 0.6992667316315335, + "eval_sts-test_pearson_cosine": 0.9024404439516205, + "eval_sts-test_pearson_dot": 0.8941554024395801, + "eval_sts-test_pearson_euclidean": 0.9164387414940136, + "eval_sts-test_pearson_manhattan": 0.9158406136217865, + "eval_sts-test_pearson_max": 0.9164387414940136, + "eval_sts-test_spearman_cosine": 0.9081672412582797, + "eval_sts-test_spearman_dot": 0.8855668249766109, + "eval_sts-test_spearman_euclidean": 0.9076990535616029, + "eval_sts-test_spearman_manhattan": 0.907316833038059, + "eval_sts-test_spearman_max": 0.9081672412582797, + "eval_vitaminc-pairs_loss": 3.8795158863067627, + "eval_vitaminc-pairs_runtime": 4.6143, + "eval_vitaminc-pairs_samples_per_second": 27.74, + "eval_vitaminc-pairs_steps_per_second": 0.217, + "step": 130 + }, + { + "epoch": 0.2506426735218509, + "eval_negation-triplets_loss": 1.1134737730026245, + "eval_negation-triplets_runtime": 3.4103, + "eval_negation-triplets_samples_per_second": 37.533, + "eval_negation-triplets_steps_per_second": 0.293, + "step": 130 + }, + { + "epoch": 0.2506426735218509, + "eval_scitail-pairs-pos_loss": 0.045735035091638565, + "eval_scitail-pairs-pos_runtime": 2.7664, + "eval_scitail-pairs-pos_samples_per_second": 46.27, + "eval_scitail-pairs-pos_steps_per_second": 0.361, + "step": 130 + }, + { + "epoch": 0.2506426735218509, + "eval_scitail-pairs-qa_loss": 0.036478620022535324, + "eval_scitail-pairs-qa_runtime": 2.2905, + "eval_scitail-pairs-qa_samples_per_second": 55.884, + "eval_scitail-pairs-qa_steps_per_second": 0.437, + "step": 130 + }, + { + "epoch": 0.2506426735218509, + "eval_xsum-pairs_loss": 0.05133158713579178, + "eval_xsum-pairs_runtime": 3.1323, + "eval_xsum-pairs_samples_per_second": 40.865, + "eval_xsum-pairs_steps_per_second": 0.319, + "step": 130 + }, + { + "epoch": 0.2506426735218509, + "eval_sciq_pairs_loss": 0.12314368784427643, + "eval_sciq_pairs_runtime": 4.8477, + "eval_sciq_pairs_samples_per_second": 26.404, + "eval_sciq_pairs_steps_per_second": 0.206, + "step": 130 + }, + { + "epoch": 0.2506426735218509, + "eval_qasc_pairs_loss": 0.22055241465568542, + "eval_qasc_pairs_runtime": 2.1243, + "eval_qasc_pairs_samples_per_second": 60.255, + "eval_qasc_pairs_steps_per_second": 0.471, + "step": 130 + }, + { + "epoch": 0.2506426735218509, + "eval_openbookqa_pairs_loss": 0.4567987024784088, + "eval_openbookqa_pairs_runtime": 2.3078, + "eval_openbookqa_pairs_samples_per_second": 55.465, + "eval_openbookqa_pairs_steps_per_second": 0.433, + "step": 130 + }, + { + "epoch": 0.2506426735218509, + "eval_msmarco_pairs_loss": 0.21620944142341614, + "eval_msmarco_pairs_runtime": 2.3399, + "eval_msmarco_pairs_samples_per_second": 54.703, + "eval_msmarco_pairs_steps_per_second": 0.427, + "step": 130 + }, + { + "epoch": 0.2506426735218509, + "eval_nq_pairs_loss": 0.26776355504989624, + "eval_nq_pairs_runtime": 3.7038, + "eval_nq_pairs_samples_per_second": 34.559, + "eval_nq_pairs_steps_per_second": 0.27, + "step": 130 + }, + { + "epoch": 0.2506426735218509, + "eval_trivia_pairs_loss": 0.15397313237190247, + "eval_trivia_pairs_runtime": 3.1256, + "eval_trivia_pairs_samples_per_second": 40.952, + "eval_trivia_pairs_steps_per_second": 0.32, + "step": 130 + }, + { + "epoch": 0.2506426735218509, + "eval_gooaq_pairs_loss": 0.12057099491357803, + "eval_gooaq_pairs_runtime": 2.1825, + "eval_gooaq_pairs_samples_per_second": 58.648, + "eval_gooaq_pairs_steps_per_second": 0.458, + "step": 130 + }, + { + "epoch": 0.2506426735218509, + "eval_paws-pos_loss": 0.026105860248208046, + "eval_paws-pos_runtime": 2.2995, + "eval_paws-pos_samples_per_second": 55.665, + "eval_paws-pos_steps_per_second": 0.435, + "step": 130 + }, + { + "epoch": 0.2506426735218509, + "eval_global_dataset_loss": 0.3591004014015198, + "eval_global_dataset_runtime": 8.4153, + "eval_global_dataset_samples_per_second": 38.62, + "eval_global_dataset_steps_per_second": 0.356, + "step": 130 + }, + { + "epoch": 0.2544987146529563, + "grad_norm": 11.167648315429688, + "learning_rate": 7.065637065637066e-06, + "loss": 0.3394, + "step": 132 + }, + { + "epoch": 0.2602827763496144, + "grad_norm": 3.4576871395111084, + "learning_rate": 7.239382239382239e-06, + "loss": 0.1276, + "step": 135 + }, + { + "epoch": 0.2660668380462725, + "grad_norm": 9.594191551208496, + "learning_rate": 7.413127413127414e-06, + "loss": 0.3569, + "step": 138 + }, + { + "epoch": 0.2718508997429306, + "grad_norm": 4.255222320556641, + "learning_rate": 7.586872586872587e-06, + "loss": 0.1231, + "step": 141 + }, + { + "epoch": 0.2776349614395887, + "grad_norm": 6.074405670166016, + "learning_rate": 7.76061776061776e-06, + "loss": 0.3086, + "step": 144 + }, + { + "epoch": 0.2834190231362468, + "grad_norm": 19.442323684692383, + "learning_rate": 7.934362934362936e-06, + "loss": 0.3541, + "step": 147 + }, + { + "epoch": 0.2892030848329049, + "grad_norm": 5.141862392425537, + "learning_rate": 8.108108108108109e-06, + "loss": 0.2597, + "step": 150 + }, + { + "epoch": 0.294987146529563, + "grad_norm": 5.942401885986328, + "learning_rate": 8.281853281853282e-06, + "loss": 0.1585, + "step": 153 + }, + { + "epoch": 0.30077120822622105, + "grad_norm": 97.85575103759766, + "learning_rate": 8.455598455598455e-06, + "loss": 0.1436, + "step": 156 + }, + { + "epoch": 0.30077120822622105, + "eval_Qnli-dev_cosine_accuracy": 0.693359375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.7395063638687134, + "eval_Qnli-dev_cosine_ap": 0.7187155891006661, + "eval_Qnli-dev_cosine_f1": 0.6837606837606838, + "eval_Qnli-dev_cosine_f1_threshold": 0.660994291305542, + "eval_Qnli-dev_cosine_precision": 0.5730659025787965, + "eval_Qnli-dev_cosine_recall": 0.847457627118644, + "eval_Qnli-dev_dot_accuracy": 0.669921875, + "eval_Qnli-dev_dot_accuracy_threshold": 701.7633056640625, + "eval_Qnli-dev_dot_ap": 0.6888124099944863, + "eval_Qnli-dev_dot_f1": 0.6812816188870151, + "eval_Qnli-dev_dot_f1_threshold": 633.31884765625, + "eval_Qnli-dev_dot_precision": 0.5658263305322129, + "eval_Qnli-dev_dot_recall": 0.8559322033898306, + "eval_Qnli-dev_euclidean_accuracy": 0.697265625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 22.954120635986328, + "eval_Qnli-dev_euclidean_ap": 0.721593279891293, + "eval_Qnli-dev_euclidean_f1": 0.6762075134168157, + "eval_Qnli-dev_euclidean_f1_threshold": 24.89653968811035, + "eval_Qnli-dev_euclidean_precision": 0.5851393188854489, + "eval_Qnli-dev_euclidean_recall": 0.8008474576271186, + "eval_Qnli-dev_manhattan_accuracy": 0.6953125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 718.2297973632812, + "eval_Qnli-dev_manhattan_ap": 0.7239405100010182, + "eval_Qnli-dev_manhattan_f1": 0.6807017543859649, + "eval_Qnli-dev_manhattan_f1_threshold": 784.701171875, + "eval_Qnli-dev_manhattan_precision": 0.5808383233532934, + "eval_Qnli-dev_manhattan_recall": 0.8220338983050848, + "eval_Qnli-dev_max_accuracy": 0.697265625, + "eval_Qnli-dev_max_accuracy_threshold": 718.2297973632812, + "eval_Qnli-dev_max_ap": 0.7239405100010182, + "eval_Qnli-dev_max_f1": 0.6837606837606838, + "eval_Qnli-dev_max_f1_threshold": 784.701171875, + "eval_Qnli-dev_max_precision": 0.5851393188854489, + "eval_Qnli-dev_max_recall": 0.8559322033898306, + "eval_allNLI-dev_cosine_accuracy": 0.71875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8460206985473633, + "eval_allNLI-dev_cosine_ap": 0.5796621470925665, + "eval_allNLI-dev_cosine_f1": 0.6139954853273137, + "eval_allNLI-dev_cosine_f1_threshold": 0.7097446322441101, + "eval_allNLI-dev_cosine_precision": 0.5037037037037037, + "eval_allNLI-dev_cosine_recall": 0.7861271676300579, + "eval_allNLI-dev_dot_accuracy": 0.7109375, + "eval_allNLI-dev_dot_accuracy_threshold": 837.8741455078125, + "eval_allNLI-dev_dot_ap": 0.5763352242724786, + "eval_allNLI-dev_dot_f1": 0.6179245283018867, + "eval_allNLI-dev_dot_f1_threshold": 712.9136352539062, + "eval_allNLI-dev_dot_precision": 0.5219123505976095, + "eval_allNLI-dev_dot_recall": 0.7572254335260116, + "eval_allNLI-dev_euclidean_accuracy": 0.724609375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 18.82543182373047, + "eval_allNLI-dev_euclidean_ap": 0.5803015452539461, + "eval_allNLI-dev_euclidean_f1": 0.6085011185682326, + "eval_allNLI-dev_euclidean_f1_threshold": 24.0869197845459, + "eval_allNLI-dev_euclidean_precision": 0.49635036496350365, + "eval_allNLI-dev_euclidean_recall": 0.7861271676300579, + "eval_allNLI-dev_manhattan_accuracy": 0.720703125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 586.8082275390625, + "eval_allNLI-dev_manhattan_ap": 0.5806697040885319, + "eval_allNLI-dev_manhattan_f1": 0.6035242290748899, + "eval_allNLI-dev_manhattan_f1_threshold": 756.8704833984375, + "eval_allNLI-dev_manhattan_precision": 0.4875444839857651, + "eval_allNLI-dev_manhattan_recall": 0.791907514450867, + "eval_allNLI-dev_max_accuracy": 0.724609375, + "eval_allNLI-dev_max_accuracy_threshold": 837.8741455078125, + "eval_allNLI-dev_max_ap": 0.5806697040885319, + "eval_allNLI-dev_max_f1": 0.6179245283018867, + "eval_allNLI-dev_max_f1_threshold": 756.8704833984375, + "eval_allNLI-dev_max_precision": 0.5219123505976095, + "eval_allNLI-dev_max_recall": 0.791907514450867, + "eval_sequential_score": 0.7239405100010182, + "eval_sts-test_pearson_cosine": 0.9086022204177822, + "eval_sts-test_pearson_dot": 0.8963707820565703, + "eval_sts-test_pearson_euclidean": 0.9226263845809013, + "eval_sts-test_pearson_manhattan": 0.9216624653503607, + "eval_sts-test_pearson_max": 0.9226263845809013, + "eval_sts-test_spearman_cosine": 0.9176886091344368, + "eval_sts-test_spearman_dot": 0.8941249219513181, + "eval_sts-test_spearman_euclidean": 0.9176283739366038, + "eval_sts-test_spearman_manhattan": 0.9167815458172944, + "eval_sts-test_spearman_max": 0.9176886091344368, + "eval_vitaminc-pairs_loss": 3.8346316814422607, + "eval_vitaminc-pairs_runtime": 4.4216, + "eval_vitaminc-pairs_samples_per_second": 28.949, + "eval_vitaminc-pairs_steps_per_second": 0.226, + "step": 156 + }, + { + "epoch": 0.30077120822622105, + "eval_negation-triplets_loss": 0.9441184997558594, + "eval_negation-triplets_runtime": 3.3236, + "eval_negation-triplets_samples_per_second": 38.512, + "eval_negation-triplets_steps_per_second": 0.301, + "step": 156 + }, + { + "epoch": 0.30077120822622105, + "eval_scitail-pairs-pos_loss": 0.03759249672293663, + "eval_scitail-pairs-pos_runtime": 2.6436, + "eval_scitail-pairs-pos_samples_per_second": 48.418, + "eval_scitail-pairs-pos_steps_per_second": 0.378, + "step": 156 + }, + { + "epoch": 0.30077120822622105, + "eval_scitail-pairs-qa_loss": 0.04104204475879669, + "eval_scitail-pairs-qa_runtime": 2.205, + "eval_scitail-pairs-qa_samples_per_second": 58.05, + "eval_scitail-pairs-qa_steps_per_second": 0.454, + "step": 156 + }, + { + "epoch": 0.30077120822622105, + "eval_xsum-pairs_loss": 0.05032652989029884, + "eval_xsum-pairs_runtime": 3.0619, + "eval_xsum-pairs_samples_per_second": 41.805, + "eval_xsum-pairs_steps_per_second": 0.327, + "step": 156 + }, + { + "epoch": 0.30077120822622105, + "eval_sciq_pairs_loss": 0.08709195256233215, + "eval_sciq_pairs_runtime": 4.8405, + "eval_sciq_pairs_samples_per_second": 26.444, + "eval_sciq_pairs_steps_per_second": 0.207, + "step": 156 + }, + { + "epoch": 0.30077120822622105, + "eval_qasc_pairs_loss": 0.17168982326984406, + "eval_qasc_pairs_runtime": 2.078, + "eval_qasc_pairs_samples_per_second": 61.599, + "eval_qasc_pairs_steps_per_second": 0.481, + "step": 156 + }, + { + "epoch": 0.30077120822622105, + "eval_openbookqa_pairs_loss": 0.4304167926311493, + "eval_openbookqa_pairs_runtime": 2.1991, + "eval_openbookqa_pairs_samples_per_second": 58.205, + "eval_openbookqa_pairs_steps_per_second": 0.455, + "step": 156 + }, + { + "epoch": 0.30077120822622105, + "eval_msmarco_pairs_loss": 0.13501481711864471, + "eval_msmarco_pairs_runtime": 2.2186, + "eval_msmarco_pairs_samples_per_second": 57.694, + "eval_msmarco_pairs_steps_per_second": 0.451, + "step": 156 + }, + { + "epoch": 0.30077120822622105, + "eval_nq_pairs_loss": 0.20035718381404877, + "eval_nq_pairs_runtime": 3.6367, + "eval_nq_pairs_samples_per_second": 35.197, + "eval_nq_pairs_steps_per_second": 0.275, + "step": 156 + }, + { + "epoch": 0.30077120822622105, + "eval_trivia_pairs_loss": 0.08228898048400879, + "eval_trivia_pairs_runtime": 3.086, + "eval_trivia_pairs_samples_per_second": 41.477, + "eval_trivia_pairs_steps_per_second": 0.324, + "step": 156 + }, + { + "epoch": 0.30077120822622105, + "eval_gooaq_pairs_loss": 0.0767039954662323, + "eval_gooaq_pairs_runtime": 2.1123, + "eval_gooaq_pairs_samples_per_second": 60.599, + "eval_gooaq_pairs_steps_per_second": 0.473, + "step": 156 + }, + { + "epoch": 0.30077120822622105, + "eval_paws-pos_loss": 0.02134825475513935, + "eval_paws-pos_runtime": 2.2234, + "eval_paws-pos_samples_per_second": 57.571, + "eval_paws-pos_steps_per_second": 0.45, + "step": 156 + }, + { + "epoch": 0.30077120822622105, + "eval_global_dataset_loss": 0.2958550453186035, + "eval_global_dataset_runtime": 8.3542, + "eval_global_dataset_samples_per_second": 38.903, + "eval_global_dataset_steps_per_second": 0.359, + "step": 156 + }, + { + "epoch": 0.3065552699228792, + "grad_norm": 47.52979278564453, + "learning_rate": 8.62934362934363e-06, + "loss": 0.1941, + "step": 159 + }, + { + "epoch": 0.31233933161953725, + "grad_norm": 1234.25927734375, + "learning_rate": 8.803088803088804e-06, + "loss": 0.3041, + "step": 162 + }, + { + "epoch": 0.3181233933161954, + "grad_norm": 185.80238342285156, + "learning_rate": 8.976833976833977e-06, + "loss": 0.2358, + "step": 165 + }, + { + "epoch": 0.32390745501285345, + "grad_norm": 7.233102798461914, + "learning_rate": 9.15057915057915e-06, + "loss": 0.2148, + "step": 168 + }, + { + "epoch": 0.3296915167095116, + "grad_norm": 342.77685546875, + "learning_rate": 9.324324324324325e-06, + "loss": 0.8567, + "step": 171 + }, + { + "epoch": 0.33547557840616965, + "grad_norm": 181.07650756835938, + "learning_rate": 9.498069498069498e-06, + "loss": 0.3668, + "step": 174 + }, + { + "epoch": 0.3412596401028278, + "grad_norm": 198.4124755859375, + "learning_rate": 9.671814671814673e-06, + "loss": 0.3278, + "step": 177 + }, + { + "epoch": 0.34704370179948585, + "grad_norm": 91.80472564697266, + "learning_rate": 9.845559845559846e-06, + "loss": 0.474, + "step": 180 + }, + { + "epoch": 0.3508997429305913, + "eval_Qnli-dev_cosine_accuracy": 0.6953125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.7320406436920166, + "eval_Qnli-dev_cosine_ap": 0.7163901550352688, + "eval_Qnli-dev_cosine_f1": 0.6857142857142857, + "eval_Qnli-dev_cosine_f1_threshold": 0.6526604890823364, + "eval_Qnli-dev_cosine_precision": 0.5682451253481894, + "eval_Qnli-dev_cosine_recall": 0.864406779661017, + "eval_Qnli-dev_dot_accuracy": 0.67578125, + "eval_Qnli-dev_dot_accuracy_threshold": 759.5008544921875, + "eval_Qnli-dev_dot_ap": 0.6857309318375049, + "eval_Qnli-dev_dot_f1": 0.6857142857142857, + "eval_Qnli-dev_dot_f1_threshold": 653.0701904296875, + "eval_Qnli-dev_dot_precision": 0.5682451253481894, + "eval_Qnli-dev_dot_recall": 0.864406779661017, + "eval_Qnli-dev_euclidean_accuracy": 0.689453125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 23.348251342773438, + "eval_Qnli-dev_euclidean_ap": 0.7186338110228241, + "eval_Qnli-dev_euclidean_f1": 0.6819672131147542, + "eval_Qnli-dev_euclidean_f1_threshold": 26.874767303466797, + "eval_Qnli-dev_euclidean_precision": 0.5561497326203209, + "eval_Qnli-dev_euclidean_recall": 0.8813559322033898, + "eval_Qnli-dev_manhattan_accuracy": 0.6875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 695.7578125, + "eval_Qnli-dev_manhattan_ap": 0.7187237394606566, + "eval_Qnli-dev_manhattan_f1": 0.6815286624203821, + "eval_Qnli-dev_manhattan_f1_threshold": 856.7666015625, + "eval_Qnli-dev_manhattan_precision": 0.5459183673469388, + "eval_Qnli-dev_manhattan_recall": 0.9067796610169492, + "eval_Qnli-dev_max_accuracy": 0.6953125, + "eval_Qnli-dev_max_accuracy_threshold": 759.5008544921875, + "eval_Qnli-dev_max_ap": 0.7187237394606566, + "eval_Qnli-dev_max_f1": 0.6857142857142857, + "eval_Qnli-dev_max_f1_threshold": 856.7666015625, + "eval_Qnli-dev_max_precision": 0.5682451253481894, + "eval_Qnli-dev_max_recall": 0.9067796610169492, + "eval_allNLI-dev_cosine_accuracy": 0.705078125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8531240224838257, + "eval_allNLI-dev_cosine_ap": 0.5555367525973895, + "eval_allNLI-dev_cosine_f1": 0.5972222222222222, + "eval_allNLI-dev_cosine_f1_threshold": 0.7072098255157471, + "eval_allNLI-dev_cosine_precision": 0.4980694980694981, + "eval_allNLI-dev_cosine_recall": 0.7456647398843931, + "eval_allNLI-dev_dot_accuracy": 0.705078125, + "eval_allNLI-dev_dot_accuracy_threshold": 858.6700439453125, + "eval_allNLI-dev_dot_ap": 0.5501313597611072, + "eval_allNLI-dev_dot_f1": 0.5992217898832685, + "eval_allNLI-dev_dot_f1_threshold": 657.9625244140625, + "eval_allNLI-dev_dot_precision": 0.45161290322580644, + "eval_allNLI-dev_dot_recall": 0.8901734104046243, + "eval_allNLI-dev_euclidean_accuracy": 0.705078125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 16.42839813232422, + "eval_allNLI-dev_euclidean_ap": 0.5519081828104085, + "eval_allNLI-dev_euclidean_f1": 0.5919282511210763, + "eval_allNLI-dev_euclidean_f1_threshold": 24.696884155273438, + "eval_allNLI-dev_euclidean_precision": 0.4835164835164835, + "eval_allNLI-dev_euclidean_recall": 0.7630057803468208, + "eval_allNLI-dev_manhattan_accuracy": 0.705078125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 517.3240356445312, + "eval_allNLI-dev_manhattan_ap": 0.5527648330520126, + "eval_allNLI-dev_manhattan_f1": 0.5900900900900902, + "eval_allNLI-dev_manhattan_f1_threshold": 765.703857421875, + "eval_allNLI-dev_manhattan_precision": 0.4833948339483395, + "eval_allNLI-dev_manhattan_recall": 0.7572254335260116, + "eval_allNLI-dev_max_accuracy": 0.705078125, + "eval_allNLI-dev_max_accuracy_threshold": 858.6700439453125, + "eval_allNLI-dev_max_ap": 0.5555367525973895, + "eval_allNLI-dev_max_f1": 0.5992217898832685, + "eval_allNLI-dev_max_f1_threshold": 765.703857421875, + "eval_allNLI-dev_max_precision": 0.4980694980694981, + "eval_allNLI-dev_max_recall": 0.8901734104046243, + "eval_sequential_score": 0.7187237394606566, + "eval_sts-test_pearson_cosine": 0.8999110842263669, + "eval_sts-test_pearson_dot": 0.8830335713257453, + "eval_sts-test_pearson_euclidean": 0.9123289195839425, + "eval_sts-test_pearson_manhattan": 0.913131760355957, + "eval_sts-test_pearson_max": 0.913131760355957, + "eval_sts-test_spearman_cosine": 0.9102556931248936, + "eval_sts-test_spearman_dot": 0.8782251567890449, + "eval_sts-test_spearman_euclidean": 0.9078544746924153, + "eval_sts-test_spearman_manhattan": 0.9084604962741777, + "eval_sts-test_spearman_max": 0.9102556931248936, + "eval_vitaminc-pairs_loss": 3.8443844318389893, + "eval_vitaminc-pairs_runtime": 4.62, + "eval_vitaminc-pairs_samples_per_second": 27.706, + "eval_vitaminc-pairs_steps_per_second": 0.216, + "step": 182 + }, + { + "epoch": 0.3508997429305913, + "eval_negation-triplets_loss": 0.8706199526786804, + "eval_negation-triplets_runtime": 3.4865, + "eval_negation-triplets_samples_per_second": 36.714, + "eval_negation-triplets_steps_per_second": 0.287, + "step": 182 + }, + { + "epoch": 0.3508997429305913, + "eval_scitail-pairs-pos_loss": 0.05496061593294144, + "eval_scitail-pairs-pos_runtime": 2.8238, + "eval_scitail-pairs-pos_samples_per_second": 45.328, + "eval_scitail-pairs-pos_steps_per_second": 0.354, + "step": 182 + }, + { + "epoch": 0.3508997429305913, + "eval_scitail-pairs-qa_loss": 0.12511999905109406, + "eval_scitail-pairs-qa_runtime": 2.2972, + "eval_scitail-pairs-qa_samples_per_second": 55.721, + "eval_scitail-pairs-qa_steps_per_second": 0.435, + "step": 182 + }, + { + "epoch": 0.3508997429305913, + "eval_xsum-pairs_loss": 0.03821578621864319, + "eval_xsum-pairs_runtime": 3.1015, + "eval_xsum-pairs_samples_per_second": 41.27, + "eval_xsum-pairs_steps_per_second": 0.322, + "step": 182 + }, + { + "epoch": 0.3508997429305913, + "eval_sciq_pairs_loss": 0.08246304094791412, + "eval_sciq_pairs_runtime": 4.8955, + "eval_sciq_pairs_samples_per_second": 26.147, + "eval_sciq_pairs_steps_per_second": 0.204, + "step": 182 + }, + { + "epoch": 0.3508997429305913, + "eval_qasc_pairs_loss": 0.15105530619621277, + "eval_qasc_pairs_runtime": 2.2079, + "eval_qasc_pairs_samples_per_second": 57.972, + "eval_qasc_pairs_steps_per_second": 0.453, + "step": 182 + }, + { + "epoch": 0.3508997429305913, + "eval_openbookqa_pairs_loss": 0.7893790602684021, + "eval_openbookqa_pairs_runtime": 2.2579, + "eval_openbookqa_pairs_samples_per_second": 56.69, + "eval_openbookqa_pairs_steps_per_second": 0.443, + "step": 182 + }, + { + "epoch": 0.3508997429305913, + "eval_msmarco_pairs_loss": 0.41756388545036316, + "eval_msmarco_pairs_runtime": 2.2908, + "eval_msmarco_pairs_samples_per_second": 55.877, + "eval_msmarco_pairs_steps_per_second": 0.437, + "step": 182 + }, + { + "epoch": 0.3508997429305913, + "eval_nq_pairs_loss": 0.2524656355381012, + "eval_nq_pairs_runtime": 3.6631, + "eval_nq_pairs_samples_per_second": 34.943, + "eval_nq_pairs_steps_per_second": 0.273, + "step": 182 + }, + { + "epoch": 0.3508997429305913, + "eval_trivia_pairs_loss": 0.062444861978292465, + "eval_trivia_pairs_runtime": 3.115, + "eval_trivia_pairs_samples_per_second": 41.091, + "eval_trivia_pairs_steps_per_second": 0.321, + "step": 182 + }, + { + "epoch": 0.3508997429305913, + "eval_gooaq_pairs_loss": 0.12118146568536758, + "eval_gooaq_pairs_runtime": 2.236, + "eval_gooaq_pairs_samples_per_second": 57.245, + "eval_gooaq_pairs_steps_per_second": 0.447, + "step": 182 + }, + { + "epoch": 0.3508997429305913, + "eval_paws-pos_loss": 0.020900897681713104, + "eval_paws-pos_runtime": 2.2981, + "eval_paws-pos_samples_per_second": 55.698, + "eval_paws-pos_steps_per_second": 0.435, + "step": 182 + }, + { + "epoch": 0.3508997429305913, + "eval_global_dataset_loss": 0.38996464014053345, + "eval_global_dataset_runtime": 8.423, + "eval_global_dataset_samples_per_second": 38.585, + "eval_global_dataset_steps_per_second": 0.356, + "step": 182 + }, + { + "epoch": 0.352827763496144, + "grad_norm": 412.62738037109375, + "learning_rate": 1.0019305019305019e-05, + "loss": 0.5365, + "step": 183 + }, + { + "epoch": 0.35861182519280205, + "grad_norm": 693.9551391601562, + "learning_rate": 1.0193050193050193e-05, + "loss": 0.6902, + "step": 186 + }, + { + "epoch": 0.3643958868894602, + "grad_norm": 739.5757446289062, + "learning_rate": 1.0366795366795366e-05, + "loss": 0.4105, + "step": 189 + }, + { + "epoch": 0.37017994858611825, + "grad_norm": 174.94882202148438, + "learning_rate": 1.0540540540540541e-05, + "loss": 0.2434, + "step": 192 + }, + { + "epoch": 0.3759640102827764, + "grad_norm": 5.913711071014404, + "learning_rate": 1.0714285714285714e-05, + "loss": 0.1521, + "step": 195 + }, + { + "epoch": 0.38174807197943444, + "grad_norm": 34.247581481933594, + "learning_rate": 1.0888030888030889e-05, + "loss": 0.1878, + "step": 198 + }, + { + "epoch": 0.38753213367609257, + "grad_norm": 50.26247787475586, + "learning_rate": 1.1061776061776062e-05, + "loss": 0.3544, + "step": 201 + }, + { + "epoch": 0.39331619537275064, + "grad_norm": 227.15286254882812, + "learning_rate": 1.1235521235521235e-05, + "loss": 0.1397, + "step": 204 + }, + { + "epoch": 0.3991002570694087, + "grad_norm": 95.1203384399414, + "learning_rate": 1.140926640926641e-05, + "loss": 0.2982, + "step": 207 + }, + { + "epoch": 0.40102827763496146, + "eval_Qnli-dev_cosine_accuracy": 0.685546875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.715149462223053, + "eval_Qnli-dev_cosine_ap": 0.7217325275419479, + "eval_Qnli-dev_cosine_f1": 0.6854130052724077, + "eval_Qnli-dev_cosine_f1_threshold": 0.658507227897644, + "eval_Qnli-dev_cosine_precision": 0.5855855855855856, + "eval_Qnli-dev_cosine_recall": 0.826271186440678, + "eval_Qnli-dev_dot_accuracy": 0.669921875, + "eval_Qnli-dev_dot_accuracy_threshold": 717.3741455078125, + "eval_Qnli-dev_dot_ap": 0.6857348319823914, + "eval_Qnli-dev_dot_f1": 0.6757679180887372, + "eval_Qnli-dev_dot_f1_threshold": 627.569580078125, + "eval_Qnli-dev_dot_precision": 0.5657142857142857, + "eval_Qnli-dev_dot_recall": 0.8389830508474576, + "eval_Qnli-dev_euclidean_accuracy": 0.68359375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 23.550647735595703, + "eval_Qnli-dev_euclidean_ap": 0.7266186529674616, + "eval_Qnli-dev_euclidean_f1": 0.6959578207381371, + "eval_Qnli-dev_euclidean_f1_threshold": 25.634620666503906, + "eval_Qnli-dev_euclidean_precision": 0.5945945945945946, + "eval_Qnli-dev_euclidean_recall": 0.8389830508474576, + "eval_Qnli-dev_manhattan_accuracy": 0.6796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 681.6762084960938, + "eval_Qnli-dev_manhattan_ap": 0.7264245248641615, + "eval_Qnli-dev_manhattan_f1": 0.6938053097345133, + "eval_Qnli-dev_manhattan_f1_threshold": 793.83837890625, + "eval_Qnli-dev_manhattan_precision": 0.5957446808510638, + "eval_Qnli-dev_manhattan_recall": 0.8305084745762712, + "eval_Qnli-dev_max_accuracy": 0.685546875, + "eval_Qnli-dev_max_accuracy_threshold": 717.3741455078125, + "eval_Qnli-dev_max_ap": 0.7266186529674616, + "eval_Qnli-dev_max_f1": 0.6959578207381371, + "eval_Qnli-dev_max_f1_threshold": 793.83837890625, + "eval_Qnli-dev_max_precision": 0.5957446808510638, + "eval_Qnli-dev_max_recall": 0.8389830508474576, + "eval_allNLI-dev_cosine_accuracy": 0.720703125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8279609680175781, + "eval_allNLI-dev_cosine_ap": 0.5853174176053627, + "eval_allNLI-dev_cosine_f1": 0.6189376443418013, + "eval_allNLI-dev_cosine_f1_threshold": 0.7112143039703369, + "eval_allNLI-dev_cosine_precision": 0.5153846153846153, + "eval_allNLI-dev_cosine_recall": 0.7745664739884393, + "eval_allNLI-dev_dot_accuracy": 0.708984375, + "eval_allNLI-dev_dot_accuracy_threshold": 844.9638061523438, + "eval_allNLI-dev_dot_ap": 0.5767480495013521, + "eval_allNLI-dev_dot_f1": 0.620525059665871, + "eval_allNLI-dev_dot_f1_threshold": 698.743408203125, + "eval_allNLI-dev_dot_precision": 0.5284552845528455, + "eval_allNLI-dev_dot_recall": 0.7514450867052023, + "eval_allNLI-dev_euclidean_accuracy": 0.71875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 17.55453109741211, + "eval_allNLI-dev_euclidean_ap": 0.5835774601767848, + "eval_allNLI-dev_euclidean_f1": 0.6063569682151589, + "eval_allNLI-dev_euclidean_f1_threshold": 22.958927154541016, + "eval_allNLI-dev_euclidean_precision": 0.5254237288135594, + "eval_allNLI-dev_euclidean_recall": 0.7167630057803468, + "eval_allNLI-dev_manhattan_accuracy": 0.71875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 546.237060546875, + "eval_allNLI-dev_manhattan_ap": 0.5834282573106326, + "eval_allNLI-dev_manhattan_f1": 0.6093023255813954, + "eval_allNLI-dev_manhattan_f1_threshold": 738.0977172851562, + "eval_allNLI-dev_manhattan_precision": 0.5097276264591439, + "eval_allNLI-dev_manhattan_recall": 0.7572254335260116, + "eval_allNLI-dev_max_accuracy": 0.720703125, + "eval_allNLI-dev_max_accuracy_threshold": 844.9638061523438, + "eval_allNLI-dev_max_ap": 0.5853174176053627, + "eval_allNLI-dev_max_f1": 0.620525059665871, + "eval_allNLI-dev_max_f1_threshold": 738.0977172851562, + "eval_allNLI-dev_max_precision": 0.5284552845528455, + "eval_allNLI-dev_max_recall": 0.7745664739884393, + "eval_sequential_score": 0.7266186529674616, + "eval_sts-test_pearson_cosine": 0.9110444991658042, + "eval_sts-test_pearson_dot": 0.9027007693117498, + "eval_sts-test_pearson_euclidean": 0.9255983263878029, + "eval_sts-test_pearson_manhattan": 0.925028098980626, + "eval_sts-test_pearson_max": 0.9255983263878029, + "eval_sts-test_spearman_cosine": 0.9211899030744385, + "eval_sts-test_spearman_dot": 0.9034949858127107, + "eval_sts-test_spearman_euclidean": 0.9209135640353466, + "eval_sts-test_spearman_manhattan": 0.9206623304538951, + "eval_sts-test_spearman_max": 0.9211899030744385, + "eval_vitaminc-pairs_loss": 3.210468292236328, + "eval_vitaminc-pairs_runtime": 4.4941, + "eval_vitaminc-pairs_samples_per_second": 28.482, + "eval_vitaminc-pairs_steps_per_second": 0.223, + "step": 208 + }, + { + "epoch": 0.40102827763496146, + "eval_negation-triplets_loss": 0.6793166995048523, + "eval_negation-triplets_runtime": 3.2613, + "eval_negation-triplets_samples_per_second": 39.249, + "eval_negation-triplets_steps_per_second": 0.307, + "step": 208 + }, + { + "epoch": 0.40102827763496146, + "eval_scitail-pairs-pos_loss": 0.020910607650876045, + "eval_scitail-pairs-pos_runtime": 2.6259, + "eval_scitail-pairs-pos_samples_per_second": 48.745, + "eval_scitail-pairs-pos_steps_per_second": 0.381, + "step": 208 + }, + { + "epoch": 0.40102827763496146, + "eval_scitail-pairs-qa_loss": 0.024287551641464233, + "eval_scitail-pairs-qa_runtime": 2.2326, + "eval_scitail-pairs-qa_samples_per_second": 57.331, + "eval_scitail-pairs-qa_steps_per_second": 0.448, + "step": 208 + }, + { + "epoch": 0.40102827763496146, + "eval_xsum-pairs_loss": 0.028975818306207657, + "eval_xsum-pairs_runtime": 3.0795, + "eval_xsum-pairs_samples_per_second": 41.565, + "eval_xsum-pairs_steps_per_second": 0.325, + "step": 208 + }, + { + "epoch": 0.40102827763496146, + "eval_sciq_pairs_loss": 0.08505217730998993, + "eval_sciq_pairs_runtime": 4.7361, + "eval_sciq_pairs_samples_per_second": 27.026, + "eval_sciq_pairs_steps_per_second": 0.211, + "step": 208 + }, + { + "epoch": 0.40102827763496146, + "eval_qasc_pairs_loss": 0.1334802210330963, + "eval_qasc_pairs_runtime": 2.0711, + "eval_qasc_pairs_samples_per_second": 61.804, + "eval_qasc_pairs_steps_per_second": 0.483, + "step": 208 + }, + { + "epoch": 0.40102827763496146, + "eval_openbookqa_pairs_loss": 0.4361065924167633, + "eval_openbookqa_pairs_runtime": 2.2402, + "eval_openbookqa_pairs_samples_per_second": 57.137, + "eval_openbookqa_pairs_steps_per_second": 0.446, + "step": 208 + }, + { + "epoch": 0.40102827763496146, + "eval_msmarco_pairs_loss": 0.09543590992689133, + "eval_msmarco_pairs_runtime": 2.2223, + "eval_msmarco_pairs_samples_per_second": 57.599, + "eval_msmarco_pairs_steps_per_second": 0.45, + "step": 208 + }, + { + "epoch": 0.40102827763496146, + "eval_nq_pairs_loss": 0.24286797642707825, + "eval_nq_pairs_runtime": 3.617, + "eval_nq_pairs_samples_per_second": 35.389, + "eval_nq_pairs_steps_per_second": 0.276, + "step": 208 + }, + { + "epoch": 0.40102827763496146, + "eval_trivia_pairs_loss": 0.053676724433898926, + "eval_trivia_pairs_runtime": 3.0737, + "eval_trivia_pairs_samples_per_second": 41.644, + "eval_trivia_pairs_steps_per_second": 0.325, + "step": 208 + }, + { + "epoch": 0.40102827763496146, + "eval_gooaq_pairs_loss": 0.07278460264205933, + "eval_gooaq_pairs_runtime": 2.1378, + "eval_gooaq_pairs_samples_per_second": 59.874, + "eval_gooaq_pairs_steps_per_second": 0.468, + "step": 208 + }, + { + "epoch": 0.40102827763496146, + "eval_paws-pos_loss": 0.022397885099053383, + "eval_paws-pos_runtime": 2.253, + "eval_paws-pos_samples_per_second": 56.812, + "eval_paws-pos_steps_per_second": 0.444, + "step": 208 + }, + { + "epoch": 0.40102827763496146, + "eval_global_dataset_loss": 0.250370055437088, + "eval_global_dataset_runtime": 8.321, + "eval_global_dataset_samples_per_second": 39.058, + "eval_global_dataset_steps_per_second": 0.361, + "step": 208 + }, + { + "epoch": 0.40488431876606684, + "grad_norm": 10.200668334960938, + "learning_rate": 1.1583011583011584e-05, + "loss": 0.1875, + "step": 210 + }, + { + "epoch": 0.4106683804627249, + "grad_norm": 25.712194442749023, + "learning_rate": 1.1756756756756757e-05, + "loss": 0.169, + "step": 213 + }, + { + "epoch": 0.41645244215938304, + "grad_norm": 48.711666107177734, + "learning_rate": 1.193050193050193e-05, + "loss": 0.2341, + "step": 216 + }, + { + "epoch": 0.4222365038560411, + "grad_norm": 13.946932792663574, + "learning_rate": 1.2104247104247105e-05, + "loss": 0.1806, + "step": 219 + }, + { + "epoch": 0.42802056555269924, + "grad_norm": 6.094577312469482, + "learning_rate": 1.227799227799228e-05, + "loss": 0.2736, + "step": 222 + }, + { + "epoch": 0.4338046272493573, + "grad_norm": 6.224375247955322, + "learning_rate": 1.2451737451737452e-05, + "loss": 0.1772, + "step": 225 + }, + { + "epoch": 0.43958868894601544, + "grad_norm": 34.21040725708008, + "learning_rate": 1.2625482625482625e-05, + "loss": 0.131, + "step": 228 + }, + { + "epoch": 0.4453727506426735, + "grad_norm": 3.1434619426727295, + "learning_rate": 1.2799227799227798e-05, + "loss": 0.0825, + "step": 231 + }, + { + "epoch": 0.45115681233933164, + "grad_norm": 11.084990501403809, + "learning_rate": 1.2972972972972975e-05, + "loss": 0.2745, + "step": 234 + }, + { + "epoch": 0.45115681233933164, + "eval_Qnli-dev_cosine_accuracy": 0.681640625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.7570502758026123, + "eval_Qnli-dev_cosine_ap": 0.7209954522759622, + "eval_Qnli-dev_cosine_f1": 0.6952054794520549, + "eval_Qnli-dev_cosine_f1_threshold": 0.6451430320739746, + "eval_Qnli-dev_cosine_precision": 0.5833333333333334, + "eval_Qnli-dev_cosine_recall": 0.8601694915254238, + "eval_Qnli-dev_dot_accuracy": 0.689453125, + "eval_Qnli-dev_dot_accuracy_threshold": 690.3031616210938, + "eval_Qnli-dev_dot_ap": 0.6892678092033461, + "eval_Qnli-dev_dot_f1": 0.6910994764397904, + "eval_Qnli-dev_dot_f1_threshold": 616.900390625, + "eval_Qnli-dev_dot_precision": 0.5875370919881305, + "eval_Qnli-dev_dot_recall": 0.8389830508474576, + "eval_Qnli-dev_euclidean_accuracy": 0.685546875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 22.101879119873047, + "eval_Qnli-dev_euclidean_ap": 0.725163804769685, + "eval_Qnli-dev_euclidean_f1": 0.6919275123558485, + "eval_Qnli-dev_euclidean_f1_threshold": 26.458560943603516, + "eval_Qnli-dev_euclidean_precision": 0.5660377358490566, + "eval_Qnli-dev_euclidean_recall": 0.8898305084745762, + "eval_Qnli-dev_manhattan_accuracy": 0.685546875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 683.6396484375, + "eval_Qnli-dev_manhattan_ap": 0.7242560148825838, + "eval_Qnli-dev_manhattan_f1": 0.6919275123558485, + "eval_Qnli-dev_manhattan_f1_threshold": 830.1368408203125, + "eval_Qnli-dev_manhattan_precision": 0.5660377358490566, + "eval_Qnli-dev_manhattan_recall": 0.8898305084745762, + "eval_Qnli-dev_max_accuracy": 0.689453125, + "eval_Qnli-dev_max_accuracy_threshold": 690.3031616210938, + "eval_Qnli-dev_max_ap": 0.725163804769685, + "eval_Qnli-dev_max_f1": 0.6952054794520549, + "eval_Qnli-dev_max_f1_threshold": 830.1368408203125, + "eval_Qnli-dev_max_precision": 0.5875370919881305, + "eval_Qnli-dev_max_recall": 0.8898305084745762, + "eval_allNLI-dev_cosine_accuracy": 0.72265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8315098285675049, + "eval_allNLI-dev_cosine_ap": 0.579111348491065, + "eval_allNLI-dev_cosine_f1": 0.6045454545454545, + "eval_allNLI-dev_cosine_f1_threshold": 0.7200944423675537, + "eval_allNLI-dev_cosine_precision": 0.49812734082397003, + "eval_allNLI-dev_cosine_recall": 0.7687861271676301, + "eval_allNLI-dev_dot_accuracy": 0.7109375, + "eval_allNLI-dev_dot_accuracy_threshold": 813.0138549804688, + "eval_allNLI-dev_dot_ap": 0.5580493245872203, + "eval_allNLI-dev_dot_f1": 0.6124401913875598, + "eval_allNLI-dev_dot_f1_threshold": 703.9876708984375, + "eval_allNLI-dev_dot_precision": 0.5224489795918368, + "eval_allNLI-dev_dot_recall": 0.7398843930635838, + "eval_allNLI-dev_euclidean_accuracy": 0.732421875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 18.200069427490234, + "eval_allNLI-dev_euclidean_ap": 0.579946296621263, + "eval_allNLI-dev_euclidean_f1": 0.6084905660377359, + "eval_allNLI-dev_euclidean_f1_threshold": 22.654571533203125, + "eval_allNLI-dev_euclidean_precision": 0.5139442231075697, + "eval_allNLI-dev_euclidean_recall": 0.7456647398843931, + "eval_allNLI-dev_manhattan_accuracy": 0.728515625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 571.3721313476562, + "eval_allNLI-dev_manhattan_ap": 0.5810042621969582, + "eval_allNLI-dev_manhattan_f1": 0.6150234741784038, + "eval_allNLI-dev_manhattan_f1_threshold": 708.9692993164062, + "eval_allNLI-dev_manhattan_precision": 0.5177865612648221, + "eval_allNLI-dev_manhattan_recall": 0.7572254335260116, + "eval_allNLI-dev_max_accuracy": 0.732421875, + "eval_allNLI-dev_max_accuracy_threshold": 813.0138549804688, + "eval_allNLI-dev_max_ap": 0.5810042621969582, + "eval_allNLI-dev_max_f1": 0.6150234741784038, + "eval_allNLI-dev_max_f1_threshold": 708.9692993164062, + "eval_allNLI-dev_max_precision": 0.5224489795918368, + "eval_allNLI-dev_max_recall": 0.7687861271676301, + "eval_sequential_score": 0.725163804769685, + "eval_sts-test_pearson_cosine": 0.9079355833912655, + "eval_sts-test_pearson_dot": 0.8900251358598403, + "eval_sts-test_pearson_euclidean": 0.9252346568054488, + "eval_sts-test_pearson_manhattan": 0.9248533505767098, + "eval_sts-test_pearson_max": 0.9252346568054488, + "eval_sts-test_spearman_cosine": 0.91858412662614, + "eval_sts-test_spearman_dot": 0.8870906501782496, + "eval_sts-test_spearman_euclidean": 0.919076345750184, + "eval_sts-test_spearman_manhattan": 0.9185557095751564, + "eval_sts-test_spearman_max": 0.919076345750184, + "eval_vitaminc-pairs_loss": 3.4478557109832764, + "eval_vitaminc-pairs_runtime": 4.5288, + "eval_vitaminc-pairs_samples_per_second": 28.264, + "eval_vitaminc-pairs_steps_per_second": 0.221, + "step": 234 + }, + { + "epoch": 0.45115681233933164, + "eval_negation-triplets_loss": 0.636644184589386, + "eval_negation-triplets_runtime": 3.402, + "eval_negation-triplets_samples_per_second": 37.625, + "eval_negation-triplets_steps_per_second": 0.294, + "step": 234 + }, + { + "epoch": 0.45115681233933164, + "eval_scitail-pairs-pos_loss": 0.026369448751211166, + "eval_scitail-pairs-pos_runtime": 2.7594, + "eval_scitail-pairs-pos_samples_per_second": 46.386, + "eval_scitail-pairs-pos_steps_per_second": 0.362, + "step": 234 + }, + { + "epoch": 0.45115681233933164, + "eval_scitail-pairs-qa_loss": 0.023621411994099617, + "eval_scitail-pairs-qa_runtime": 2.2993, + "eval_scitail-pairs-qa_samples_per_second": 55.669, + "eval_scitail-pairs-qa_steps_per_second": 0.435, + "step": 234 + }, + { + "epoch": 0.45115681233933164, + "eval_xsum-pairs_loss": 0.03148656710982323, + "eval_xsum-pairs_runtime": 3.1312, + "eval_xsum-pairs_samples_per_second": 40.878, + "eval_xsum-pairs_steps_per_second": 0.319, + "step": 234 + }, + { + "epoch": 0.45115681233933164, + "eval_sciq_pairs_loss": 0.08541633188724518, + "eval_sciq_pairs_runtime": 4.8654, + "eval_sciq_pairs_samples_per_second": 26.308, + "eval_sciq_pairs_steps_per_second": 0.206, + "step": 234 + }, + { + "epoch": 0.45115681233933164, + "eval_qasc_pairs_loss": 0.12370316684246063, + "eval_qasc_pairs_runtime": 2.2058, + "eval_qasc_pairs_samples_per_second": 58.029, + "eval_qasc_pairs_steps_per_second": 0.453, + "step": 234 + }, + { + "epoch": 0.45115681233933164, + "eval_openbookqa_pairs_loss": 0.41401803493499756, + "eval_openbookqa_pairs_runtime": 2.338, + "eval_openbookqa_pairs_samples_per_second": 54.747, + "eval_openbookqa_pairs_steps_per_second": 0.428, + "step": 234 + }, + { + "epoch": 0.45115681233933164, + "eval_msmarco_pairs_loss": 0.031952742487192154, + "eval_msmarco_pairs_runtime": 2.3775, + "eval_msmarco_pairs_samples_per_second": 53.839, + "eval_msmarco_pairs_steps_per_second": 0.421, + "step": 234 + }, + { + "epoch": 0.45115681233933164, + "eval_nq_pairs_loss": 0.20804765820503235, + "eval_nq_pairs_runtime": 3.7039, + "eval_nq_pairs_samples_per_second": 34.558, + "eval_nq_pairs_steps_per_second": 0.27, + "step": 234 + }, + { + "epoch": 0.45115681233933164, + "eval_trivia_pairs_loss": 0.053435422480106354, + "eval_trivia_pairs_runtime": 3.1705, + "eval_trivia_pairs_samples_per_second": 40.372, + "eval_trivia_pairs_steps_per_second": 0.315, + "step": 234 + }, + { + "epoch": 0.45115681233933164, + "eval_gooaq_pairs_loss": 0.06229045242071152, + "eval_gooaq_pairs_runtime": 2.2703, + "eval_gooaq_pairs_samples_per_second": 56.38, + "eval_gooaq_pairs_steps_per_second": 0.44, + "step": 234 + }, + { + "epoch": 0.45115681233933164, + "eval_paws-pos_loss": 0.018832631409168243, + "eval_paws-pos_runtime": 2.3285, + "eval_paws-pos_samples_per_second": 54.972, + "eval_paws-pos_steps_per_second": 0.429, + "step": 234 + }, + { + "epoch": 0.45115681233933164, + "eval_global_dataset_loss": 0.24586248397827148, + "eval_global_dataset_runtime": 8.4553, + "eval_global_dataset_samples_per_second": 38.438, + "eval_global_dataset_steps_per_second": 0.355, + "step": 234 + }, + { + "epoch": 0.4569408740359897, + "grad_norm": 12.329869270324707, + "learning_rate": 1.3146718146718148e-05, + "loss": 0.131, + "step": 237 + }, + { + "epoch": 0.46272493573264784, + "grad_norm": 19.341089248657227, + "learning_rate": 1.332046332046332e-05, + "loss": 0.2012, + "step": 240 + }, + { + "epoch": 0.4685089974293059, + "grad_norm": 2.5328197479248047, + "learning_rate": 1.3494208494208494e-05, + "loss": 0.0855, + "step": 243 + }, + { + "epoch": 0.47429305912596403, + "grad_norm": 7.81564474105835, + "learning_rate": 1.3667953667953668e-05, + "loss": 0.1181, + "step": 246 + }, + { + "epoch": 0.4800771208226221, + "grad_norm": 4.904431343078613, + "learning_rate": 1.3841698841698843e-05, + "loss": 0.0992, + "step": 249 + }, + { + "epoch": 0.48586118251928023, + "grad_norm": 2.585756540298462, + "learning_rate": 1.4015444015444016e-05, + "loss": 0.0375, + "step": 252 + }, + { + "epoch": 0.4916452442159383, + "grad_norm": 67.9809799194336, + "learning_rate": 1.4189189189189189e-05, + "loss": 0.1503, + "step": 255 + }, + { + "epoch": 0.4974293059125964, + "grad_norm": 8.545217514038086, + "learning_rate": 1.4362934362934362e-05, + "loss": 0.1239, + "step": 258 + }, + { + "epoch": 0.5012853470437018, + "eval_Qnli-dev_cosine_accuracy": 0.69140625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6921107769012451, + "eval_Qnli-dev_cosine_ap": 0.7392224932862383, + "eval_Qnli-dev_cosine_f1": 0.6944444444444444, + "eval_Qnli-dev_cosine_f1_threshold": 0.6230607032775879, + "eval_Qnli-dev_cosine_precision": 0.5882352941176471, + "eval_Qnli-dev_cosine_recall": 0.847457627118644, + "eval_Qnli-dev_dot_accuracy": 0.6875, + "eval_Qnli-dev_dot_accuracy_threshold": 697.8347778320312, + "eval_Qnli-dev_dot_ap": 0.7145385015601986, + "eval_Qnli-dev_dot_f1": 0.6952054794520549, + "eval_Qnli-dev_dot_f1_threshold": 581.857666015625, + "eval_Qnli-dev_dot_precision": 0.5833333333333334, + "eval_Qnli-dev_dot_recall": 0.8601694915254238, + "eval_Qnli-dev_euclidean_accuracy": 0.689453125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 23.611530303955078, + "eval_Qnli-dev_euclidean_ap": 0.7374932215461754, + "eval_Qnli-dev_euclidean_f1": 0.697594501718213, + "eval_Qnli-dev_euclidean_f1_threshold": 26.889122009277344, + "eval_Qnli-dev_euclidean_precision": 0.5867052023121387, + "eval_Qnli-dev_euclidean_recall": 0.8601694915254238, + "eval_Qnli-dev_manhattan_accuracy": 0.6875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 731.7583618164062, + "eval_Qnli-dev_manhattan_ap": 0.737286744524734, + "eval_Qnli-dev_manhattan_f1": 0.6998223801065719, + "eval_Qnli-dev_manhattan_f1_threshold": 821.986328125, + "eval_Qnli-dev_manhattan_precision": 0.6024464831804281, + "eval_Qnli-dev_manhattan_recall": 0.8347457627118644, + "eval_Qnli-dev_max_accuracy": 0.69140625, + "eval_Qnli-dev_max_accuracy_threshold": 731.7583618164062, + "eval_Qnli-dev_max_ap": 0.7392224932862383, + "eval_Qnli-dev_max_f1": 0.6998223801065719, + "eval_Qnli-dev_max_f1_threshold": 821.986328125, + "eval_Qnli-dev_max_precision": 0.6024464831804281, + "eval_Qnli-dev_max_recall": 0.8601694915254238, + "eval_allNLI-dev_cosine_accuracy": 0.716796875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.842121958732605, + "eval_allNLI-dev_cosine_ap": 0.5873681355801845, + "eval_allNLI-dev_cosine_f1": 0.6113744075829384, + "eval_allNLI-dev_cosine_f1_threshold": 0.7030021548271179, + "eval_allNLI-dev_cosine_precision": 0.5180722891566265, + "eval_allNLI-dev_cosine_recall": 0.7456647398843931, + "eval_allNLI-dev_dot_accuracy": 0.720703125, + "eval_allNLI-dev_dot_accuracy_threshold": 750.5220947265625, + "eval_allNLI-dev_dot_ap": 0.5799311198634003, + "eval_allNLI-dev_dot_f1": 0.6201923076923077, + "eval_allNLI-dev_dot_f1_threshold": 695.0645751953125, + "eval_allNLI-dev_dot_precision": 0.5308641975308642, + "eval_allNLI-dev_dot_recall": 0.7456647398843931, + "eval_allNLI-dev_euclidean_accuracy": 0.71875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 17.74507713317871, + "eval_allNLI-dev_euclidean_ap": 0.5836049221301078, + "eval_allNLI-dev_euclidean_f1": 0.6175771971496437, + "eval_allNLI-dev_euclidean_f1_threshold": 23.809024810791016, + "eval_allNLI-dev_euclidean_precision": 0.5241935483870968, + "eval_allNLI-dev_euclidean_recall": 0.7514450867052023, + "eval_allNLI-dev_manhattan_accuracy": 0.716796875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 556.4411010742188, + "eval_allNLI-dev_manhattan_ap": 0.5840998034727366, + "eval_allNLI-dev_manhattan_f1": 0.6146572104018913, + "eval_allNLI-dev_manhattan_f1_threshold": 746.3275756835938, + "eval_allNLI-dev_manhattan_precision": 0.52, + "eval_allNLI-dev_manhattan_recall": 0.7514450867052023, + "eval_allNLI-dev_max_accuracy": 0.720703125, + "eval_allNLI-dev_max_accuracy_threshold": 750.5220947265625, + "eval_allNLI-dev_max_ap": 0.5873681355801845, + "eval_allNLI-dev_max_f1": 0.6201923076923077, + "eval_allNLI-dev_max_f1_threshold": 746.3275756835938, + "eval_allNLI-dev_max_precision": 0.5308641975308642, + "eval_allNLI-dev_max_recall": 0.7514450867052023, + "eval_sequential_score": 0.7392224932862383, + "eval_sts-test_pearson_cosine": 0.9106833903275314, + "eval_sts-test_pearson_dot": 0.8992284856445683, + "eval_sts-test_pearson_euclidean": 0.9250592315631505, + "eval_sts-test_pearson_manhattan": 0.9246839540884049, + "eval_sts-test_pearson_max": 0.9250592315631505, + "eval_sts-test_spearman_cosine": 0.920613820086547, + "eval_sts-test_spearman_dot": 0.8966107650190875, + "eval_sts-test_spearman_euclidean": 0.9195501721057961, + "eval_sts-test_spearman_manhattan": 0.9193704510400483, + "eval_sts-test_spearman_max": 0.920613820086547, + "eval_vitaminc-pairs_loss": 2.7884156703948975, + "eval_vitaminc-pairs_runtime": 4.4544, + "eval_vitaminc-pairs_samples_per_second": 28.736, + "eval_vitaminc-pairs_steps_per_second": 0.224, + "step": 260 + }, + { + "epoch": 0.5012853470437018, + "eval_negation-triplets_loss": 0.6253538727760315, + "eval_negation-triplets_runtime": 3.2937, + "eval_negation-triplets_samples_per_second": 38.862, + "eval_negation-triplets_steps_per_second": 0.304, + "step": 260 + }, + { + "epoch": 0.5012853470437018, + "eval_scitail-pairs-pos_loss": 0.01729702576994896, + "eval_scitail-pairs-pos_runtime": 2.6618, + "eval_scitail-pairs-pos_samples_per_second": 48.088, + "eval_scitail-pairs-pos_steps_per_second": 0.376, + "step": 260 + }, + { + "epoch": 0.5012853470437018, + "eval_scitail-pairs-qa_loss": 0.025592364370822906, + "eval_scitail-pairs-qa_runtime": 2.204, + "eval_scitail-pairs-qa_samples_per_second": 58.075, + "eval_scitail-pairs-qa_steps_per_second": 0.454, + "step": 260 + }, + { + "epoch": 0.5012853470437018, + "eval_xsum-pairs_loss": 0.012989089824259281, + "eval_xsum-pairs_runtime": 3.0884, + "eval_xsum-pairs_samples_per_second": 41.445, + "eval_xsum-pairs_steps_per_second": 0.324, + "step": 260 + }, + { + "epoch": 0.5012853470437018, + "eval_sciq_pairs_loss": 0.0916769877076149, + "eval_sciq_pairs_runtime": 4.7492, + "eval_sciq_pairs_samples_per_second": 26.952, + "eval_sciq_pairs_steps_per_second": 0.211, + "step": 260 + }, + { + "epoch": 0.5012853470437018, + "eval_qasc_pairs_loss": 0.10503728687763214, + "eval_qasc_pairs_runtime": 2.1169, + "eval_qasc_pairs_samples_per_second": 60.466, + "eval_qasc_pairs_steps_per_second": 0.472, + "step": 260 + }, + { + "epoch": 0.5012853470437018, + "eval_openbookqa_pairs_loss": 0.439107745885849, + "eval_openbookqa_pairs_runtime": 2.1928, + "eval_openbookqa_pairs_samples_per_second": 58.372, + "eval_openbookqa_pairs_steps_per_second": 0.456, + "step": 260 + }, + { + "epoch": 0.5012853470437018, + "eval_msmarco_pairs_loss": 0.4504755735397339, + "eval_msmarco_pairs_runtime": 2.2269, + "eval_msmarco_pairs_samples_per_second": 57.48, + "eval_msmarco_pairs_steps_per_second": 0.449, + "step": 260 + }, + { + "epoch": 0.5012853470437018, + "eval_nq_pairs_loss": 0.14197981357574463, + "eval_nq_pairs_runtime": 3.633, + "eval_nq_pairs_samples_per_second": 35.232, + "eval_nq_pairs_steps_per_second": 0.275, + "step": 260 + }, + { + "epoch": 0.5012853470437018, + "eval_trivia_pairs_loss": 0.05544201284646988, + "eval_trivia_pairs_runtime": 3.0694, + "eval_trivia_pairs_samples_per_second": 41.701, + "eval_trivia_pairs_steps_per_second": 0.326, + "step": 260 + }, + { + "epoch": 0.5012853470437018, + "eval_gooaq_pairs_loss": 0.0432414636015892, + "eval_gooaq_pairs_runtime": 2.096, + "eval_gooaq_pairs_samples_per_second": 61.069, + "eval_gooaq_pairs_steps_per_second": 0.477, + "step": 260 + }, + { + "epoch": 0.5012853470437018, + "eval_paws-pos_loss": 0.02002798579633236, + "eval_paws-pos_runtime": 2.2043, + "eval_paws-pos_samples_per_second": 58.068, + "eval_paws-pos_steps_per_second": 0.454, + "step": 260 + }, + { + "epoch": 0.5012853470437018, + "eval_global_dataset_loss": 0.2467704862356186, + "eval_global_dataset_runtime": 8.3491, + "eval_global_dataset_samples_per_second": 38.926, + "eval_global_dataset_steps_per_second": 0.359, + "step": 260 + }, + { + "epoch": 0.5032133676092545, + "grad_norm": 7.293272495269775, + "learning_rate": 1.4536679536679537e-05, + "loss": 0.1849, + "step": 261 + }, + { + "epoch": 0.5089974293059126, + "grad_norm": 4.900815010070801, + "learning_rate": 1.4710424710424711e-05, + "loss": 0.0453, + "step": 264 + }, + { + "epoch": 0.5147814910025706, + "grad_norm": 130.37289428710938, + "learning_rate": 1.4884169884169884e-05, + "loss": 0.118, + "step": 267 + }, + { + "epoch": 0.5205655526992288, + "grad_norm": 139.6493377685547, + "learning_rate": 1.4999944826454462e-05, + "loss": 0.2285, + "step": 270 + }, + { + "epoch": 0.5263496143958869, + "grad_norm": 8.455353736877441, + "learning_rate": 1.4999117241310476e-05, + "loss": 0.072, + "step": 273 + }, + { + "epoch": 0.532133676092545, + "grad_norm": 37.64395523071289, + "learning_rate": 1.4997296673048672e-05, + "loss": 0.1449, + "step": 276 + }, + { + "epoch": 0.537917737789203, + "grad_norm": 35.890071868896484, + "learning_rate": 1.4994483389525345e-05, + "loss": 0.2179, + "step": 279 + }, + { + "epoch": 0.5437017994858612, + "grad_norm": 4.571140289306641, + "learning_rate": 1.4990677804652856e-05, + "loss": 0.0687, + "step": 282 + }, + { + "epoch": 0.5494858611825193, + "grad_norm": 4.596847057342529, + "learning_rate": 1.4985880478338728e-05, + "loss": 0.071, + "step": 285 + }, + { + "epoch": 0.5514138817480719, + "eval_Qnli-dev_cosine_accuracy": 0.6953125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.7168486714363098, + "eval_Qnli-dev_cosine_ap": 0.7438559407478242, + "eval_Qnli-dev_cosine_f1": 0.7035175879396984, + "eval_Qnli-dev_cosine_f1_threshold": 0.6049751043319702, + "eval_Qnli-dev_cosine_precision": 0.5817174515235457, + "eval_Qnli-dev_cosine_recall": 0.8898305084745762, + "eval_Qnli-dev_dot_accuracy": 0.689453125, + "eval_Qnli-dev_dot_accuracy_threshold": 723.0029296875, + "eval_Qnli-dev_dot_ap": 0.715173938792371, + "eval_Qnli-dev_dot_f1": 0.6942148760330578, + "eval_Qnli-dev_dot_f1_threshold": 577.0636596679688, + "eval_Qnli-dev_dot_precision": 0.5691056910569106, + "eval_Qnli-dev_dot_recall": 0.8898305084745762, + "eval_Qnli-dev_euclidean_accuracy": 0.693359375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 23.278547286987305, + "eval_Qnli-dev_euclidean_ap": 0.7463211078255994, + "eval_Qnli-dev_euclidean_f1": 0.6970227670753064, + "eval_Qnli-dev_euclidean_f1_threshold": 27.014936447143555, + "eval_Qnli-dev_euclidean_precision": 0.5940298507462687, + "eval_Qnli-dev_euclidean_recall": 0.8432203389830508, + "eval_Qnli-dev_manhattan_accuracy": 0.693359375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 722.95458984375, + "eval_Qnli-dev_manhattan_ap": 0.7459762632488858, + "eval_Qnli-dev_manhattan_f1": 0.7010309278350516, + "eval_Qnli-dev_manhattan_f1_threshold": 851.98779296875, + "eval_Qnli-dev_manhattan_precision": 0.5895953757225434, + "eval_Qnli-dev_manhattan_recall": 0.864406779661017, + "eval_Qnli-dev_max_accuracy": 0.6953125, + "eval_Qnli-dev_max_accuracy_threshold": 723.0029296875, + "eval_Qnli-dev_max_ap": 0.7463211078255994, + "eval_Qnli-dev_max_f1": 0.7035175879396984, + "eval_Qnli-dev_max_f1_threshold": 851.98779296875, + "eval_Qnli-dev_max_precision": 0.5940298507462687, + "eval_Qnli-dev_max_recall": 0.8898305084745762, + "eval_allNLI-dev_cosine_accuracy": 0.716796875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8314377069473267, + "eval_allNLI-dev_cosine_ap": 0.5914689901349341, + "eval_allNLI-dev_cosine_f1": 0.6323529411764707, + "eval_allNLI-dev_cosine_f1_threshold": 0.7238098978996277, + "eval_allNLI-dev_cosine_precision": 0.548936170212766, + "eval_allNLI-dev_cosine_recall": 0.7456647398843931, + "eval_allNLI-dev_dot_accuracy": 0.71484375, + "eval_allNLI-dev_dot_accuracy_threshold": 810.9595947265625, + "eval_allNLI-dev_dot_ap": 0.5839807817494149, + "eval_allNLI-dev_dot_f1": 0.6192468619246863, + "eval_allNLI-dev_dot_f1_threshold": 654.8411254882812, + "eval_allNLI-dev_dot_precision": 0.4852459016393443, + "eval_allNLI-dev_dot_recall": 0.8554913294797688, + "eval_allNLI-dev_euclidean_accuracy": 0.720703125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 18.951290130615234, + "eval_allNLI-dev_euclidean_ap": 0.5919673869334235, + "eval_allNLI-dev_euclidean_f1": 0.6300715990453462, + "eval_allNLI-dev_euclidean_f1_threshold": 23.576871871948242, + "eval_allNLI-dev_euclidean_precision": 0.5365853658536586, + "eval_allNLI-dev_euclidean_recall": 0.7630057803468208, + "eval_allNLI-dev_manhattan_accuracy": 0.72265625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 587.756103515625, + "eval_allNLI-dev_manhattan_ap": 0.5937706599658953, + "eval_allNLI-dev_manhattan_f1": 0.631578947368421, + "eval_allNLI-dev_manhattan_f1_threshold": 734.0830078125, + "eval_allNLI-dev_manhattan_precision": 0.5387755102040817, + "eval_allNLI-dev_manhattan_recall": 0.7630057803468208, + "eval_allNLI-dev_max_accuracy": 0.72265625, + "eval_allNLI-dev_max_accuracy_threshold": 810.9595947265625, + "eval_allNLI-dev_max_ap": 0.5937706599658953, + "eval_allNLI-dev_max_f1": 0.6323529411764707, + "eval_allNLI-dev_max_f1_threshold": 734.0830078125, + "eval_allNLI-dev_max_precision": 0.548936170212766, + "eval_allNLI-dev_max_recall": 0.8554913294797688, + "eval_sequential_score": 0.7463211078255994, + "eval_sts-test_pearson_cosine": 0.9045390674341153, + "eval_sts-test_pearson_dot": 0.8941766888408116, + "eval_sts-test_pearson_euclidean": 0.9222411970618043, + "eval_sts-test_pearson_manhattan": 0.9215732932835548, + "eval_sts-test_pearson_max": 0.9222411970618043, + "eval_sts-test_spearman_cosine": 0.9179395294570584, + "eval_sts-test_spearman_dot": 0.8987898829412804, + "eval_sts-test_spearman_euclidean": 0.9189799962844084, + "eval_sts-test_spearman_manhattan": 0.9181845426131761, + "eval_sts-test_spearman_max": 0.9189799962844084, + "eval_vitaminc-pairs_loss": 3.9507529735565186, + "eval_vitaminc-pairs_runtime": 4.5266, + "eval_vitaminc-pairs_samples_per_second": 28.277, + "eval_vitaminc-pairs_steps_per_second": 0.221, + "step": 286 + }, + { + "epoch": 0.5514138817480719, + "eval_negation-triplets_loss": 0.6451127529144287, + "eval_negation-triplets_runtime": 3.3941, + "eval_negation-triplets_samples_per_second": 37.713, + "eval_negation-triplets_steps_per_second": 0.295, + "step": 286 + }, + { + "epoch": 0.5514138817480719, + "eval_scitail-pairs-pos_loss": 0.018346838653087616, + "eval_scitail-pairs-pos_runtime": 2.7523, + "eval_scitail-pairs-pos_samples_per_second": 46.506, + "eval_scitail-pairs-pos_steps_per_second": 0.363, + "step": 286 + }, + { + "epoch": 0.5514138817480719, + "eval_scitail-pairs-qa_loss": 0.021733948960900307, + "eval_scitail-pairs-qa_runtime": 2.3512, + "eval_scitail-pairs-qa_samples_per_second": 54.439, + "eval_scitail-pairs-qa_steps_per_second": 0.425, + "step": 286 + }, + { + "epoch": 0.5514138817480719, + "eval_xsum-pairs_loss": 0.006104973144829273, + "eval_xsum-pairs_runtime": 3.1114, + "eval_xsum-pairs_samples_per_second": 41.139, + "eval_xsum-pairs_steps_per_second": 0.321, + "step": 286 + }, + { + "epoch": 0.5514138817480719, + "eval_sciq_pairs_loss": 0.09672096371650696, + "eval_sciq_pairs_runtime": 4.7885, + "eval_sciq_pairs_samples_per_second": 26.731, + "eval_sciq_pairs_steps_per_second": 0.209, + "step": 286 + }, + { + "epoch": 0.5514138817480719, + "eval_qasc_pairs_loss": 0.12122269719839096, + "eval_qasc_pairs_runtime": 2.2255, + "eval_qasc_pairs_samples_per_second": 57.515, + "eval_qasc_pairs_steps_per_second": 0.449, + "step": 286 + }, + { + "epoch": 0.5514138817480719, + "eval_openbookqa_pairs_loss": 0.4551953971385956, + "eval_openbookqa_pairs_runtime": 2.3841, + "eval_openbookqa_pairs_samples_per_second": 53.688, + "eval_openbookqa_pairs_steps_per_second": 0.419, + "step": 286 + }, + { + "epoch": 0.5514138817480719, + "eval_msmarco_pairs_loss": 0.029557496309280396, + "eval_msmarco_pairs_runtime": 2.2846, + "eval_msmarco_pairs_samples_per_second": 56.027, + "eval_msmarco_pairs_steps_per_second": 0.438, + "step": 286 + }, + { + "epoch": 0.5514138817480719, + "eval_nq_pairs_loss": 0.16617023944854736, + "eval_nq_pairs_runtime": 3.6481, + "eval_nq_pairs_samples_per_second": 35.087, + "eval_nq_pairs_steps_per_second": 0.274, + "step": 286 + }, + { + "epoch": 0.5514138817480719, + "eval_trivia_pairs_loss": 0.06793519854545593, + "eval_trivia_pairs_runtime": 3.1576, + "eval_trivia_pairs_samples_per_second": 40.537, + "eval_trivia_pairs_steps_per_second": 0.317, + "step": 286 + }, + { + "epoch": 0.5514138817480719, + "eval_gooaq_pairs_loss": 0.05080220475792885, + "eval_gooaq_pairs_runtime": 2.2483, + "eval_gooaq_pairs_samples_per_second": 56.931, + "eval_gooaq_pairs_steps_per_second": 0.445, + "step": 286 + }, + { + "epoch": 0.5514138817480719, + "eval_paws-pos_loss": 0.02359372191131115, + "eval_paws-pos_runtime": 2.2829, + "eval_paws-pos_samples_per_second": 56.07, + "eval_paws-pos_steps_per_second": 0.438, + "step": 286 + }, + { + "epoch": 0.5514138817480719, + "eval_global_dataset_loss": 0.2985983192920685, + "eval_global_dataset_runtime": 8.4308, + "eval_global_dataset_samples_per_second": 38.549, + "eval_global_dataset_steps_per_second": 0.356, + "step": 286 + }, + { + "epoch": 0.5552699228791774, + "grad_norm": 5.148947238922119, + "learning_rate": 1.4980092116403276e-05, + "loss": 0.0855, + "step": 288 + }, + { + "epoch": 0.5610539845758354, + "grad_norm": 13.268424034118652, + "learning_rate": 1.4973313570475753e-05, + "loss": 0.103, + "step": 291 + }, + { + "epoch": 0.5668380462724936, + "grad_norm": 6.761672496795654, + "learning_rate": 1.4965545837869058e-05, + "loss": 0.1555, + "step": 294 + }, + { + "epoch": 0.5726221079691517, + "grad_norm": 3.2632031440734863, + "learning_rate": 1.4956790061432996e-05, + "loss": 0.1069, + "step": 297 + }, + { + "epoch": 0.5784061696658098, + "grad_norm": 7.57591438293457, + "learning_rate": 1.494704752938614e-05, + "loss": 0.2014, + "step": 300 + }, + { + "epoch": 0.5841902313624678, + "grad_norm": 3.9399232864379883, + "learning_rate": 1.49363196751263e-05, + "loss": 0.1028, + "step": 303 + }, + { + "epoch": 0.589974293059126, + "grad_norm": 6.177419662475586, + "learning_rate": 1.4924608077019618e-05, + "loss": 0.2425, + "step": 306 + }, + { + "epoch": 0.5957583547557841, + "grad_norm": 3.4256973266601562, + "learning_rate": 1.4911914458168361e-05, + "loss": 0.1639, + "step": 309 + }, + { + "epoch": 0.6015424164524421, + "grad_norm": 3.3635151386260986, + "learning_rate": 1.4898240686157396e-05, + "loss": 0.099, + "step": 312 + }, + { + "epoch": 0.6015424164524421, + "eval_Qnli-dev_cosine_accuracy": 0.69921875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.7296463251113892, + "eval_Qnli-dev_cosine_ap": 0.7499128775998081, + "eval_Qnli-dev_cosine_f1": 0.700525394045534, + "eval_Qnli-dev_cosine_f1_threshold": 0.6390261650085449, + "eval_Qnli-dev_cosine_precision": 0.5970149253731343, + "eval_Qnli-dev_cosine_recall": 0.847457627118644, + "eval_Qnli-dev_dot_accuracy": 0.685546875, + "eval_Qnli-dev_dot_accuracy_threshold": 729.7808837890625, + "eval_Qnli-dev_dot_ap": 0.7267806921631229, + "eval_Qnli-dev_dot_f1": 0.6901408450704225, + "eval_Qnli-dev_dot_f1_threshold": 627.7941284179688, + "eval_Qnli-dev_dot_precision": 0.5903614457831325, + "eval_Qnli-dev_dot_recall": 0.8305084745762712, + "eval_Qnli-dev_euclidean_accuracy": 0.703125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 23.62594223022461, + "eval_Qnli-dev_euclidean_ap": 0.7513960328625653, + "eval_Qnli-dev_euclidean_f1": 0.6930693069306931, + "eval_Qnli-dev_euclidean_f1_threshold": 27.77108383178711, + "eval_Qnli-dev_euclidean_precision": 0.5675675675675675, + "eval_Qnli-dev_euclidean_recall": 0.8898305084745762, + "eval_Qnli-dev_manhattan_accuracy": 0.703125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 719.592529296875, + "eval_Qnli-dev_manhattan_ap": 0.7518755537519403, + "eval_Qnli-dev_manhattan_f1": 0.6963696369636965, + "eval_Qnli-dev_manhattan_f1_threshold": 864.5909423828125, + "eval_Qnli-dev_manhattan_precision": 0.5702702702702702, + "eval_Qnli-dev_manhattan_recall": 0.8940677966101694, + "eval_Qnli-dev_max_accuracy": 0.703125, + "eval_Qnli-dev_max_accuracy_threshold": 729.7808837890625, + "eval_Qnli-dev_max_ap": 0.7518755537519403, + "eval_Qnli-dev_max_f1": 0.700525394045534, + "eval_Qnli-dev_max_f1_threshold": 864.5909423828125, + "eval_Qnli-dev_max_precision": 0.5970149253731343, + "eval_Qnli-dev_max_recall": 0.8940677966101694, + "eval_allNLI-dev_cosine_accuracy": 0.716796875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.744063138961792, + "eval_allNLI-dev_cosine_ap": 0.5872524450594222, + "eval_allNLI-dev_cosine_f1": 0.6287128712871288, + "eval_allNLI-dev_cosine_f1_threshold": 0.7192724943161011, + "eval_allNLI-dev_cosine_precision": 0.5497835497835498, + "eval_allNLI-dev_cosine_recall": 0.7341040462427746, + "eval_allNLI-dev_dot_accuracy": 0.716796875, + "eval_allNLI-dev_dot_accuracy_threshold": 740.0152587890625, + "eval_allNLI-dev_dot_ap": 0.5749750816372319, + "eval_allNLI-dev_dot_f1": 0.6201923076923077, + "eval_allNLI-dev_dot_f1_threshold": 690.6895751953125, + "eval_allNLI-dev_dot_precision": 0.5308641975308642, + "eval_allNLI-dev_dot_recall": 0.7456647398843931, + "eval_allNLI-dev_euclidean_accuracy": 0.71484375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 22.877256393432617, + "eval_allNLI-dev_euclidean_ap": 0.587161106665968, + "eval_allNLI-dev_euclidean_f1": 0.6275510204081632, + "eval_allNLI-dev_euclidean_f1_threshold": 22.877256393432617, + "eval_allNLI-dev_euclidean_precision": 0.5616438356164384, + "eval_allNLI-dev_euclidean_recall": 0.7109826589595376, + "eval_allNLI-dev_manhattan_accuracy": 0.71484375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 710.319091796875, + "eval_allNLI-dev_manhattan_ap": 0.588942704782977, + "eval_allNLI-dev_manhattan_f1": 0.6294416243654821, + "eval_allNLI-dev_manhattan_f1_threshold": 713.067138671875, + "eval_allNLI-dev_manhattan_precision": 0.5610859728506787, + "eval_allNLI-dev_manhattan_recall": 0.7167630057803468, + "eval_allNLI-dev_max_accuracy": 0.716796875, + "eval_allNLI-dev_max_accuracy_threshold": 740.0152587890625, + "eval_allNLI-dev_max_ap": 0.588942704782977, + "eval_allNLI-dev_max_f1": 0.6294416243654821, + "eval_allNLI-dev_max_f1_threshold": 713.067138671875, + "eval_allNLI-dev_max_precision": 0.5616438356164384, + "eval_allNLI-dev_max_recall": 0.7456647398843931, + "eval_sequential_score": 0.7518755537519403, + "eval_sts-test_pearson_cosine": 0.9043388760494557, + "eval_sts-test_pearson_dot": 0.8886027908855041, + "eval_sts-test_pearson_euclidean": 0.9241059779670555, + "eval_sts-test_pearson_manhattan": 0.9234022512717295, + "eval_sts-test_pearson_max": 0.9241059779670555, + "eval_sts-test_spearman_cosine": 0.919076121993877, + "eval_sts-test_spearman_dot": 0.8892642189440298, + "eval_sts-test_spearman_euclidean": 0.9203842013643476, + "eval_sts-test_spearman_manhattan": 0.9199692676687267, + "eval_sts-test_spearman_max": 0.9203842013643476, + "eval_vitaminc-pairs_loss": 3.3625028133392334, + "eval_vitaminc-pairs_runtime": 4.4127, + "eval_vitaminc-pairs_samples_per_second": 29.007, + "eval_vitaminc-pairs_steps_per_second": 0.227, + "step": 312 + }, + { + "epoch": 0.6015424164524421, + "eval_negation-triplets_loss": 0.5403304696083069, + "eval_negation-triplets_runtime": 3.3216, + "eval_negation-triplets_samples_per_second": 38.535, + "eval_negation-triplets_steps_per_second": 0.301, + "step": 312 + }, + { + "epoch": 0.6015424164524421, + "eval_scitail-pairs-pos_loss": 0.007662723306566477, + "eval_scitail-pairs-pos_runtime": 2.6317, + "eval_scitail-pairs-pos_samples_per_second": 48.637, + "eval_scitail-pairs-pos_steps_per_second": 0.38, + "step": 312 + }, + { + "epoch": 0.6015424164524421, + "eval_scitail-pairs-qa_loss": 0.014555388130247593, + "eval_scitail-pairs-qa_runtime": 2.1734, + "eval_scitail-pairs-qa_samples_per_second": 58.895, + "eval_scitail-pairs-qa_steps_per_second": 0.46, + "step": 312 + }, + { + "epoch": 0.6015424164524421, + "eval_xsum-pairs_loss": 0.004975470248609781, + "eval_xsum-pairs_runtime": 3.0489, + "eval_xsum-pairs_samples_per_second": 41.982, + "eval_xsum-pairs_steps_per_second": 0.328, + "step": 312 + }, + { + "epoch": 0.6015424164524421, + "eval_sciq_pairs_loss": 0.08704940229654312, + "eval_sciq_pairs_runtime": 4.7639, + "eval_sciq_pairs_samples_per_second": 26.869, + "eval_sciq_pairs_steps_per_second": 0.21, + "step": 312 + }, + { + "epoch": 0.6015424164524421, + "eval_qasc_pairs_loss": 0.15072381496429443, + "eval_qasc_pairs_runtime": 2.0766, + "eval_qasc_pairs_samples_per_second": 61.638, + "eval_qasc_pairs_steps_per_second": 0.482, + "step": 312 + }, + { + "epoch": 0.6015424164524421, + "eval_openbookqa_pairs_loss": 0.4214599132537842, + "eval_openbookqa_pairs_runtime": 2.1857, + "eval_openbookqa_pairs_samples_per_second": 58.564, + "eval_openbookqa_pairs_steps_per_second": 0.458, + "step": 312 + }, + { + "epoch": 0.6015424164524421, + "eval_msmarco_pairs_loss": 0.04896237328648567, + "eval_msmarco_pairs_runtime": 2.2125, + "eval_msmarco_pairs_samples_per_second": 57.853, + "eval_msmarco_pairs_steps_per_second": 0.452, + "step": 312 + }, + { + "epoch": 0.6015424164524421, + "eval_nq_pairs_loss": 0.08789835125207901, + "eval_nq_pairs_runtime": 3.6548, + "eval_nq_pairs_samples_per_second": 35.022, + "eval_nq_pairs_steps_per_second": 0.274, + "step": 312 + }, + { + "epoch": 0.6015424164524421, + "eval_trivia_pairs_loss": 0.04830099269747734, + "eval_trivia_pairs_runtime": 3.0616, + "eval_trivia_pairs_samples_per_second": 41.808, + "eval_trivia_pairs_steps_per_second": 0.327, + "step": 312 + }, + { + "epoch": 0.6015424164524421, + "eval_gooaq_pairs_loss": 0.04206620529294014, + "eval_gooaq_pairs_runtime": 2.1174, + "eval_gooaq_pairs_samples_per_second": 60.452, + "eval_gooaq_pairs_steps_per_second": 0.472, + "step": 312 + }, + { + "epoch": 0.6015424164524421, + "eval_paws-pos_loss": 0.024975091218948364, + "eval_paws-pos_runtime": 2.201, + "eval_paws-pos_samples_per_second": 58.154, + "eval_paws-pos_steps_per_second": 0.454, + "step": 312 + }, + { + "epoch": 0.6015424164524421, + "eval_global_dataset_loss": 0.23649312555789948, + "eval_global_dataset_runtime": 8.3828, + "eval_global_dataset_samples_per_second": 38.77, + "eval_global_dataset_steps_per_second": 0.358, + "step": 312 + }, + { + "epoch": 0.6073264781491002, + "grad_norm": 2.4958090782165527, + "learning_rate": 1.4883588772779416e-05, + "loss": 0.1218, + "step": 315 + }, + { + "epoch": 0.6131105398457584, + "grad_norm": 3.888019323348999, + "learning_rate": 1.4867960873738955e-05, + "loss": 0.1575, + "step": 318 + }, + { + "epoch": 0.6188946015424165, + "grad_norm": 5.112899303436279, + "learning_rate": 1.485135928833522e-05, + "loss": 0.178, + "step": 321 + }, + { + "epoch": 0.6246786632390745, + "grad_norm": 2.7898402214050293, + "learning_rate": 1.4833786459123798e-05, + "loss": 0.0777, + "step": 324 + }, + { + "epoch": 0.6304627249357326, + "grad_norm": 6.2421488761901855, + "learning_rate": 1.4815244971557292e-05, + "loss": 0.0696, + "step": 327 + }, + { + "epoch": 0.6362467866323908, + "grad_norm": 133.5791015625, + "learning_rate": 1.4795737553604924e-05, + "loss": 0.1206, + "step": 330 + }, + { + "epoch": 0.6420308483290489, + "grad_norm": 9.727315902709961, + "learning_rate": 1.477526707535119e-05, + "loss": 0.0926, + "step": 333 + }, + { + "epoch": 0.6478149100257069, + "grad_norm": 6.652628421783447, + "learning_rate": 1.4753836548573564e-05, + "loss": 0.1357, + "step": 336 + }, + { + "epoch": 0.6516709511568124, + "eval_Qnli-dev_cosine_accuracy": 0.703125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.7029188275337219, + "eval_Qnli-dev_cosine_ap": 0.7383384263937729, + "eval_Qnli-dev_cosine_f1": 0.7160940325497287, + "eval_Qnli-dev_cosine_f1_threshold": 0.6273617744445801, + "eval_Qnli-dev_cosine_precision": 0.6246056782334385, + "eval_Qnli-dev_cosine_recall": 0.8389830508474576, + "eval_Qnli-dev_dot_accuracy": 0.69921875, + "eval_Qnli-dev_dot_accuracy_threshold": 647.5323486328125, + "eval_Qnli-dev_dot_ap": 0.7214295624305991, + "eval_Qnli-dev_dot_f1": 0.6965517241379311, + "eval_Qnli-dev_dot_f1_threshold": 570.4014282226562, + "eval_Qnli-dev_dot_precision": 0.5872093023255814, + "eval_Qnli-dev_dot_recall": 0.8559322033898306, + "eval_Qnli-dev_euclidean_accuracy": 0.70703125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 23.107650756835938, + "eval_Qnli-dev_euclidean_ap": 0.7395303199325498, + "eval_Qnli-dev_euclidean_f1": 0.7045871559633027, + "eval_Qnli-dev_euclidean_f1_threshold": 26.295127868652344, + "eval_Qnli-dev_euclidean_precision": 0.6213592233009708, + "eval_Qnli-dev_euclidean_recall": 0.8135593220338984, + "eval_Qnli-dev_manhattan_accuracy": 0.701171875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 702.7601928710938, + "eval_Qnli-dev_manhattan_ap": 0.7390235736074865, + "eval_Qnli-dev_manhattan_f1": 0.7094474153297684, + "eval_Qnli-dev_manhattan_f1_threshold": 828.67724609375, + "eval_Qnli-dev_manhattan_precision": 0.6123076923076923, + "eval_Qnli-dev_manhattan_recall": 0.8432203389830508, + "eval_Qnli-dev_max_accuracy": 0.70703125, + "eval_Qnli-dev_max_accuracy_threshold": 702.7601928710938, + "eval_Qnli-dev_max_ap": 0.7395303199325498, + "eval_Qnli-dev_max_f1": 0.7160940325497287, + "eval_Qnli-dev_max_f1_threshold": 828.67724609375, + "eval_Qnli-dev_max_precision": 0.6246056782334385, + "eval_Qnli-dev_max_recall": 0.8559322033898306, + "eval_allNLI-dev_cosine_accuracy": 0.716796875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8458511233329773, + "eval_allNLI-dev_cosine_ap": 0.596148092754136, + "eval_allNLI-dev_cosine_f1": 0.6157517899761337, + "eval_allNLI-dev_cosine_f1_threshold": 0.7034751176834106, + "eval_allNLI-dev_cosine_precision": 0.524390243902439, + "eval_allNLI-dev_cosine_recall": 0.7456647398843931, + "eval_allNLI-dev_dot_accuracy": 0.7109375, + "eval_allNLI-dev_dot_accuracy_threshold": 726.9520263671875, + "eval_allNLI-dev_dot_ap": 0.5809559752676249, + "eval_allNLI-dev_dot_f1": 0.6153846153846154, + "eval_allNLI-dev_dot_f1_threshold": 624.0245361328125, + "eval_allNLI-dev_dot_precision": 0.488135593220339, + "eval_allNLI-dev_dot_recall": 0.8323699421965318, + "eval_allNLI-dev_euclidean_accuracy": 0.716796875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 17.03449249267578, + "eval_allNLI-dev_euclidean_ap": 0.5966915737527797, + "eval_allNLI-dev_euclidean_f1": 0.6265060240963856, + "eval_allNLI-dev_euclidean_f1_threshold": 23.620059967041016, + "eval_allNLI-dev_euclidean_precision": 0.5371900826446281, + "eval_allNLI-dev_euclidean_recall": 0.7514450867052023, + "eval_allNLI-dev_manhattan_accuracy": 0.716796875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 531.2730712890625, + "eval_allNLI-dev_manhattan_ap": 0.598298778219752, + "eval_allNLI-dev_manhattan_f1": 0.6231884057971014, + "eval_allNLI-dev_manhattan_f1_threshold": 734.1275634765625, + "eval_allNLI-dev_manhattan_precision": 0.5352697095435685, + "eval_allNLI-dev_manhattan_recall": 0.7456647398843931, + "eval_allNLI-dev_max_accuracy": 0.716796875, + "eval_allNLI-dev_max_accuracy_threshold": 726.9520263671875, + "eval_allNLI-dev_max_ap": 0.598298778219752, + "eval_allNLI-dev_max_f1": 0.6265060240963856, + "eval_allNLI-dev_max_f1_threshold": 734.1275634765625, + "eval_allNLI-dev_max_precision": 0.5371900826446281, + "eval_allNLI-dev_max_recall": 0.8323699421965318, + "eval_sequential_score": 0.7395303199325498, + "eval_sts-test_pearson_cosine": 0.9132656286571847, + "eval_sts-test_pearson_dot": 0.900423684668929, + "eval_sts-test_pearson_euclidean": 0.930599039345619, + "eval_sts-test_pearson_manhattan": 0.9300256250399432, + "eval_sts-test_pearson_max": 0.930599039345619, + "eval_sts-test_spearman_cosine": 0.9248599987724085, + "eval_sts-test_spearman_dot": 0.8983566907310115, + "eval_sts-test_spearman_euclidean": 0.926492435285444, + "eval_sts-test_spearman_manhattan": 0.9259926291445923, + "eval_sts-test_spearman_max": 0.926492435285444, + "eval_vitaminc-pairs_loss": 3.174186944961548, + "eval_vitaminc-pairs_runtime": 4.5721, + "eval_vitaminc-pairs_samples_per_second": 27.996, + "eval_vitaminc-pairs_steps_per_second": 0.219, + "step": 338 + }, + { + "epoch": 0.6516709511568124, + "eval_negation-triplets_loss": 0.5640384554862976, + "eval_negation-triplets_runtime": 3.4331, + "eval_negation-triplets_samples_per_second": 37.285, + "eval_negation-triplets_steps_per_second": 0.291, + "step": 338 + }, + { + "epoch": 0.6516709511568124, + "eval_scitail-pairs-pos_loss": 0.010583501309156418, + "eval_scitail-pairs-pos_runtime": 2.746, + "eval_scitail-pairs-pos_samples_per_second": 46.613, + "eval_scitail-pairs-pos_steps_per_second": 0.364, + "step": 338 + }, + { + "epoch": 0.6516709511568124, + "eval_scitail-pairs-qa_loss": 0.0013579510850831866, + "eval_scitail-pairs-qa_runtime": 2.2771, + "eval_scitail-pairs-qa_samples_per_second": 56.211, + "eval_scitail-pairs-qa_steps_per_second": 0.439, + "step": 338 + }, + { + "epoch": 0.6516709511568124, + "eval_xsum-pairs_loss": 0.005872013047337532, + "eval_xsum-pairs_runtime": 3.1545, + "eval_xsum-pairs_samples_per_second": 40.577, + "eval_xsum-pairs_steps_per_second": 0.317, + "step": 338 + }, + { + "epoch": 0.6516709511568124, + "eval_sciq_pairs_loss": 0.0752023309469223, + "eval_sciq_pairs_runtime": 4.8292, + "eval_sciq_pairs_samples_per_second": 26.505, + "eval_sciq_pairs_steps_per_second": 0.207, + "step": 338 + }, + { + "epoch": 0.6516709511568124, + "eval_qasc_pairs_loss": 0.1399243026971817, + "eval_qasc_pairs_runtime": 2.1004, + "eval_qasc_pairs_samples_per_second": 60.942, + "eval_qasc_pairs_steps_per_second": 0.476, + "step": 338 + }, + { + "epoch": 0.6516709511568124, + "eval_openbookqa_pairs_loss": 0.4193750321865082, + "eval_openbookqa_pairs_runtime": 2.2958, + "eval_openbookqa_pairs_samples_per_second": 55.754, + "eval_openbookqa_pairs_steps_per_second": 0.436, + "step": 338 + }, + { + "epoch": 0.6516709511568124, + "eval_msmarco_pairs_loss": 0.037960685789585114, + "eval_msmarco_pairs_runtime": 2.2487, + "eval_msmarco_pairs_samples_per_second": 56.922, + "eval_msmarco_pairs_steps_per_second": 0.445, + "step": 338 + }, + { + "epoch": 0.6516709511568124, + "eval_nq_pairs_loss": 0.11820158362388611, + "eval_nq_pairs_runtime": 3.6602, + "eval_nq_pairs_samples_per_second": 34.971, + "eval_nq_pairs_steps_per_second": 0.273, + "step": 338 + }, + { + "epoch": 0.6516709511568124, + "eval_trivia_pairs_loss": 0.028769271448254585, + "eval_trivia_pairs_runtime": 3.1269, + "eval_trivia_pairs_samples_per_second": 40.935, + "eval_trivia_pairs_steps_per_second": 0.32, + "step": 338 + }, + { + "epoch": 0.6516709511568124, + "eval_gooaq_pairs_loss": 0.0657971203327179, + "eval_gooaq_pairs_runtime": 2.1776, + "eval_gooaq_pairs_samples_per_second": 58.779, + "eval_gooaq_pairs_steps_per_second": 0.459, + "step": 338 + }, + { + "epoch": 0.6516709511568124, + "eval_paws-pos_loss": 0.02519122324883938, + "eval_paws-pos_runtime": 2.3434, + "eval_paws-pos_samples_per_second": 54.622, + "eval_paws-pos_steps_per_second": 0.427, + "step": 338 + }, + { + "epoch": 0.6516709511568124, + "eval_global_dataset_loss": 0.19101615250110626, + "eval_global_dataset_runtime": 8.4497, + "eval_global_dataset_samples_per_second": 38.463, + "eval_global_dataset_steps_per_second": 0.355, + "step": 338 + }, + { + "epoch": 0.653598971722365, + "grad_norm": 45.49017333984375, + "learning_rate": 1.4731449126299404e-05, + "loss": 0.1234, + "step": 339 + }, + { + "epoch": 0.6593830334190232, + "grad_norm": 7.209679126739502, + "learning_rate": 1.4708108102342047e-05, + "loss": 0.1056, + "step": 342 + }, + { + "epoch": 0.6651670951156813, + "grad_norm": 6.202587604522705, + "learning_rate": 1.468381691081619e-05, + "loss": 0.1525, + "step": 345 + }, + { + "epoch": 0.6709511568123393, + "grad_norm": 7.8031535148620605, + "learning_rate": 1.4658579125632652e-05, + "loss": 0.1425, + "step": 348 + }, + { + "epoch": 0.6767352185089974, + "grad_norm": 4.834568500518799, + "learning_rate": 1.4632398459972538e-05, + "loss": 0.1401, + "step": 351 + }, + { + "epoch": 0.6825192802056556, + "grad_norm": 5.427046298980713, + "learning_rate": 1.4605278765740929e-05, + "loss": 0.1271, + "step": 354 + }, + { + "epoch": 0.6883033419023136, + "grad_norm": 1.4662452936172485, + "learning_rate": 1.457722403300017e-05, + "loss": 0.0598, + "step": 357 + }, + { + "epoch": 0.6940874035989717, + "grad_norm": 2.993671417236328, + "learning_rate": 1.4548238389382802e-05, + "loss": 0.0681, + "step": 360 + }, + { + "epoch": 0.6998714652956298, + "grad_norm": 10.515313148498535, + "learning_rate": 1.4518326099484293e-05, + "loss": 0.1132, + "step": 363 + }, + { + "epoch": 0.7017994858611826, + "eval_Qnli-dev_cosine_accuracy": 0.69921875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.7072766423225403, + "eval_Qnli-dev_cosine_ap": 0.7245987545496009, + "eval_Qnli-dev_cosine_f1": 0.7016274864376131, + "eval_Qnli-dev_cosine_f1_threshold": 0.6138921976089478, + "eval_Qnli-dev_cosine_precision": 0.61198738170347, + "eval_Qnli-dev_cosine_recall": 0.8220338983050848, + "eval_Qnli-dev_dot_accuracy": 0.68359375, + "eval_Qnli-dev_dot_accuracy_threshold": 638.5958251953125, + "eval_Qnli-dev_dot_ap": 0.6982591186827876, + "eval_Qnli-dev_dot_f1": 0.6946308724832215, + "eval_Qnli-dev_dot_f1_threshold": 531.0111083984375, + "eval_Qnli-dev_dot_precision": 0.575, + "eval_Qnli-dev_dot_recall": 0.8771186440677966, + "eval_Qnli-dev_euclidean_accuracy": 0.70703125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 23.832258224487305, + "eval_Qnli-dev_euclidean_ap": 0.7287738238939392, + "eval_Qnli-dev_euclidean_f1": 0.6975088967971529, + "eval_Qnli-dev_euclidean_f1_threshold": 27.3236083984375, + "eval_Qnli-dev_euclidean_precision": 0.6012269938650306, + "eval_Qnli-dev_euclidean_recall": 0.8305084745762712, + "eval_Qnli-dev_manhattan_accuracy": 0.7109375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 737.96826171875, + "eval_Qnli-dev_manhattan_ap": 0.7286289416450926, + "eval_Qnli-dev_manhattan_f1": 0.7001795332136445, + "eval_Qnli-dev_manhattan_f1_threshold": 847.50732421875, + "eval_Qnli-dev_manhattan_precision": 0.6074766355140186, + "eval_Qnli-dev_manhattan_recall": 0.826271186440678, + "eval_Qnli-dev_max_accuracy": 0.7109375, + "eval_Qnli-dev_max_accuracy_threshold": 737.96826171875, + "eval_Qnli-dev_max_ap": 0.7287738238939392, + "eval_Qnli-dev_max_f1": 0.7016274864376131, + "eval_Qnli-dev_max_f1_threshold": 847.50732421875, + "eval_Qnli-dev_max_precision": 0.61198738170347, + "eval_Qnli-dev_max_recall": 0.8771186440677966, + "eval_allNLI-dev_cosine_accuracy": 0.716796875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.793137788772583, + "eval_allNLI-dev_cosine_ap": 0.5846176608505069, + "eval_allNLI-dev_cosine_f1": 0.6329670329670329, + "eval_allNLI-dev_cosine_f1_threshold": 0.6309775114059448, + "eval_allNLI-dev_cosine_precision": 0.5106382978723404, + "eval_allNLI-dev_cosine_recall": 0.8323699421965318, + "eval_allNLI-dev_dot_accuracy": 0.712890625, + "eval_allNLI-dev_dot_accuracy_threshold": 794.71240234375, + "eval_allNLI-dev_dot_ap": 0.5801144123861134, + "eval_allNLI-dev_dot_f1": 0.6205357142857143, + "eval_allNLI-dev_dot_f1_threshold": 594.3692626953125, + "eval_allNLI-dev_dot_precision": 0.5054545454545455, + "eval_allNLI-dev_dot_recall": 0.8034682080924855, + "eval_allNLI-dev_euclidean_accuracy": 0.716796875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 18.532958984375, + "eval_allNLI-dev_euclidean_ap": 0.5858640058497049, + "eval_allNLI-dev_euclidean_f1": 0.6361556064073227, + "eval_allNLI-dev_euclidean_f1_threshold": 25.700441360473633, + "eval_allNLI-dev_euclidean_precision": 0.5265151515151515, + "eval_allNLI-dev_euclidean_recall": 0.8034682080924855, + "eval_allNLI-dev_manhattan_accuracy": 0.716796875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 577.34521484375, + "eval_allNLI-dev_manhattan_ap": 0.585729227241877, + "eval_allNLI-dev_manhattan_f1": 0.6339285714285714, + "eval_allNLI-dev_manhattan_f1_threshold": 813.3930053710938, + "eval_allNLI-dev_manhattan_precision": 0.5163636363636364, + "eval_allNLI-dev_manhattan_recall": 0.8208092485549133, + "eval_allNLI-dev_max_accuracy": 0.716796875, + "eval_allNLI-dev_max_accuracy_threshold": 794.71240234375, + "eval_allNLI-dev_max_ap": 0.5858640058497049, + "eval_allNLI-dev_max_f1": 0.6361556064073227, + "eval_allNLI-dev_max_f1_threshold": 813.3930053710938, + "eval_allNLI-dev_max_precision": 0.5265151515151515, + "eval_allNLI-dev_max_recall": 0.8323699421965318, + "eval_sequential_score": 0.7287738238939392, + "eval_sts-test_pearson_cosine": 0.9168734909278741, + "eval_sts-test_pearson_dot": 0.9060722657495308, + "eval_sts-test_pearson_euclidean": 0.9292788993338806, + "eval_sts-test_pearson_manhattan": 0.9286667562357451, + "eval_sts-test_pearson_max": 0.9292788993338806, + "eval_sts-test_spearman_cosine": 0.923743231044386, + "eval_sts-test_spearman_dot": 0.9006938701084388, + "eval_sts-test_spearman_euclidean": 0.9245718006490479, + "eval_sts-test_spearman_manhattan": 0.9238214114980368, + "eval_sts-test_spearman_max": 0.9245718006490479, + "eval_vitaminc-pairs_loss": 3.7763519287109375, + "eval_vitaminc-pairs_runtime": 4.4308, + "eval_vitaminc-pairs_samples_per_second": 28.889, + "eval_vitaminc-pairs_steps_per_second": 0.226, + "step": 364 + }, + { + "epoch": 0.7017994858611826, + "eval_negation-triplets_loss": 0.5828245282173157, + "eval_negation-triplets_runtime": 3.3268, + "eval_negation-triplets_samples_per_second": 38.476, + "eval_negation-triplets_steps_per_second": 0.301, + "step": 364 + }, + { + "epoch": 0.7017994858611826, + "eval_scitail-pairs-pos_loss": 0.011268602684140205, + "eval_scitail-pairs-pos_runtime": 2.6437, + "eval_scitail-pairs-pos_samples_per_second": 48.416, + "eval_scitail-pairs-pos_steps_per_second": 0.378, + "step": 364 + }, + { + "epoch": 0.7017994858611826, + "eval_scitail-pairs-qa_loss": 0.0031084802467375994, + "eval_scitail-pairs-qa_runtime": 2.2151, + "eval_scitail-pairs-qa_samples_per_second": 57.785, + "eval_scitail-pairs-qa_steps_per_second": 0.451, + "step": 364 + }, + { + "epoch": 0.7017994858611826, + "eval_xsum-pairs_loss": 0.01026161853224039, + "eval_xsum-pairs_runtime": 3.0546, + "eval_xsum-pairs_samples_per_second": 41.904, + "eval_xsum-pairs_steps_per_second": 0.327, + "step": 364 + }, + { + "epoch": 0.7017994858611826, + "eval_sciq_pairs_loss": 0.07422904670238495, + "eval_sciq_pairs_runtime": 4.7698, + "eval_sciq_pairs_samples_per_second": 26.836, + "eval_sciq_pairs_steps_per_second": 0.21, + "step": 364 + }, + { + "epoch": 0.7017994858611826, + "eval_qasc_pairs_loss": 0.06358247995376587, + "eval_qasc_pairs_runtime": 2.0983, + "eval_qasc_pairs_samples_per_second": 61.002, + "eval_qasc_pairs_steps_per_second": 0.477, + "step": 364 + }, + { + "epoch": 0.7017994858611826, + "eval_openbookqa_pairs_loss": 0.34939464926719666, + "eval_openbookqa_pairs_runtime": 2.1938, + "eval_openbookqa_pairs_samples_per_second": 58.348, + "eval_openbookqa_pairs_steps_per_second": 0.456, + "step": 364 + }, + { + "epoch": 0.7017994858611826, + "eval_msmarco_pairs_loss": 0.02596309781074524, + "eval_msmarco_pairs_runtime": 2.2095, + "eval_msmarco_pairs_samples_per_second": 57.931, + "eval_msmarco_pairs_steps_per_second": 0.453, + "step": 364 + }, + { + "epoch": 0.7017994858611826, + "eval_nq_pairs_loss": 0.026090342551469803, + "eval_nq_pairs_runtime": 3.6272, + "eval_nq_pairs_samples_per_second": 35.289, + "eval_nq_pairs_steps_per_second": 0.276, + "step": 364 + }, + { + "epoch": 0.7017994858611826, + "eval_trivia_pairs_loss": 0.019855637103319168, + "eval_trivia_pairs_runtime": 3.0864, + "eval_trivia_pairs_samples_per_second": 41.472, + "eval_trivia_pairs_steps_per_second": 0.324, + "step": 364 + }, + { + "epoch": 0.7017994858611826, + "eval_gooaq_pairs_loss": 0.07295343279838562, + "eval_gooaq_pairs_runtime": 2.1333, + "eval_gooaq_pairs_samples_per_second": 60.0, + "eval_gooaq_pairs_steps_per_second": 0.469, + "step": 364 + }, + { + "epoch": 0.7017994858611826, + "eval_paws-pos_loss": 0.02318389154970646, + "eval_paws-pos_runtime": 2.2203, + "eval_paws-pos_samples_per_second": 57.65, + "eval_paws-pos_steps_per_second": 0.45, + "step": 364 + }, + { + "epoch": 0.7017994858611826, + "eval_global_dataset_loss": 0.24246200919151306, + "eval_global_dataset_runtime": 8.3489, + "eval_global_dataset_samples_per_second": 38.927, + "eval_global_dataset_steps_per_second": 0.359, + "step": 364 + }, + { + "epoch": 0.705655526992288, + "grad_norm": 3.3271055221557617, + "learning_rate": 1.448749156423558e-05, + "loss": 0.0714, + "step": 366 + }, + { + "epoch": 0.711439588688946, + "grad_norm": 5.8299336433410645, + "learning_rate": 1.4455739320255581e-05, + "loss": 0.1816, + "step": 369 + }, + { + "epoch": 0.7172236503856041, + "grad_norm": 5.665037155151367, + "learning_rate": 1.4423074039183725e-05, + "loss": 0.1609, + "step": 372 + }, + { + "epoch": 0.7230077120822622, + "grad_norm": 3.767242193222046, + "learning_rate": 1.4389500526992632e-05, + "loss": 0.1123, + "step": 375 + }, + { + "epoch": 0.7287917737789203, + "grad_norm": 6.201554298400879, + "learning_rate": 1.4355023723281005e-05, + "loss": 0.1906, + "step": 378 + }, + { + "epoch": 0.7345758354755784, + "grad_norm": 6.10802698135376, + "learning_rate": 1.4319648700546897e-05, + "loss": 0.0689, + "step": 381 + }, + { + "epoch": 0.7403598971722365, + "grad_norm": 3.1739063262939453, + "learning_rate": 1.4283380663441395e-05, + "loss": 0.1897, + "step": 384 + }, + { + "epoch": 0.7461439588688946, + "grad_norm": 4.408695220947266, + "learning_rate": 1.4246224948002864e-05, + "loss": 0.1268, + "step": 387 + }, + { + "epoch": 0.7519280205655527, + "grad_norm": 4.3629350662231445, + "learning_rate": 1.4208187020871878e-05, + "loss": 0.1256, + "step": 390 + }, + { + "epoch": 0.7519280205655527, + "eval_Qnli-dev_cosine_accuracy": 0.71875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.687468409538269, + "eval_Qnli-dev_cosine_ap": 0.7382336178190824, + "eval_Qnli-dev_cosine_f1": 0.7061224489795919, + "eval_Qnli-dev_cosine_f1_threshold": 0.6693283319473267, + "eval_Qnli-dev_cosine_precision": 0.6811023622047244, + "eval_Qnli-dev_cosine_recall": 0.7330508474576272, + "eval_Qnli-dev_dot_accuracy": 0.703125, + "eval_Qnli-dev_dot_accuracy_threshold": 671.2976684570312, + "eval_Qnli-dev_dot_ap": 0.7084828885948017, + "eval_Qnli-dev_dot_f1": 0.6944908180300502, + "eval_Qnli-dev_dot_f1_threshold": 556.4210205078125, + "eval_Qnli-dev_dot_precision": 0.5730027548209367, + "eval_Qnli-dev_dot_recall": 0.8813559322033898, + "eval_Qnli-dev_euclidean_accuracy": 0.71875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 24.559717178344727, + "eval_Qnli-dev_euclidean_ap": 0.739908359646434, + "eval_Qnli-dev_euclidean_f1": 0.705223880597015, + "eval_Qnli-dev_euclidean_f1_threshold": 25.997222900390625, + "eval_Qnli-dev_euclidean_precision": 0.63, + "eval_Qnli-dev_euclidean_recall": 0.8008474576271186, + "eval_Qnli-dev_manhattan_accuracy": 0.716796875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 754.3623046875, + "eval_Qnli-dev_manhattan_ap": 0.7392277169186197, + "eval_Qnli-dev_manhattan_f1": 0.7067669172932329, + "eval_Qnli-dev_manhattan_f1_threshold": 808.9886474609375, + "eval_Qnli-dev_manhattan_precision": 0.6351351351351351, + "eval_Qnli-dev_manhattan_recall": 0.7966101694915254, + "eval_Qnli-dev_max_accuracy": 0.71875, + "eval_Qnli-dev_max_accuracy_threshold": 754.3623046875, + "eval_Qnli-dev_max_ap": 0.739908359646434, + "eval_Qnli-dev_max_f1": 0.7067669172932329, + "eval_Qnli-dev_max_f1_threshold": 808.9886474609375, + "eval_Qnli-dev_max_precision": 0.6811023622047244, + "eval_Qnli-dev_max_recall": 0.8813559322033898, + "eval_allNLI-dev_cosine_accuracy": 0.720703125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8184170126914978, + "eval_allNLI-dev_cosine_ap": 0.5969272075792788, + "eval_allNLI-dev_cosine_f1": 0.6310904872389791, + "eval_allNLI-dev_cosine_f1_threshold": 0.6732250452041626, + "eval_allNLI-dev_cosine_precision": 0.5271317829457365, + "eval_allNLI-dev_cosine_recall": 0.7861271676300579, + "eval_allNLI-dev_dot_accuracy": 0.7265625, + "eval_allNLI-dev_dot_accuracy_threshold": 728.9887084960938, + "eval_allNLI-dev_dot_ap": 0.5950023435373721, + "eval_allNLI-dev_dot_f1": 0.6263498920086393, + "eval_allNLI-dev_dot_f1_threshold": 599.548583984375, + "eval_allNLI-dev_dot_precision": 0.5, + "eval_allNLI-dev_dot_recall": 0.838150289017341, + "eval_allNLI-dev_euclidean_accuracy": 0.720703125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 19.761653900146484, + "eval_allNLI-dev_euclidean_ap": 0.5951396590656841, + "eval_allNLI-dev_euclidean_f1": 0.631336405529954, + "eval_allNLI-dev_euclidean_f1_threshold": 24.669166564941406, + "eval_allNLI-dev_euclidean_precision": 0.524904214559387, + "eval_allNLI-dev_euclidean_recall": 0.791907514450867, + "eval_allNLI-dev_manhattan_accuracy": 0.720703125, + "eval_allNLI-dev_manhattan_accuracy_threshold": 621.3643798828125, + "eval_allNLI-dev_manhattan_ap": 0.5949760338392235, + "eval_allNLI-dev_manhattan_f1": 0.6415094339622641, + "eval_allNLI-dev_manhattan_f1_threshold": 759.7421264648438, + "eval_allNLI-dev_manhattan_precision": 0.5418326693227091, + "eval_allNLI-dev_manhattan_recall": 0.7861271676300579, + "eval_allNLI-dev_max_accuracy": 0.7265625, + "eval_allNLI-dev_max_accuracy_threshold": 728.9887084960938, + "eval_allNLI-dev_max_ap": 0.5969272075792788, + "eval_allNLI-dev_max_f1": 0.6415094339622641, + "eval_allNLI-dev_max_f1_threshold": 759.7421264648438, + "eval_allNLI-dev_max_precision": 0.5418326693227091, + "eval_allNLI-dev_max_recall": 0.838150289017341, + "eval_sequential_score": 0.739908359646434, + "eval_sts-test_pearson_cosine": 0.9171550146572618, + "eval_sts-test_pearson_dot": 0.9078360871894045, + "eval_sts-test_pearson_euclidean": 0.9297785604600608, + "eval_sts-test_pearson_manhattan": 0.9289608083051153, + "eval_sts-test_pearson_max": 0.9297785604600608, + "eval_sts-test_spearman_cosine": 0.9273106673487241, + "eval_sts-test_spearman_dot": 0.9100156921100516, + "eval_sts-test_spearman_euclidean": 0.927380166057665, + "eval_sts-test_spearman_manhattan": 0.9263161600668233, + "eval_sts-test_spearman_max": 0.927380166057665, + "eval_vitaminc-pairs_loss": 3.1046688556671143, + "eval_vitaminc-pairs_runtime": 4.5292, + "eval_vitaminc-pairs_samples_per_second": 28.261, + "eval_vitaminc-pairs_steps_per_second": 0.221, + "step": 390 + }, + { + "epoch": 0.7519280205655527, + "eval_negation-triplets_loss": 0.575215756893158, + "eval_negation-triplets_runtime": 3.4581, + "eval_negation-triplets_samples_per_second": 37.014, + "eval_negation-triplets_steps_per_second": 0.289, + "step": 390 + }, + { + "epoch": 0.7519280205655527, + "eval_scitail-pairs-pos_loss": 0.008397433906793594, + "eval_scitail-pairs-pos_runtime": 2.7763, + "eval_scitail-pairs-pos_samples_per_second": 46.104, + "eval_scitail-pairs-pos_steps_per_second": 0.36, + "step": 390 + }, + { + "epoch": 0.7519280205655527, + "eval_scitail-pairs-qa_loss": 0.0021291954908519983, + "eval_scitail-pairs-qa_runtime": 2.2643, + "eval_scitail-pairs-qa_samples_per_second": 56.53, + "eval_scitail-pairs-qa_steps_per_second": 0.442, + "step": 390 + }, + { + "epoch": 0.7519280205655527, + "eval_xsum-pairs_loss": 0.006353999022394419, + "eval_xsum-pairs_runtime": 3.1689, + "eval_xsum-pairs_samples_per_second": 40.393, + "eval_xsum-pairs_steps_per_second": 0.316, + "step": 390 + }, + { + "epoch": 0.7519280205655527, + "eval_sciq_pairs_loss": 0.06674782186746597, + "eval_sciq_pairs_runtime": 4.8334, + "eval_sciq_pairs_samples_per_second": 26.483, + "eval_sciq_pairs_steps_per_second": 0.207, + "step": 390 + }, + { + "epoch": 0.7519280205655527, + "eval_qasc_pairs_loss": 0.0450245626270771, + "eval_qasc_pairs_runtime": 2.1448, + "eval_qasc_pairs_samples_per_second": 59.68, + "eval_qasc_pairs_steps_per_second": 0.466, + "step": 390 + }, + { + "epoch": 0.7519280205655527, + "eval_openbookqa_pairs_loss": 0.34624671936035156, + "eval_openbookqa_pairs_runtime": 2.2437, + "eval_openbookqa_pairs_samples_per_second": 57.048, + "eval_openbookqa_pairs_steps_per_second": 0.446, + "step": 390 + }, + { + "epoch": 0.7519280205655527, + "eval_msmarco_pairs_loss": 0.0238041914999485, + "eval_msmarco_pairs_runtime": 2.222, + "eval_msmarco_pairs_samples_per_second": 57.606, + "eval_msmarco_pairs_steps_per_second": 0.45, + "step": 390 + }, + { + "epoch": 0.7519280205655527, + "eval_nq_pairs_loss": 0.04486466199159622, + "eval_nq_pairs_runtime": 3.632, + "eval_nq_pairs_samples_per_second": 35.243, + "eval_nq_pairs_steps_per_second": 0.275, + "step": 390 + }, + { + "epoch": 0.7519280205655527, + "eval_trivia_pairs_loss": 0.02883004955947399, + "eval_trivia_pairs_runtime": 3.1146, + "eval_trivia_pairs_samples_per_second": 41.097, + "eval_trivia_pairs_steps_per_second": 0.321, + "step": 390 + }, + { + "epoch": 0.7519280205655527, + "eval_gooaq_pairs_loss": 0.05297350883483887, + "eval_gooaq_pairs_runtime": 2.1744, + "eval_gooaq_pairs_samples_per_second": 58.867, + "eval_gooaq_pairs_steps_per_second": 0.46, + "step": 390 + }, + { + "epoch": 0.7519280205655527, + "eval_paws-pos_loss": 0.024740247055888176, + "eval_paws-pos_runtime": 2.3365, + "eval_paws-pos_samples_per_second": 54.782, + "eval_paws-pos_steps_per_second": 0.428, + "step": 390 + }, + { + "epoch": 0.7519280205655527, + "eval_global_dataset_loss": 0.18419289588928223, + "eval_global_dataset_runtime": 8.4479, + "eval_global_dataset_samples_per_second": 38.471, + "eval_global_dataset_steps_per_second": 0.355, + "step": 390 + }, + { + "epoch": 0.7577120822622108, + "grad_norm": 4.638159275054932, + "learning_rate": 1.416927247848693e-05, + "loss": 0.1201, + "step": 393 + }, + { + "epoch": 0.7634961439588689, + "grad_norm": 2.0510141849517822, + "learning_rate": 1.4129487046261021e-05, + "loss": 0.0995, + "step": 396 + }, + { + "epoch": 0.769280205655527, + "grad_norm": 3.916620969772339, + "learning_rate": 1.4088836577739306e-05, + "loss": 0.0825, + "step": 399 + }, + { + "epoch": 0.7750642673521851, + "grad_norm": 1.7055885791778564, + "learning_rate": 1.4047327053737877e-05, + "loss": 0.0778, + "step": 402 + }, + { + "epoch": 0.7808483290488432, + "grad_norm": 4.174785614013672, + "learning_rate": 1.4004964581463804e-05, + "loss": 0.0696, + "step": 405 + }, + { + "epoch": 0.7866323907455013, + "grad_norm": 4.478493690490723, + "learning_rate": 1.3961755393616608e-05, + "loss": 0.1332, + "step": 408 + }, + { + "epoch": 0.7924164524421594, + "grad_norm": 116.97537994384766, + "learning_rate": 1.3917705847471257e-05, + "loss": 0.0684, + "step": 411 + }, + { + "epoch": 0.7982005141388174, + "grad_norm": 7.2861433029174805, + "learning_rate": 1.3872822423942823e-05, + "loss": 0.2002, + "step": 414 + }, + { + "epoch": 0.8020565552699229, + "eval_Qnli-dev_cosine_accuracy": 0.71875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.7014390230178833, + "eval_Qnli-dev_cosine_ap": 0.7508144524343823, + "eval_Qnli-dev_cosine_f1": 0.70703125, + "eval_Qnli-dev_cosine_f1_threshold": 0.6568057537078857, + "eval_Qnli-dev_cosine_precision": 0.6557971014492754, + "eval_Qnli-dev_cosine_recall": 0.7669491525423728, + "eval_Qnli-dev_dot_accuracy": 0.703125, + "eval_Qnli-dev_dot_accuracy_threshold": 700.2732543945312, + "eval_Qnli-dev_dot_ap": 0.7261430213596531, + "eval_Qnli-dev_dot_f1": 0.7074380165289257, + "eval_Qnli-dev_dot_f1_threshold": 554.1097412109375, + "eval_Qnli-dev_dot_precision": 0.5799457994579946, + "eval_Qnli-dev_dot_recall": 0.9067796610169492, + "eval_Qnli-dev_euclidean_accuracy": 0.720703125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 23.440505981445312, + "eval_Qnli-dev_euclidean_ap": 0.7517128865086153, + "eval_Qnli-dev_euclidean_f1": 0.7045454545454546, + "eval_Qnli-dev_euclidean_f1_threshold": 25.780654907226562, + "eval_Qnli-dev_euclidean_precision": 0.636986301369863, + "eval_Qnli-dev_euclidean_recall": 0.788135593220339, + "eval_Qnli-dev_manhattan_accuracy": 0.720703125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 725.7116088867188, + "eval_Qnli-dev_manhattan_ap": 0.7520524924838988, + "eval_Qnli-dev_manhattan_f1": 0.7114624505928854, + "eval_Qnli-dev_manhattan_f1_threshold": 791.727294921875, + "eval_Qnli-dev_manhattan_precision": 0.6666666666666666, + "eval_Qnli-dev_manhattan_recall": 0.7627118644067796, + "eval_Qnli-dev_max_accuracy": 0.720703125, + "eval_Qnli-dev_max_accuracy_threshold": 725.7116088867188, + "eval_Qnli-dev_max_ap": 0.7520524924838988, + "eval_Qnli-dev_max_f1": 0.7114624505928854, + "eval_Qnli-dev_max_f1_threshold": 791.727294921875, + "eval_Qnli-dev_max_precision": 0.6666666666666666, + "eval_Qnli-dev_max_recall": 0.9067796610169492, + "eval_allNLI-dev_cosine_accuracy": 0.7265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7971610426902771, + "eval_allNLI-dev_cosine_ap": 0.6029939953883172, + "eval_allNLI-dev_cosine_f1": 0.6285714285714286, + "eval_allNLI-dev_cosine_f1_threshold": 0.6836909055709839, + "eval_allNLI-dev_cosine_precision": 0.5344129554655871, + "eval_allNLI-dev_cosine_recall": 0.7630057803468208, + "eval_allNLI-dev_dot_accuracy": 0.736328125, + "eval_allNLI-dev_dot_accuracy_threshold": 756.0765380859375, + "eval_allNLI-dev_dot_ap": 0.6094208828475387, + "eval_allNLI-dev_dot_f1": 0.6301969365426696, + "eval_allNLI-dev_dot_f1_threshold": 619.8289794921875, + "eval_allNLI-dev_dot_precision": 0.5070422535211268, + "eval_allNLI-dev_dot_recall": 0.8323699421965318, + "eval_allNLI-dev_euclidean_accuracy": 0.724609375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 19.75946807861328, + "eval_allNLI-dev_euclidean_ap": 0.5979771944382657, + "eval_allNLI-dev_euclidean_f1": 0.6224256292906178, + "eval_allNLI-dev_euclidean_f1_threshold": 24.92770767211914, + "eval_allNLI-dev_euclidean_precision": 0.5151515151515151, + "eval_allNLI-dev_euclidean_recall": 0.7861271676300579, + "eval_allNLI-dev_manhattan_accuracy": 0.72265625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 590.9429321289062, + "eval_allNLI-dev_manhattan_ap": 0.5975736439492428, + "eval_allNLI-dev_manhattan_f1": 0.6261261261261263, + "eval_allNLI-dev_manhattan_f1_threshold": 779.2568359375, + "eval_allNLI-dev_manhattan_precision": 0.5129151291512916, + "eval_allNLI-dev_manhattan_recall": 0.8034682080924855, + "eval_allNLI-dev_max_accuracy": 0.736328125, + "eval_allNLI-dev_max_accuracy_threshold": 756.0765380859375, + "eval_allNLI-dev_max_ap": 0.6094208828475387, + "eval_allNLI-dev_max_f1": 0.6301969365426696, + "eval_allNLI-dev_max_f1_threshold": 779.2568359375, + "eval_allNLI-dev_max_precision": 0.5344129554655871, + "eval_allNLI-dev_max_recall": 0.8323699421965318, + "eval_sequential_score": 0.7520524924838988, + "eval_sts-test_pearson_cosine": 0.9150548197025481, + "eval_sts-test_pearson_dot": 0.9041984235640055, + "eval_sts-test_pearson_euclidean": 0.929928477203424, + "eval_sts-test_pearson_manhattan": 0.9292292029223367, + "eval_sts-test_pearson_max": 0.929928477203424, + "eval_sts-test_spearman_cosine": 0.9273011800813092, + "eval_sts-test_spearman_dot": 0.9082826547614081, + "eval_sts-test_spearman_euclidean": 0.9273154557336931, + "eval_sts-test_spearman_manhattan": 0.9263550489129724, + "eval_sts-test_spearman_max": 0.9273154557336931, + "eval_vitaminc-pairs_loss": 2.4590253829956055, + "eval_vitaminc-pairs_runtime": 4.4297, + "eval_vitaminc-pairs_samples_per_second": 28.896, + "eval_vitaminc-pairs_steps_per_second": 0.226, + "step": 416 + }, + { + "epoch": 0.8020565552699229, + "eval_negation-triplets_loss": 0.5903122425079346, + "eval_negation-triplets_runtime": 3.2993, + "eval_negation-triplets_samples_per_second": 38.796, + "eval_negation-triplets_steps_per_second": 0.303, + "step": 416 + }, + { + "epoch": 0.8020565552699229, + "eval_scitail-pairs-pos_loss": 0.008961405605077744, + "eval_scitail-pairs-pos_runtime": 2.6102, + "eval_scitail-pairs-pos_samples_per_second": 49.039, + "eval_scitail-pairs-pos_steps_per_second": 0.383, + "step": 416 + }, + { + "epoch": 0.8020565552699229, + "eval_scitail-pairs-qa_loss": 0.0034254209604114294, + "eval_scitail-pairs-qa_runtime": 2.2159, + "eval_scitail-pairs-qa_samples_per_second": 57.763, + "eval_scitail-pairs-qa_steps_per_second": 0.451, + "step": 416 + }, + { + "epoch": 0.8020565552699229, + "eval_xsum-pairs_loss": 0.00398410065099597, + "eval_xsum-pairs_runtime": 3.0705, + "eval_xsum-pairs_samples_per_second": 41.687, + "eval_xsum-pairs_steps_per_second": 0.326, + "step": 416 + }, + { + "epoch": 0.8020565552699229, + "eval_sciq_pairs_loss": 0.07794960588216782, + "eval_sciq_pairs_runtime": 4.7751, + "eval_sciq_pairs_samples_per_second": 26.806, + "eval_sciq_pairs_steps_per_second": 0.209, + "step": 416 + }, + { + "epoch": 0.8020565552699229, + "eval_qasc_pairs_loss": 0.08308541029691696, + "eval_qasc_pairs_runtime": 2.0713, + "eval_qasc_pairs_samples_per_second": 61.797, + "eval_qasc_pairs_steps_per_second": 0.483, + "step": 416 + }, + { + "epoch": 0.8020565552699229, + "eval_openbookqa_pairs_loss": 0.36576414108276367, + "eval_openbookqa_pairs_runtime": 2.2061, + "eval_openbookqa_pairs_samples_per_second": 58.022, + "eval_openbookqa_pairs_steps_per_second": 0.453, + "step": 416 + }, + { + "epoch": 0.8020565552699229, + "eval_msmarco_pairs_loss": 0.01674315147101879, + "eval_msmarco_pairs_runtime": 2.2271, + "eval_msmarco_pairs_samples_per_second": 57.473, + "eval_msmarco_pairs_steps_per_second": 0.449, + "step": 416 + }, + { + "epoch": 0.8020565552699229, + "eval_nq_pairs_loss": 0.05656155198812485, + "eval_nq_pairs_runtime": 3.648, + "eval_nq_pairs_samples_per_second": 35.088, + "eval_nq_pairs_steps_per_second": 0.274, + "step": 416 + }, + { + "epoch": 0.8020565552699229, + "eval_trivia_pairs_loss": 0.029444541782140732, + "eval_trivia_pairs_runtime": 3.0578, + "eval_trivia_pairs_samples_per_second": 41.86, + "eval_trivia_pairs_steps_per_second": 0.327, + "step": 416 + }, + { + "epoch": 0.8020565552699229, + "eval_gooaq_pairs_loss": 0.0676947683095932, + "eval_gooaq_pairs_runtime": 2.107, + "eval_gooaq_pairs_samples_per_second": 60.75, + "eval_gooaq_pairs_steps_per_second": 0.475, + "step": 416 + }, + { + "epoch": 0.8020565552699229, + "eval_paws-pos_loss": 0.02595301903784275, + "eval_paws-pos_runtime": 2.2011, + "eval_paws-pos_samples_per_second": 58.152, + "eval_paws-pos_steps_per_second": 0.454, + "step": 416 + }, + { + "epoch": 0.8020565552699229, + "eval_global_dataset_loss": 0.14871826767921448, + "eval_global_dataset_runtime": 8.3457, + "eval_global_dataset_samples_per_second": 38.942, + "eval_global_dataset_steps_per_second": 0.359, + "step": 416 + }, + { + "epoch": 0.8039845758354756, + "grad_norm": 2.4795706272125244, + "learning_rate": 1.3827111726632967e-05, + "loss": 0.0455, + "step": 417 + }, + { + "epoch": 0.8097686375321337, + "grad_norm": 8.65510082244873, + "learning_rate": 1.3780580480858366e-05, + "loss": 0.15, + "step": 420 + }, + { + "epoch": 0.8155526992287918, + "grad_norm": 6.457799434661865, + "learning_rate": 1.373323553266123e-05, + "loss": 0.1232, + "step": 423 + }, + { + "epoch": 0.8213367609254498, + "grad_norm": 4.571555137634277, + "learning_rate": 1.3685083847802057e-05, + "loss": 0.193, + "step": 426 + }, + { + "epoch": 0.827120822622108, + "grad_norm": 6.843802452087402, + "learning_rate": 1.3636132510734774e-05, + "loss": 0.1856, + "step": 429 + }, + { + "epoch": 0.8329048843187661, + "grad_norm": 10.976419448852539, + "learning_rate": 1.3586388723564424e-05, + "loss": 0.0898, + "step": 432 + }, + { + "epoch": 0.8386889460154242, + "grad_norm": 0.5411176085472107, + "learning_rate": 1.3535859804987526e-05, + "loss": 0.0242, + "step": 435 + }, + { + "epoch": 0.8444730077120822, + "grad_norm": 5.699883937835693, + "learning_rate": 1.34845531892153e-05, + "loss": 0.0693, + "step": 438 + }, + { + "epoch": 0.8502570694087404, + "grad_norm": 5.475403308868408, + "learning_rate": 1.3432476424879881e-05, + "loss": 0.1245, + "step": 441 + }, + { + "epoch": 0.8521850899742931, + "eval_Qnli-dev_cosine_accuracy": 0.720703125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6795437335968018, + "eval_Qnli-dev_cosine_ap": 0.7439191659873163, + "eval_Qnli-dev_cosine_f1": 0.7072691552062867, + "eval_Qnli-dev_cosine_f1_threshold": 0.6607608795166016, + "eval_Qnli-dev_cosine_precision": 0.6593406593406593, + "eval_Qnli-dev_cosine_recall": 0.7627118644067796, + "eval_Qnli-dev_dot_accuracy": 0.701171875, + "eval_Qnli-dev_dot_accuracy_threshold": 683.7807006835938, + "eval_Qnli-dev_dot_ap": 0.7138952944956131, + "eval_Qnli-dev_dot_f1": 0.6950092421441774, + "eval_Qnli-dev_dot_f1_threshold": 608.1612548828125, + "eval_Qnli-dev_dot_precision": 0.6163934426229508, + "eval_Qnli-dev_dot_recall": 0.7966101694915254, + "eval_Qnli-dev_euclidean_accuracy": 0.71875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 23.998138427734375, + "eval_Qnli-dev_euclidean_ap": 0.7473177380956941, + "eval_Qnli-dev_euclidean_f1": 0.7063339731285989, + "eval_Qnli-dev_euclidean_f1_threshold": 25.749244689941406, + "eval_Qnli-dev_euclidean_precision": 0.6456140350877193, + "eval_Qnli-dev_euclidean_recall": 0.7796610169491526, + "eval_Qnli-dev_manhattan_accuracy": 0.720703125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 744.572509765625, + "eval_Qnli-dev_manhattan_ap": 0.7478733392923025, + "eval_Qnli-dev_manhattan_f1": 0.7135506003430532, + "eval_Qnli-dev_manhattan_f1_threshold": 845.7078857421875, + "eval_Qnli-dev_manhattan_precision": 0.5994236311239193, + "eval_Qnli-dev_manhattan_recall": 0.8813559322033898, + "eval_Qnli-dev_max_accuracy": 0.720703125, + "eval_Qnli-dev_max_accuracy_threshold": 744.572509765625, + "eval_Qnli-dev_max_ap": 0.7478733392923025, + "eval_Qnli-dev_max_f1": 0.7135506003430532, + "eval_Qnli-dev_max_f1_threshold": 845.7078857421875, + "eval_Qnli-dev_max_precision": 0.6593406593406593, + "eval_Qnli-dev_max_recall": 0.8813559322033898, + "eval_allNLI-dev_cosine_accuracy": 0.7265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7983643412590027, + "eval_allNLI-dev_cosine_ap": 0.6021691080084255, + "eval_allNLI-dev_cosine_f1": 0.6365795724465557, + "eval_allNLI-dev_cosine_f1_threshold": 0.6869878172874451, + "eval_allNLI-dev_cosine_precision": 0.5403225806451613, + "eval_allNLI-dev_cosine_recall": 0.7745664739884393, + "eval_allNLI-dev_dot_accuracy": 0.728515625, + "eval_allNLI-dev_dot_accuracy_threshold": 743.5009765625, + "eval_allNLI-dev_dot_ap": 0.6046964324835868, + "eval_allNLI-dev_dot_f1": 0.6241457858769931, + "eval_allNLI-dev_dot_f1_threshold": 630.758056640625, + "eval_allNLI-dev_dot_precision": 0.5150375939849624, + "eval_allNLI-dev_dot_recall": 0.791907514450867, + "eval_allNLI-dev_euclidean_accuracy": 0.724609375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 19.04987144470215, + "eval_allNLI-dev_euclidean_ap": 0.5993095223669318, + "eval_allNLI-dev_euclidean_f1": 0.6382978723404256, + "eval_allNLI-dev_euclidean_f1_threshold": 24.33264923095703, + "eval_allNLI-dev_euclidean_precision": 0.54, + "eval_allNLI-dev_euclidean_recall": 0.7803468208092486, + "eval_allNLI-dev_manhattan_accuracy": 0.72265625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 599.6419677734375, + "eval_allNLI-dev_manhattan_ap": 0.6009156396302207, + "eval_allNLI-dev_manhattan_f1": 0.6400000000000001, + "eval_allNLI-dev_manhattan_f1_threshold": 763.5076904296875, + "eval_allNLI-dev_manhattan_precision": 0.5396825396825397, + "eval_allNLI-dev_manhattan_recall": 0.7861271676300579, + "eval_allNLI-dev_max_accuracy": 0.728515625, + "eval_allNLI-dev_max_accuracy_threshold": 743.5009765625, + "eval_allNLI-dev_max_ap": 0.6046964324835868, + "eval_allNLI-dev_max_f1": 0.6400000000000001, + "eval_allNLI-dev_max_f1_threshold": 763.5076904296875, + "eval_allNLI-dev_max_precision": 0.5403225806451613, + "eval_allNLI-dev_max_recall": 0.791907514450867, + "eval_sequential_score": 0.7478733392923025, + "eval_sts-test_pearson_cosine": 0.9138402547612807, + "eval_sts-test_pearson_dot": 0.9059823959468752, + "eval_sts-test_pearson_euclidean": 0.9303888224436159, + "eval_sts-test_pearson_manhattan": 0.9299076277394032, + "eval_sts-test_pearson_max": 0.9303888224436159, + "eval_sts-test_spearman_cosine": 0.926297498790823, + "eval_sts-test_spearman_dot": 0.9113238162317836, + "eval_sts-test_spearman_euclidean": 0.9254560408228805, + "eval_sts-test_spearman_manhattan": 0.9249534841574552, + "eval_sts-test_spearman_max": 0.926297498790823, + "eval_vitaminc-pairs_loss": 2.7577602863311768, + "eval_vitaminc-pairs_runtime": 4.5823, + "eval_vitaminc-pairs_samples_per_second": 27.933, + "eval_vitaminc-pairs_steps_per_second": 0.218, + "step": 442 + }, + { + "epoch": 0.8521850899742931, + "eval_negation-triplets_loss": 0.5208465456962585, + "eval_negation-triplets_runtime": 3.4824, + "eval_negation-triplets_samples_per_second": 36.756, + "eval_negation-triplets_steps_per_second": 0.287, + "step": 442 + }, + { + "epoch": 0.8521850899742931, + "eval_scitail-pairs-pos_loss": 0.00855955295264721, + "eval_scitail-pairs-pos_runtime": 2.7579, + "eval_scitail-pairs-pos_samples_per_second": 46.413, + "eval_scitail-pairs-pos_steps_per_second": 0.363, + "step": 442 + }, + { + "epoch": 0.8521850899742931, + "eval_scitail-pairs-qa_loss": 0.0023300235625356436, + "eval_scitail-pairs-qa_runtime": 2.3124, + "eval_scitail-pairs-qa_samples_per_second": 55.353, + "eval_scitail-pairs-qa_steps_per_second": 0.432, + "step": 442 + }, + { + "epoch": 0.8521850899742931, + "eval_xsum-pairs_loss": 0.006901867687702179, + "eval_xsum-pairs_runtime": 3.212, + "eval_xsum-pairs_samples_per_second": 39.851, + "eval_xsum-pairs_steps_per_second": 0.311, + "step": 442 + }, + { + "epoch": 0.8521850899742931, + "eval_sciq_pairs_loss": 0.0706043541431427, + "eval_sciq_pairs_runtime": 4.8786, + "eval_sciq_pairs_samples_per_second": 26.237, + "eval_sciq_pairs_steps_per_second": 0.205, + "step": 442 + }, + { + "epoch": 0.8521850899742931, + "eval_qasc_pairs_loss": 0.08733583986759186, + "eval_qasc_pairs_runtime": 2.1645, + "eval_qasc_pairs_samples_per_second": 59.137, + "eval_qasc_pairs_steps_per_second": 0.462, + "step": 442 + }, + { + "epoch": 0.8521850899742931, + "eval_openbookqa_pairs_loss": 0.3373841345310211, + "eval_openbookqa_pairs_runtime": 2.2848, + "eval_openbookqa_pairs_samples_per_second": 56.022, + "eval_openbookqa_pairs_steps_per_second": 0.438, + "step": 442 + }, + { + "epoch": 0.8521850899742931, + "eval_msmarco_pairs_loss": 0.01373821310698986, + "eval_msmarco_pairs_runtime": 2.3098, + "eval_msmarco_pairs_samples_per_second": 55.417, + "eval_msmarco_pairs_steps_per_second": 0.433, + "step": 442 + }, + { + "epoch": 0.8521850899742931, + "eval_nq_pairs_loss": 0.04490604251623154, + "eval_nq_pairs_runtime": 3.7868, + "eval_nq_pairs_samples_per_second": 33.802, + "eval_nq_pairs_steps_per_second": 0.264, + "step": 442 + }, + { + "epoch": 0.8521850899742931, + "eval_trivia_pairs_loss": 0.01734539307653904, + "eval_trivia_pairs_runtime": 3.1386, + "eval_trivia_pairs_samples_per_second": 40.783, + "eval_trivia_pairs_steps_per_second": 0.319, + "step": 442 + }, + { + "epoch": 0.8521850899742931, + "eval_gooaq_pairs_loss": 0.06459866464138031, + "eval_gooaq_pairs_runtime": 2.2089, + "eval_gooaq_pairs_samples_per_second": 57.948, + "eval_gooaq_pairs_steps_per_second": 0.453, + "step": 442 + }, + { + "epoch": 0.8521850899742931, + "eval_paws-pos_loss": 0.025955231860280037, + "eval_paws-pos_runtime": 2.2617, + "eval_paws-pos_samples_per_second": 56.594, + "eval_paws-pos_steps_per_second": 0.442, + "step": 442 + }, + { + "epoch": 0.8521850899742931, + "eval_global_dataset_loss": 0.145456463098526, + "eval_global_dataset_runtime": 8.4881, + "eval_global_dataset_samples_per_second": 38.289, + "eval_global_dataset_steps_per_second": 0.353, + "step": 442 + }, + { + "epoch": 0.8560411311053985, + "grad_norm": 6.109774589538574, + "learning_rate": 1.3379637173923707e-05, + "loss": 0.114, + "step": 444 + }, + { + "epoch": 0.8618251928020566, + "grad_norm": 3.212393283843994, + "learning_rate": 1.3326043210472238e-05, + "loss": 0.0713, + "step": 447 + }, + { + "epoch": 0.8676092544987146, + "grad_norm": 3.1123602390289307, + "learning_rate": 1.3271702419690156e-05, + "loss": 0.1336, + "step": 450 + }, + { + "epoch": 0.8733933161953727, + "grad_norm": 2.0358967781066895, + "learning_rate": 1.3216622796621244e-05, + "loss": 0.0612, + "step": 453 + }, + { + "epoch": 0.8791773778920309, + "grad_norm": 16.21796226501465, + "learning_rate": 1.3160812445012099e-05, + "loss": 0.1083, + "step": 456 + }, + { + "epoch": 0.8849614395886889, + "grad_norm": 2.596705675125122, + "learning_rate": 1.310427957611983e-05, + "loss": 0.0584, + "step": 459 + }, + { + "epoch": 0.890745501285347, + "grad_norm": 7.7204670906066895, + "learning_rate": 1.3047032507503966e-05, + "loss": 0.0929, + "step": 462 + }, + { + "epoch": 0.8965295629820051, + "grad_norm": 1.744662880897522, + "learning_rate": 1.2989079661802703e-05, + "loss": 0.066, + "step": 465 + }, + { + "epoch": 0.9023136246786633, + "grad_norm": 3.123241424560547, + "learning_rate": 1.2930429565493705e-05, + "loss": 0.0423, + "step": 468 + }, + { + "epoch": 0.9023136246786633, + "eval_Qnli-dev_cosine_accuracy": 0.712890625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6461004018783569, + "eval_Qnli-dev_cosine_ap": 0.7429638364715181, + "eval_Qnli-dev_cosine_f1": 0.7142857142857143, + "eval_Qnli-dev_cosine_f1_threshold": 0.5756168365478516, + "eval_Qnli-dev_cosine_precision": 0.5965909090909091, + "eval_Qnli-dev_cosine_recall": 0.8898305084745762, + "eval_Qnli-dev_dot_accuracy": 0.689453125, + "eval_Qnli-dev_dot_accuracy_threshold": 560.8919067382812, + "eval_Qnli-dev_dot_ap": 0.7158571398614493, + "eval_Qnli-dev_dot_f1": 0.698292220113852, + "eval_Qnli-dev_dot_f1_threshold": 560.8919067382812, + "eval_Qnli-dev_dot_precision": 0.6323024054982818, + "eval_Qnli-dev_dot_recall": 0.7796610169491526, + "eval_Qnli-dev_euclidean_accuracy": 0.716796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 23.726337432861328, + "eval_Qnli-dev_euclidean_ap": 0.7468184659568129, + "eval_Qnli-dev_euclidean_f1": 0.7103448275862068, + "eval_Qnli-dev_euclidean_f1_threshold": 27.682994842529297, + "eval_Qnli-dev_euclidean_precision": 0.5988372093023255, + "eval_Qnli-dev_euclidean_recall": 0.8728813559322034, + "eval_Qnli-dev_manhattan_accuracy": 0.71875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 729.7203369140625, + "eval_Qnli-dev_manhattan_ap": 0.7468995752921103, + "eval_Qnli-dev_manhattan_f1": 0.7122807017543861, + "eval_Qnli-dev_manhattan_f1_threshold": 845.4459228515625, + "eval_Qnli-dev_manhattan_precision": 0.6077844311377245, + "eval_Qnli-dev_manhattan_recall": 0.8601694915254238, + "eval_Qnli-dev_max_accuracy": 0.71875, + "eval_Qnli-dev_max_accuracy_threshold": 729.7203369140625, + "eval_Qnli-dev_max_ap": 0.7468995752921103, + "eval_Qnli-dev_max_f1": 0.7142857142857143, + "eval_Qnli-dev_max_f1_threshold": 845.4459228515625, + "eval_Qnli-dev_max_precision": 0.6323024054982818, + "eval_Qnli-dev_max_recall": 0.8898305084745762, + "eval_allNLI-dev_cosine_accuracy": 0.74609375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7913638353347778, + "eval_allNLI-dev_cosine_ap": 0.6147794488224912, + "eval_allNLI-dev_cosine_f1": 0.6318289786223278, + "eval_allNLI-dev_cosine_f1_threshold": 0.6930974721908569, + "eval_allNLI-dev_cosine_precision": 0.5362903225806451, + "eval_allNLI-dev_cosine_recall": 0.7687861271676301, + "eval_allNLI-dev_dot_accuracy": 0.75390625, + "eval_allNLI-dev_dot_accuracy_threshold": 700.6183471679688, + "eval_allNLI-dev_dot_ap": 0.6091421586452705, + "eval_allNLI-dev_dot_f1": 0.6189473684210527, + "eval_allNLI-dev_dot_f1_threshold": 583.320556640625, + "eval_allNLI-dev_dot_precision": 0.4867549668874172, + "eval_allNLI-dev_dot_recall": 0.8497109826589595, + "eval_allNLI-dev_euclidean_accuracy": 0.740234375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 19.20952606201172, + "eval_allNLI-dev_euclidean_ap": 0.6122761202106319, + "eval_allNLI-dev_euclidean_f1": 0.6340326340326341, + "eval_allNLI-dev_euclidean_f1_threshold": 23.720947265625, + "eval_allNLI-dev_euclidean_precision": 0.53125, + "eval_allNLI-dev_euclidean_recall": 0.7861271676300579, + "eval_allNLI-dev_manhattan_accuracy": 0.734375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 600.646240234375, + "eval_allNLI-dev_manhattan_ap": 0.6131254272679774, + "eval_allNLI-dev_manhattan_f1": 0.6327944572748269, + "eval_allNLI-dev_manhattan_f1_threshold": 743.770263671875, + "eval_allNLI-dev_manhattan_precision": 0.5269230769230769, + "eval_allNLI-dev_manhattan_recall": 0.791907514450867, + "eval_allNLI-dev_max_accuracy": 0.75390625, + "eval_allNLI-dev_max_accuracy_threshold": 700.6183471679688, + "eval_allNLI-dev_max_ap": 0.6147794488224912, + "eval_allNLI-dev_max_f1": 0.6340326340326341, + "eval_allNLI-dev_max_f1_threshold": 743.770263671875, + "eval_allNLI-dev_max_precision": 0.5362903225806451, + "eval_allNLI-dev_max_recall": 0.8497109826589595, + "eval_sequential_score": 0.7468995752921103, + "eval_sts-test_pearson_cosine": 0.9159220241890097, + "eval_sts-test_pearson_dot": 0.9052308545844396, + "eval_sts-test_pearson_euclidean": 0.9313844833923889, + "eval_sts-test_pearson_manhattan": 0.9306550150689094, + "eval_sts-test_pearson_max": 0.9313844833923889, + "eval_sts-test_spearman_cosine": 0.9273589092085042, + "eval_sts-test_spearman_dot": 0.9088352880883306, + "eval_sts-test_spearman_euclidean": 0.9273073110041199, + "eval_sts-test_spearman_manhattan": 0.9268169266817929, + "eval_sts-test_spearman_max": 0.9273589092085042, + "eval_vitaminc-pairs_loss": 2.987377166748047, + "eval_vitaminc-pairs_runtime": 4.4371, + "eval_vitaminc-pairs_samples_per_second": 28.848, + "eval_vitaminc-pairs_steps_per_second": 0.225, + "step": 468 + }, + { + "epoch": 0.9023136246786633, + "eval_negation-triplets_loss": 0.5298277735710144, + "eval_negation-triplets_runtime": 3.2754, + "eval_negation-triplets_samples_per_second": 39.079, + "eval_negation-triplets_steps_per_second": 0.305, + "step": 468 + }, + { + "epoch": 0.9023136246786633, + "eval_scitail-pairs-pos_loss": 0.008193486370146275, + "eval_scitail-pairs-pos_runtime": 2.707, + "eval_scitail-pairs-pos_samples_per_second": 47.285, + "eval_scitail-pairs-pos_steps_per_second": 0.369, + "step": 468 + }, + { + "epoch": 0.9023136246786633, + "eval_scitail-pairs-qa_loss": 0.003911494743078947, + "eval_scitail-pairs-qa_runtime": 2.2325, + "eval_scitail-pairs-qa_samples_per_second": 57.334, + "eval_scitail-pairs-qa_steps_per_second": 0.448, + "step": 468 + }, + { + "epoch": 0.9023136246786633, + "eval_xsum-pairs_loss": 0.003330780193209648, + "eval_xsum-pairs_runtime": 3.0861, + "eval_xsum-pairs_samples_per_second": 41.476, + "eval_xsum-pairs_steps_per_second": 0.324, + "step": 468 + }, + { + "epoch": 0.9023136246786633, + "eval_sciq_pairs_loss": 0.06425436586141586, + "eval_sciq_pairs_runtime": 4.8256, + "eval_sciq_pairs_samples_per_second": 26.525, + "eval_sciq_pairs_steps_per_second": 0.207, + "step": 468 + }, + { + "epoch": 0.9023136246786633, + "eval_qasc_pairs_loss": 0.06486833840608597, + "eval_qasc_pairs_runtime": 2.1633, + "eval_qasc_pairs_samples_per_second": 59.17, + "eval_qasc_pairs_steps_per_second": 0.462, + "step": 468 + }, + { + "epoch": 0.9023136246786633, + "eval_openbookqa_pairs_loss": 0.34340742230415344, + "eval_openbookqa_pairs_runtime": 2.225, + "eval_openbookqa_pairs_samples_per_second": 57.527, + "eval_openbookqa_pairs_steps_per_second": 0.449, + "step": 468 + }, + { + "epoch": 0.9023136246786633, + "eval_msmarco_pairs_loss": 0.00849138479679823, + "eval_msmarco_pairs_runtime": 2.2296, + "eval_msmarco_pairs_samples_per_second": 57.409, + "eval_msmarco_pairs_steps_per_second": 0.449, + "step": 468 + }, + { + "epoch": 0.9023136246786633, + "eval_nq_pairs_loss": 0.03956576809287071, + "eval_nq_pairs_runtime": 3.6403, + "eval_nq_pairs_samples_per_second": 35.162, + "eval_nq_pairs_steps_per_second": 0.275, + "step": 468 + }, + { + "epoch": 0.9023136246786633, + "eval_trivia_pairs_loss": 0.018078835681080818, + "eval_trivia_pairs_runtime": 3.0876, + "eval_trivia_pairs_samples_per_second": 41.456, + "eval_trivia_pairs_steps_per_second": 0.324, + "step": 468 + }, + { + "epoch": 0.9023136246786633, + "eval_gooaq_pairs_loss": 0.07264826446771622, + "eval_gooaq_pairs_runtime": 2.1431, + "eval_gooaq_pairs_samples_per_second": 59.727, + "eval_gooaq_pairs_steps_per_second": 0.467, + "step": 468 + }, + { + "epoch": 0.9023136246786633, + "eval_paws-pos_loss": 0.026289092376828194, + "eval_paws-pos_runtime": 2.2416, + "eval_paws-pos_samples_per_second": 57.102, + "eval_paws-pos_steps_per_second": 0.446, + "step": 468 + }, + { + "epoch": 0.9023136246786633, + "eval_global_dataset_loss": 0.1634972095489502, + "eval_global_dataset_runtime": 8.3728, + "eval_global_dataset_samples_per_second": 38.816, + "eval_global_dataset_steps_per_second": 0.358, + "step": 468 + }, + { + "epoch": 0.9080976863753213, + "grad_norm": 3.4135470390319824, + "learning_rate": 1.2871090847639623e-05, + "loss": 0.0809, + "step": 471 + }, + { + "epoch": 0.9138817480719794, + "grad_norm": 9.212980270385742, + "learning_rate": 1.2811072238618512e-05, + "loss": 0.3409, + "step": 474 + }, + { + "epoch": 0.9196658097686375, + "grad_norm": 3.3372764587402344, + "learning_rate": 1.2750382568839355e-05, + "loss": 0.0726, + "step": 477 + }, + { + "epoch": 0.9254498714652957, + "grad_norm": 4.139834403991699, + "learning_rate": 1.2689030767442859e-05, + "loss": 0.101, + "step": 480 + }, + { + "epoch": 0.9312339331619537, + "grad_norm": 4.122523307800293, + "learning_rate": 1.2627025860987738e-05, + "loss": 0.0944, + "step": 483 + }, + { + "epoch": 0.9370179948586118, + "grad_norm": 4.818490982055664, + "learning_rate": 1.2564376972122648e-05, + "loss": 0.0431, + "step": 486 + }, + { + "epoch": 0.9428020565552699, + "grad_norm": 3.818354368209839, + "learning_rate": 1.2501093318243996e-05, + "loss": 0.1061, + "step": 489 + }, + { + "epoch": 0.9485861182519281, + "grad_norm": 3.3825314044952393, + "learning_rate": 1.2437184210139797e-05, + "loss": 0.0843, + "step": 492 + }, + { + "epoch": 0.9524421593830334, + "eval_Qnli-dev_cosine_accuracy": 0.70703125, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.7167578339576721, + "eval_Qnli-dev_cosine_ap": 0.7450139524438387, + "eval_Qnli-dev_cosine_f1": 0.7066895368782161, + "eval_Qnli-dev_cosine_f1_threshold": 0.5705962181091309, + "eval_Qnli-dev_cosine_precision": 0.5936599423631124, + "eval_Qnli-dev_cosine_recall": 0.8728813559322034, + "eval_Qnli-dev_dot_accuracy": 0.69140625, + "eval_Qnli-dev_dot_accuracy_threshold": 615.7021484375, + "eval_Qnli-dev_dot_ap": 0.7160223929256964, + "eval_Qnli-dev_dot_f1": 0.700348432055749, + "eval_Qnli-dev_dot_f1_threshold": 524.7015380859375, + "eval_Qnli-dev_dot_precision": 0.5946745562130178, + "eval_Qnli-dev_dot_recall": 0.8516949152542372, + "eval_Qnli-dev_euclidean_accuracy": 0.7109375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 23.455307006835938, + "eval_Qnli-dev_euclidean_ap": 0.7526592716500206, + "eval_Qnli-dev_euclidean_f1": 0.7093235831809872, + "eval_Qnli-dev_euclidean_f1_threshold": 26.879047393798828, + "eval_Qnli-dev_euclidean_precision": 0.6237942122186495, + "eval_Qnli-dev_euclidean_recall": 0.8220338983050848, + "eval_Qnli-dev_manhattan_accuracy": 0.7109375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 711.6500244140625, + "eval_Qnli-dev_manhattan_ap": 0.7530792172375838, + "eval_Qnli-dev_manhattan_f1": 0.7107750472589792, + "eval_Qnli-dev_manhattan_f1_threshold": 821.3135986328125, + "eval_Qnli-dev_manhattan_precision": 0.6416382252559727, + "eval_Qnli-dev_manhattan_recall": 0.7966101694915254, + "eval_Qnli-dev_max_accuracy": 0.7109375, + "eval_Qnli-dev_max_accuracy_threshold": 711.6500244140625, + "eval_Qnli-dev_max_ap": 0.7530792172375838, + "eval_Qnli-dev_max_f1": 0.7107750472589792, + "eval_Qnli-dev_max_f1_threshold": 821.3135986328125, + "eval_Qnli-dev_max_precision": 0.6416382252559727, + "eval_Qnli-dev_max_recall": 0.8728813559322034, + "eval_allNLI-dev_cosine_accuracy": 0.724609375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7712951898574829, + "eval_allNLI-dev_cosine_ap": 0.6069995438494482, + "eval_allNLI-dev_cosine_f1": 0.644880174291939, + "eval_allNLI-dev_cosine_f1_threshold": 0.614275336265564, + "eval_allNLI-dev_cosine_precision": 0.5174825174825175, + "eval_allNLI-dev_cosine_recall": 0.8554913294797688, + "eval_allNLI-dev_dot_accuracy": 0.740234375, + "eval_allNLI-dev_dot_accuracy_threshold": 670.4595336914062, + "eval_allNLI-dev_dot_ap": 0.6148264061774873, + "eval_allNLI-dev_dot_f1": 0.6363636363636364, + "eval_allNLI-dev_dot_f1_threshold": 541.9984741210938, + "eval_allNLI-dev_dot_precision": 0.49517684887459806, + "eval_allNLI-dev_dot_recall": 0.8901734104046243, + "eval_allNLI-dev_euclidean_accuracy": 0.720703125, + "eval_allNLI-dev_euclidean_accuracy_threshold": 22.52337646484375, + "eval_allNLI-dev_euclidean_ap": 0.6033858475727263, + "eval_allNLI-dev_euclidean_f1": 0.6389496717724288, + "eval_allNLI-dev_euclidean_f1_threshold": 26.440635681152344, + "eval_allNLI-dev_euclidean_precision": 0.5140845070422535, + "eval_allNLI-dev_euclidean_recall": 0.8439306358381503, + "eval_allNLI-dev_manhattan_accuracy": 0.724609375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 702.60302734375, + "eval_allNLI-dev_manhattan_ap": 0.60446166181456, + "eval_allNLI-dev_manhattan_f1": 0.644880174291939, + "eval_allNLI-dev_manhattan_f1_threshold": 825.4560546875, + "eval_allNLI-dev_manhattan_precision": 0.5174825174825175, + "eval_allNLI-dev_manhattan_recall": 0.8554913294797688, + "eval_allNLI-dev_max_accuracy": 0.740234375, + "eval_allNLI-dev_max_accuracy_threshold": 702.60302734375, + "eval_allNLI-dev_max_ap": 0.6148264061774873, + "eval_allNLI-dev_max_f1": 0.644880174291939, + "eval_allNLI-dev_max_f1_threshold": 825.4560546875, + "eval_allNLI-dev_max_precision": 0.5174825174825175, + "eval_allNLI-dev_max_recall": 0.8901734104046243, + "eval_sequential_score": 0.7530792172375838, + "eval_sts-test_pearson_cosine": 0.9186491762470915, + "eval_sts-test_pearson_dot": 0.9080249596015681, + "eval_sts-test_pearson_euclidean": 0.9296691988330295, + "eval_sts-test_pearson_manhattan": 0.9290762237427987, + "eval_sts-test_pearson_max": 0.9296691988330295, + "eval_sts-test_spearman_cosine": 0.9238488440212697, + "eval_sts-test_spearman_dot": 0.905394945366344, + "eval_sts-test_spearman_euclidean": 0.9232121678254545, + "eval_sts-test_spearman_manhattan": 0.9227415635606625, + "eval_sts-test_spearman_max": 0.9238488440212697, + "eval_vitaminc-pairs_loss": 3.5028598308563232, + "eval_vitaminc-pairs_runtime": 4.5728, + "eval_vitaminc-pairs_samples_per_second": 27.991, + "eval_vitaminc-pairs_steps_per_second": 0.219, + "step": 494 + }, + { + "epoch": 0.9524421593830334, + "eval_negation-triplets_loss": 0.4938853681087494, + "eval_negation-triplets_runtime": 3.3957, + "eval_negation-triplets_samples_per_second": 37.695, + "eval_negation-triplets_steps_per_second": 0.294, + "step": 494 + }, + { + "epoch": 0.9524421593830334, + "eval_scitail-pairs-pos_loss": 0.0071646529249846935, + "eval_scitail-pairs-pos_runtime": 2.76, + "eval_scitail-pairs-pos_samples_per_second": 46.377, + "eval_scitail-pairs-pos_steps_per_second": 0.362, + "step": 494 + }, + { + "epoch": 0.9524421593830334, + "eval_scitail-pairs-qa_loss": 0.012059859931468964, + "eval_scitail-pairs-qa_runtime": 2.3694, + "eval_scitail-pairs-qa_samples_per_second": 54.021, + "eval_scitail-pairs-qa_steps_per_second": 0.422, + "step": 494 + }, + { + "epoch": 0.9524421593830334, + "eval_xsum-pairs_loss": 0.0036885663866996765, + "eval_xsum-pairs_runtime": 3.1007, + "eval_xsum-pairs_samples_per_second": 41.282, + "eval_xsum-pairs_steps_per_second": 0.323, + "step": 494 + }, + { + "epoch": 0.9524421593830334, + "eval_sciq_pairs_loss": 0.06713753193616867, + "eval_sciq_pairs_runtime": 4.8664, + "eval_sciq_pairs_samples_per_second": 26.303, + "eval_sciq_pairs_steps_per_second": 0.205, + "step": 494 + }, + { + "epoch": 0.9524421593830334, + "eval_qasc_pairs_loss": 0.07108274102210999, + "eval_qasc_pairs_runtime": 2.086, + "eval_qasc_pairs_samples_per_second": 61.362, + "eval_qasc_pairs_steps_per_second": 0.479, + "step": 494 + }, + { + "epoch": 0.9524421593830334, + "eval_openbookqa_pairs_loss": 0.3318020701408386, + "eval_openbookqa_pairs_runtime": 2.1907, + "eval_openbookqa_pairs_samples_per_second": 58.428, + "eval_openbookqa_pairs_steps_per_second": 0.456, + "step": 494 + }, + { + "epoch": 0.9524421593830334, + "eval_msmarco_pairs_loss": 0.0052869198843836784, + "eval_msmarco_pairs_runtime": 2.3169, + "eval_msmarco_pairs_samples_per_second": 55.246, + "eval_msmarco_pairs_steps_per_second": 0.432, + "step": 494 + }, + { + "epoch": 0.9524421593830334, + "eval_nq_pairs_loss": 0.021715562790632248, + "eval_nq_pairs_runtime": 3.6502, + "eval_nq_pairs_samples_per_second": 35.066, + "eval_nq_pairs_steps_per_second": 0.274, + "step": 494 + }, + { + "epoch": 0.9524421593830334, + "eval_trivia_pairs_loss": 0.013279886916279793, + "eval_trivia_pairs_runtime": 3.1362, + "eval_trivia_pairs_samples_per_second": 40.813, + "eval_trivia_pairs_steps_per_second": 0.319, + "step": 494 + }, + { + "epoch": 0.9524421593830334, + "eval_gooaq_pairs_loss": 0.07737365365028381, + "eval_gooaq_pairs_runtime": 2.1849, + "eval_gooaq_pairs_samples_per_second": 58.583, + "eval_gooaq_pairs_steps_per_second": 0.458, + "step": 494 + }, + { + "epoch": 0.9524421593830334, + "eval_paws-pos_loss": 0.02541165053844452, + "eval_paws-pos_runtime": 2.3321, + "eval_paws-pos_samples_per_second": 54.885, + "eval_paws-pos_steps_per_second": 0.429, + "step": 494 + }, + { + "epoch": 0.9524421593830334, + "eval_global_dataset_loss": 0.20370937883853912, + "eval_global_dataset_runtime": 8.4936, + "eval_global_dataset_samples_per_second": 38.264, + "eval_global_dataset_steps_per_second": 0.353, + "step": 494 + }, + { + "epoch": 0.9543701799485861, + "grad_norm": 2.5691542625427246, + "learning_rate": 1.2372659050619804e-05, + "loss": 0.1177, + "step": 495 + }, + { + "epoch": 0.9601542416452442, + "grad_norm": 2.1602962017059326, + "learning_rate": 1.230752733313209e-05, + "loss": 0.1513, + "step": 498 + }, + { + "epoch": 0.9659383033419023, + "grad_norm": 5.229061126708984, + "learning_rate": 1.2241798640366303e-05, + "loss": 0.1944, + "step": 501 + }, + { + "epoch": 0.9717223650385605, + "grad_norm": 1.516933560371399, + "learning_rate": 1.2175482642843778e-05, + "loss": 0.0969, + "step": 504 + }, + { + "epoch": 0.9775064267352185, + "grad_norm": 4.720734119415283, + "learning_rate": 1.2108589097494732e-05, + "loss": 0.1293, + "step": 507 + }, + { + "epoch": 0.9832904884318766, + "grad_norm": 6.196876049041748, + "learning_rate": 1.2041127846222761e-05, + "loss": 0.15, + "step": 510 + }, + { + "epoch": 0.9890745501285347, + "grad_norm": 4.537347316741943, + "learning_rate": 1.1973108814456808e-05, + "loss": 0.046, + "step": 513 + }, + { + "epoch": 0.9948586118251928, + "grad_norm": 2.640577793121338, + "learning_rate": 1.1904542009690866e-05, + "loss": 0.1007, + "step": 516 + }, + { + "epoch": 1.000642673521851, + "grad_norm": 1.3155243396759033, + "learning_rate": 1.183543752001159e-05, + "loss": 0.0359, + "step": 519 + }, + { + "epoch": 1.0025706940874035, + "eval_Qnli-dev_cosine_accuracy": 0.712890625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6758573055267334, + "eval_Qnli-dev_cosine_ap": 0.7532921492091731, + "eval_Qnli-dev_cosine_f1": 0.7227191413237924, + "eval_Qnli-dev_cosine_f1_threshold": 0.6170268058776855, + "eval_Qnli-dev_cosine_precision": 0.6253869969040248, + "eval_Qnli-dev_cosine_recall": 0.8559322033898306, + "eval_Qnli-dev_dot_accuracy": 0.701171875, + "eval_Qnli-dev_dot_accuracy_threshold": 683.794921875, + "eval_Qnli-dev_dot_ap": 0.7244801987603469, + "eval_Qnli-dev_dot_f1": 0.7104795737122558, + "eval_Qnli-dev_dot_f1_threshold": 583.2965698242188, + "eval_Qnli-dev_dot_precision": 0.6116207951070336, + "eval_Qnli-dev_dot_recall": 0.847457627118644, + "eval_Qnli-dev_euclidean_accuracy": 0.716796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 24.123794555664062, + "eval_Qnli-dev_euclidean_ap": 0.7565081721110317, + "eval_Qnli-dev_euclidean_f1": 0.7179487179487178, + "eval_Qnli-dev_euclidean_f1_threshold": 26.894371032714844, + "eval_Qnli-dev_euclidean_precision": 0.632258064516129, + "eval_Qnli-dev_euclidean_recall": 0.8305084745762712, + "eval_Qnli-dev_manhattan_accuracy": 0.71484375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 716.9512939453125, + "eval_Qnli-dev_manhattan_ap": 0.7571063694133022, + "eval_Qnli-dev_manhattan_f1": 0.7232472324723246, + "eval_Qnli-dev_manhattan_f1_threshold": 835.429931640625, + "eval_Qnli-dev_manhattan_precision": 0.6405228758169934, + "eval_Qnli-dev_manhattan_recall": 0.8305084745762712, + "eval_Qnli-dev_max_accuracy": 0.716796875, + "eval_Qnli-dev_max_accuracy_threshold": 716.9512939453125, + "eval_Qnli-dev_max_ap": 0.7571063694133022, + "eval_Qnli-dev_max_f1": 0.7232472324723246, + "eval_Qnli-dev_max_f1_threshold": 835.429931640625, + "eval_Qnli-dev_max_precision": 0.6405228758169934, + "eval_Qnli-dev_max_recall": 0.8559322033898306, + "eval_allNLI-dev_cosine_accuracy": 0.72265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8173953294754028, + "eval_allNLI-dev_cosine_ap": 0.6081718667776024, + "eval_allNLI-dev_cosine_f1": 0.6428571428571429, + "eval_allNLI-dev_cosine_f1_threshold": 0.6594400405883789, + "eval_allNLI-dev_cosine_precision": 0.5236363636363637, + "eval_allNLI-dev_cosine_recall": 0.8323699421965318, + "eval_allNLI-dev_dot_accuracy": 0.734375, + "eval_allNLI-dev_dot_accuracy_threshold": 752.1400146484375, + "eval_allNLI-dev_dot_ap": 0.6091788780183418, + "eval_allNLI-dev_dot_f1": 0.6403508771929824, + "eval_allNLI-dev_dot_f1_threshold": 627.6685791015625, + "eval_allNLI-dev_dot_precision": 0.5159010600706714, + "eval_allNLI-dev_dot_recall": 0.8439306358381503, + "eval_allNLI-dev_euclidean_accuracy": 0.724609375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 18.706993103027344, + "eval_allNLI-dev_euclidean_ap": 0.6063452720455431, + "eval_allNLI-dev_euclidean_f1": 0.6445916114790288, + "eval_allNLI-dev_euclidean_f1_threshold": 25.848369598388672, + "eval_allNLI-dev_euclidean_precision": 0.5214285714285715, + "eval_allNLI-dev_euclidean_recall": 0.8439306358381503, + "eval_allNLI-dev_manhattan_accuracy": 0.728515625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 578.2825317382812, + "eval_allNLI-dev_manhattan_ap": 0.606964832099393, + "eval_allNLI-dev_manhattan_f1": 0.6447368421052632, + "eval_allNLI-dev_manhattan_f1_threshold": 809.925048828125, + "eval_allNLI-dev_manhattan_precision": 0.519434628975265, + "eval_allNLI-dev_manhattan_recall": 0.8497109826589595, + "eval_allNLI-dev_max_accuracy": 0.734375, + "eval_allNLI-dev_max_accuracy_threshold": 752.1400146484375, + "eval_allNLI-dev_max_ap": 0.6091788780183418, + "eval_allNLI-dev_max_f1": 0.6447368421052632, + "eval_allNLI-dev_max_f1_threshold": 809.925048828125, + "eval_allNLI-dev_max_precision": 0.5236363636363637, + "eval_allNLI-dev_max_recall": 0.8497109826589595, + "eval_sequential_score": 0.7571063694133022, + "eval_sts-test_pearson_cosine": 0.9192188691197744, + "eval_sts-test_pearson_dot": 0.9100833111963467, + "eval_sts-test_pearson_euclidean": 0.9354587253751946, + "eval_sts-test_pearson_manhattan": 0.9345466766955768, + "eval_sts-test_pearson_max": 0.9354587253751946, + "eval_sts-test_spearman_cosine": 0.9309495266662431, + "eval_sts-test_spearman_dot": 0.9116208303536382, + "eval_sts-test_spearman_euclidean": 0.9303286029144374, + "eval_sts-test_spearman_manhattan": 0.9292958333040467, + "eval_sts-test_spearman_max": 0.9309495266662431, + "eval_vitaminc-pairs_loss": 3.200331926345825, + "eval_vitaminc-pairs_runtime": 4.5357, + "eval_vitaminc-pairs_samples_per_second": 28.22, + "eval_vitaminc-pairs_steps_per_second": 0.22, + "step": 520 + }, + { + "epoch": 1.0025706940874035, + "eval_negation-triplets_loss": 0.5108892321586609, + "eval_negation-triplets_runtime": 3.3008, + "eval_negation-triplets_samples_per_second": 38.778, + "eval_negation-triplets_steps_per_second": 0.303, + "step": 520 + }, + { + "epoch": 1.0025706940874035, + "eval_scitail-pairs-pos_loss": 0.015103669837117195, + "eval_scitail-pairs-pos_runtime": 2.6781, + "eval_scitail-pairs-pos_samples_per_second": 47.795, + "eval_scitail-pairs-pos_steps_per_second": 0.373, + "step": 520 + }, + { + "epoch": 1.0025706940874035, + "eval_scitail-pairs-qa_loss": 0.01923208124935627, + "eval_scitail-pairs-qa_runtime": 2.228, + "eval_scitail-pairs-qa_samples_per_second": 57.45, + "eval_scitail-pairs-qa_steps_per_second": 0.449, + "step": 520 + }, + { + "epoch": 1.0025706940874035, + "eval_xsum-pairs_loss": 0.008190438151359558, + "eval_xsum-pairs_runtime": 3.1044, + "eval_xsum-pairs_samples_per_second": 41.232, + "eval_xsum-pairs_steps_per_second": 0.322, + "step": 520 + }, + { + "epoch": 1.0025706940874035, + "eval_sciq_pairs_loss": 0.0655350387096405, + "eval_sciq_pairs_runtime": 4.7915, + "eval_sciq_pairs_samples_per_second": 26.714, + "eval_sciq_pairs_steps_per_second": 0.209, + "step": 520 + }, + { + "epoch": 1.0025706940874035, + "eval_qasc_pairs_loss": 0.03946835547685623, + "eval_qasc_pairs_runtime": 2.0897, + "eval_qasc_pairs_samples_per_second": 61.252, + "eval_qasc_pairs_steps_per_second": 0.479, + "step": 520 + }, + { + "epoch": 1.0025706940874035, + "eval_openbookqa_pairs_loss": 0.349005788564682, + "eval_openbookqa_pairs_runtime": 2.2285, + "eval_openbookqa_pairs_samples_per_second": 57.438, + "eval_openbookqa_pairs_steps_per_second": 0.449, + "step": 520 + }, + { + "epoch": 1.0025706940874035, + "eval_msmarco_pairs_loss": 0.01249724067747593, + "eval_msmarco_pairs_runtime": 2.2794, + "eval_msmarco_pairs_samples_per_second": 56.155, + "eval_msmarco_pairs_steps_per_second": 0.439, + "step": 520 + }, + { + "epoch": 1.0025706940874035, + "eval_nq_pairs_loss": 0.03227410092949867, + "eval_nq_pairs_runtime": 3.6412, + "eval_nq_pairs_samples_per_second": 35.153, + "eval_nq_pairs_steps_per_second": 0.275, + "step": 520 + }, + { + "epoch": 1.0025706940874035, + "eval_trivia_pairs_loss": 0.021581444889307022, + "eval_trivia_pairs_runtime": 3.0866, + "eval_trivia_pairs_samples_per_second": 41.469, + "eval_trivia_pairs_steps_per_second": 0.324, + "step": 520 + }, + { + "epoch": 1.0025706940874035, + "eval_gooaq_pairs_loss": 0.06554285436868668, + "eval_gooaq_pairs_runtime": 2.1345, + "eval_gooaq_pairs_samples_per_second": 59.966, + "eval_gooaq_pairs_steps_per_second": 0.468, + "step": 520 + }, + { + "epoch": 1.0025706940874035, + "eval_paws-pos_loss": 0.02683321386575699, + "eval_paws-pos_runtime": 2.3595, + "eval_paws-pos_samples_per_second": 54.25, + "eval_paws-pos_steps_per_second": 0.424, + "step": 520 + }, + { + "epoch": 1.0025706940874035, + "eval_global_dataset_loss": 0.1803271323442459, + "eval_global_dataset_runtime": 8.3879, + "eval_global_dataset_samples_per_second": 38.746, + "eval_global_dataset_steps_per_second": 0.358, + "step": 520 + }, + { + "epoch": 1.006426735218509, + "grad_norm": 1.69589102268219, + "learning_rate": 1.1765805512614057e-05, + "loss": 0.0547, + "step": 522 + }, + { + "epoch": 1.012210796915167, + "grad_norm": 1.3169803619384766, + "learning_rate": 1.1695656232305901e-05, + "loss": 0.0669, + "step": 525 + }, + { + "epoch": 1.0179948586118253, + "grad_norm": 6.317488193511963, + "learning_rate": 1.1625000000000001e-05, + "loss": 0.0965, + "step": 528 + }, + { + "epoch": 1.0237789203084833, + "grad_norm": 2.3865275382995605, + "learning_rate": 1.1553847211195996e-05, + "loss": 0.1971, + "step": 531 + }, + { + "epoch": 1.0295629820051413, + "grad_norm": 2.723081350326538, + "learning_rate": 1.1482208334450817e-05, + "loss": 0.0637, + "step": 534 + }, + { + "epoch": 1.0353470437017995, + "grad_norm": 2.2263810634613037, + "learning_rate": 1.141009390983847e-05, + "loss": 0.0341, + "step": 537 + }, + { + "epoch": 1.0411311053984575, + "grad_norm": 2.5484347343444824, + "learning_rate": 1.1337514547399292e-05, + "loss": 0.0513, + "step": 540 + }, + { + "epoch": 1.0469151670951158, + "grad_norm": 1.4160079956054688, + "learning_rate": 1.126448092557893e-05, + "loss": 0.0435, + "step": 543 + }, + { + "epoch": 1.0526992287917738, + "grad_norm": 71.0225601196289, + "learning_rate": 1.119100378965723e-05, + "loss": 0.2092, + "step": 546 + }, + { + "epoch": 1.0526992287917738, + "eval_Qnli-dev_cosine_accuracy": 0.69921875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6603677868843079, + "eval_Qnli-dev_cosine_ap": 0.7396221999476175, + "eval_Qnli-dev_cosine_f1": 0.7132867132867132, + "eval_Qnli-dev_cosine_f1_threshold": 0.6077866554260254, + "eval_Qnli-dev_cosine_precision": 0.6071428571428571, + "eval_Qnli-dev_cosine_recall": 0.864406779661017, + "eval_Qnli-dev_dot_accuracy": 0.685546875, + "eval_Qnli-dev_dot_accuracy_threshold": 623.8295288085938, + "eval_Qnli-dev_dot_ap": 0.7098926596216567, + "eval_Qnli-dev_dot_f1": 0.7038917089678511, + "eval_Qnli-dev_dot_f1_threshold": 560.6802978515625, + "eval_Qnli-dev_dot_precision": 0.5859154929577465, + "eval_Qnli-dev_dot_recall": 0.8813559322033898, + "eval_Qnli-dev_euclidean_accuracy": 0.712890625, + "eval_Qnli-dev_euclidean_accuracy_threshold": 23.495559692382812, + "eval_Qnli-dev_euclidean_ap": 0.7427326753063308, + "eval_Qnli-dev_euclidean_f1": 0.7018121911037892, + "eval_Qnli-dev_euclidean_f1_threshold": 28.43372344970703, + "eval_Qnli-dev_euclidean_precision": 0.5741239892183289, + "eval_Qnli-dev_euclidean_recall": 0.902542372881356, + "eval_Qnli-dev_manhattan_accuracy": 0.71484375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 734.5736083984375, + "eval_Qnli-dev_manhattan_ap": 0.7440872441738853, + "eval_Qnli-dev_manhattan_f1": 0.7052980132450332, + "eval_Qnli-dev_manhattan_f1_threshold": 883.763671875, + "eval_Qnli-dev_manhattan_precision": 0.5788043478260869, + "eval_Qnli-dev_manhattan_recall": 0.902542372881356, + "eval_Qnli-dev_max_accuracy": 0.71484375, + "eval_Qnli-dev_max_accuracy_threshold": 734.5736083984375, + "eval_Qnli-dev_max_ap": 0.7440872441738853, + "eval_Qnli-dev_max_f1": 0.7132867132867132, + "eval_Qnli-dev_max_f1_threshold": 883.763671875, + "eval_Qnli-dev_max_precision": 0.6071428571428571, + "eval_Qnli-dev_max_recall": 0.902542372881356, + "eval_allNLI-dev_cosine_accuracy": 0.720703125, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8185880184173584, + "eval_allNLI-dev_cosine_ap": 0.6017568840142871, + "eval_allNLI-dev_cosine_f1": 0.6313465783664459, + "eval_allNLI-dev_cosine_f1_threshold": 0.6495623588562012, + "eval_allNLI-dev_cosine_precision": 0.5107142857142857, + "eval_allNLI-dev_cosine_recall": 0.8265895953757225, + "eval_allNLI-dev_dot_accuracy": 0.724609375, + "eval_allNLI-dev_dot_accuracy_threshold": 798.8133544921875, + "eval_allNLI-dev_dot_ap": 0.6121296986834653, + "eval_allNLI-dev_dot_f1": 0.6247191011235955, + "eval_allNLI-dev_dot_f1_threshold": 646.3211669921875, + "eval_allNLI-dev_dot_precision": 0.5110294117647058, + "eval_allNLI-dev_dot_recall": 0.8034682080924855, + "eval_allNLI-dev_euclidean_accuracy": 0.716796875, + "eval_allNLI-dev_euclidean_accuracy_threshold": 18.009429931640625, + "eval_allNLI-dev_euclidean_ap": 0.5951106462557548, + "eval_allNLI-dev_euclidean_f1": 0.6347031963470319, + "eval_allNLI-dev_euclidean_f1_threshold": 25.74187660217285, + "eval_allNLI-dev_euclidean_precision": 0.5245283018867924, + "eval_allNLI-dev_euclidean_recall": 0.8034682080924855, + "eval_allNLI-dev_manhattan_accuracy": 0.716796875, + "eval_allNLI-dev_manhattan_accuracy_threshold": 562.3073120117188, + "eval_allNLI-dev_manhattan_ap": 0.5960484641619748, + "eval_allNLI-dev_manhattan_f1": 0.6382022471910113, + "eval_allNLI-dev_manhattan_f1_threshold": 806.8055419921875, + "eval_allNLI-dev_manhattan_precision": 0.5220588235294118, + "eval_allNLI-dev_manhattan_recall": 0.8208092485549133, + "eval_allNLI-dev_max_accuracy": 0.724609375, + "eval_allNLI-dev_max_accuracy_threshold": 798.8133544921875, + "eval_allNLI-dev_max_ap": 0.6121296986834653, + "eval_allNLI-dev_max_f1": 0.6382022471910113, + "eval_allNLI-dev_max_f1_threshold": 806.8055419921875, + "eval_allNLI-dev_max_precision": 0.5245283018867924, + "eval_allNLI-dev_max_recall": 0.8265895953757225, + "eval_sequential_score": 0.7440872441738853, + "eval_sts-test_pearson_cosine": 0.9139137265176375, + "eval_sts-test_pearson_dot": 0.904456189910613, + "eval_sts-test_pearson_euclidean": 0.9315973143536354, + "eval_sts-test_pearson_manhattan": 0.9309090818905605, + "eval_sts-test_pearson_max": 0.9315973143536354, + "eval_sts-test_spearman_cosine": 0.9259321047415623, + "eval_sts-test_spearman_dot": 0.9085682125603467, + "eval_sts-test_spearman_euclidean": 0.9255636676065268, + "eval_sts-test_spearman_manhattan": 0.9248692175322552, + "eval_sts-test_spearman_max": 0.9259321047415623, + "eval_vitaminc-pairs_loss": 3.7179458141326904, + "eval_vitaminc-pairs_runtime": 4.6404, + "eval_vitaminc-pairs_samples_per_second": 27.584, + "eval_vitaminc-pairs_steps_per_second": 0.215, + "step": 546 + }, + { + "epoch": 1.0526992287917738, + "eval_negation-triplets_loss": 0.48229217529296875, + "eval_negation-triplets_runtime": 3.5048, + "eval_negation-triplets_samples_per_second": 36.521, + "eval_negation-triplets_steps_per_second": 0.285, + "step": 546 + }, + { + "epoch": 1.0526992287917738, + "eval_scitail-pairs-pos_loss": 0.009107854217290878, + "eval_scitail-pairs-pos_runtime": 2.8141, + "eval_scitail-pairs-pos_samples_per_second": 45.486, + "eval_scitail-pairs-pos_steps_per_second": 0.355, + "step": 546 + }, + { + "epoch": 1.0526992287917738, + "eval_scitail-pairs-qa_loss": 0.013651267625391483, + "eval_scitail-pairs-qa_runtime": 2.3335, + "eval_scitail-pairs-qa_samples_per_second": 54.854, + "eval_scitail-pairs-qa_steps_per_second": 0.429, + "step": 546 + }, + { + "epoch": 1.0526992287917738, + "eval_xsum-pairs_loss": 0.005041995085775852, + "eval_xsum-pairs_runtime": 3.1099, + "eval_xsum-pairs_samples_per_second": 41.158, + "eval_xsum-pairs_steps_per_second": 0.322, + "step": 546 + }, + { + "epoch": 1.0526992287917738, + "eval_sciq_pairs_loss": 0.06929144263267517, + "eval_sciq_pairs_runtime": 4.8403, + "eval_sciq_pairs_samples_per_second": 26.445, + "eval_sciq_pairs_steps_per_second": 0.207, + "step": 546 + }, + { + "epoch": 1.0526992287917738, + "eval_qasc_pairs_loss": 0.05941355228424072, + "eval_qasc_pairs_runtime": 2.154, + "eval_qasc_pairs_samples_per_second": 59.426, + "eval_qasc_pairs_steps_per_second": 0.464, + "step": 546 + }, + { + "epoch": 1.0526992287917738, + "eval_openbookqa_pairs_loss": 0.4173741340637207, + "eval_openbookqa_pairs_runtime": 2.2586, + "eval_openbookqa_pairs_samples_per_second": 56.671, + "eval_openbookqa_pairs_steps_per_second": 0.443, + "step": 546 + }, + { + "epoch": 1.0526992287917738, + "eval_msmarco_pairs_loss": 0.015324967913329601, + "eval_msmarco_pairs_runtime": 2.2897, + "eval_msmarco_pairs_samples_per_second": 55.902, + "eval_msmarco_pairs_steps_per_second": 0.437, + "step": 546 + }, + { + "epoch": 1.0526992287917738, + "eval_nq_pairs_loss": 0.03324040398001671, + "eval_nq_pairs_runtime": 3.6951, + "eval_nq_pairs_samples_per_second": 34.64, + "eval_nq_pairs_steps_per_second": 0.271, + "step": 546 + }, + { + "epoch": 1.0526992287917738, + "eval_trivia_pairs_loss": 0.0088445283472538, + "eval_trivia_pairs_runtime": 3.1799, + "eval_trivia_pairs_samples_per_second": 40.253, + "eval_trivia_pairs_steps_per_second": 0.314, + "step": 546 + }, + { + "epoch": 1.0526992287917738, + "eval_gooaq_pairs_loss": 0.07060601562261581, + "eval_gooaq_pairs_runtime": 2.1993, + "eval_gooaq_pairs_samples_per_second": 58.199, + "eval_gooaq_pairs_steps_per_second": 0.455, + "step": 546 + }, + { + "epoch": 1.0526992287917738, + "eval_paws-pos_loss": 0.02692914567887783, + "eval_paws-pos_runtime": 2.3486, + "eval_paws-pos_samples_per_second": 54.501, + "eval_paws-pos_steps_per_second": 0.426, + "step": 546 + }, + { + "epoch": 1.0526992287917738, + "eval_global_dataset_loss": 0.22938521206378937, + "eval_global_dataset_runtime": 8.4105, + "eval_global_dataset_samples_per_second": 38.642, + "eval_global_dataset_steps_per_second": 0.357, + "step": 546 + }, + { + "epoch": 1.0584832904884318, + "grad_norm": 0.8804048299789429, + "learning_rate": 1.1117093950167325e-05, + "loss": 0.054, + "step": 549 + }, + { + "epoch": 1.06426735218509, + "grad_norm": 2.0877773761749268, + "learning_rate": 1.1042762281305083e-05, + "loss": 0.046, + "step": 552 + }, + { + "epoch": 1.070051413881748, + "grad_norm": 2.892855167388916, + "learning_rate": 1.0968019719329226e-05, + "loss": 0.0982, + "step": 555 + }, + { + "epoch": 1.075835475578406, + "grad_norm": 2.433546304702759, + "learning_rate": 1.0892877260952303e-05, + "loss": 0.0441, + "step": 558 + }, + { + "epoch": 1.0816195372750643, + "grad_norm": 10.981830596923828, + "learning_rate": 1.081734596172275e-05, + "loss": 0.1141, + "step": 561 + }, + { + "epoch": 1.0874035989717223, + "grad_norm": 3.4110443592071533, + "learning_rate": 1.0741436934398335e-05, + "loss": 0.1131, + "step": 564 + }, + { + "epoch": 1.0931876606683804, + "grad_norm": 6.111934661865234, + "learning_rate": 1.0665161347311134e-05, + "loss": 0.0423, + "step": 567 + }, + { + "epoch": 1.0989717223650386, + "grad_norm": 4.0002312660217285, + "learning_rate": 1.0588530422724387e-05, + "loss": 0.1062, + "step": 570 + }, + { + "epoch": 1.1028277634961439, + "eval_Qnli-dev_cosine_accuracy": 0.73046875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6712266206741333, + "eval_Qnli-dev_cosine_ap": 0.763778545105543, + "eval_Qnli-dev_cosine_f1": 0.7204502814258913, + "eval_Qnli-dev_cosine_f1_threshold": 0.6122102737426758, + "eval_Qnli-dev_cosine_precision": 0.6464646464646465, + "eval_Qnli-dev_cosine_recall": 0.8135593220338984, + "eval_Qnli-dev_dot_accuracy": 0.708984375, + "eval_Qnli-dev_dot_accuracy_threshold": 623.3515625, + "eval_Qnli-dev_dot_ap": 0.7357525473143589, + "eval_Qnli-dev_dot_f1": 0.6990654205607476, + "eval_Qnli-dev_dot_f1_threshold": 565.19580078125, + "eval_Qnli-dev_dot_precision": 0.6254180602006689, + "eval_Qnli-dev_dot_recall": 0.7923728813559322, + "eval_Qnli-dev_euclidean_accuracy": 0.732421875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 24.461320877075195, + "eval_Qnli-dev_euclidean_ap": 0.7695553886748087, + "eval_Qnli-dev_euclidean_f1": 0.7195357833655707, + "eval_Qnli-dev_euclidean_f1_threshold": 26.384815216064453, + "eval_Qnli-dev_euclidean_precision": 0.6619217081850534, + "eval_Qnli-dev_euclidean_recall": 0.788135593220339, + "eval_Qnli-dev_manhattan_accuracy": 0.73046875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 759.6908569335938, + "eval_Qnli-dev_manhattan_ap": 0.769070403388376, + "eval_Qnli-dev_manhattan_f1": 0.7200000000000001, + "eval_Qnli-dev_manhattan_f1_threshold": 828.003173828125, + "eval_Qnli-dev_manhattan_precision": 0.6539792387543253, + "eval_Qnli-dev_manhattan_recall": 0.8008474576271186, + "eval_Qnli-dev_max_accuracy": 0.732421875, + "eval_Qnli-dev_max_accuracy_threshold": 759.6908569335938, + "eval_Qnli-dev_max_ap": 0.7695553886748087, + "eval_Qnli-dev_max_f1": 0.7204502814258913, + "eval_Qnli-dev_max_f1_threshold": 828.003173828125, + "eval_Qnli-dev_max_precision": 0.6619217081850534, + "eval_Qnli-dev_max_recall": 0.8135593220338984, + "eval_allNLI-dev_cosine_accuracy": 0.73046875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8079162836074829, + "eval_allNLI-dev_cosine_ap": 0.6161924609767827, + "eval_allNLI-dev_cosine_f1": 0.6376146788990825, + "eval_allNLI-dev_cosine_f1_threshold": 0.6470102071762085, + "eval_allNLI-dev_cosine_precision": 0.5285171102661597, + "eval_allNLI-dev_cosine_recall": 0.8034682080924855, + "eval_allNLI-dev_dot_accuracy": 0.73046875, + "eval_allNLI-dev_dot_accuracy_threshold": 802.7283935546875, + "eval_allNLI-dev_dot_ap": 0.6270154596584896, + "eval_allNLI-dev_dot_f1": 0.6347826086956523, + "eval_allNLI-dev_dot_f1_threshold": 603.6707763671875, + "eval_allNLI-dev_dot_precision": 0.5087108013937283, + "eval_allNLI-dev_dot_recall": 0.8439306358381503, + "eval_allNLI-dev_euclidean_accuracy": 0.72265625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 18.285118103027344, + "eval_allNLI-dev_euclidean_ap": 0.6109553174348793, + "eval_allNLI-dev_euclidean_f1": 0.6387665198237885, + "eval_allNLI-dev_euclidean_f1_threshold": 26.536678314208984, + "eval_allNLI-dev_euclidean_precision": 0.5160142348754448, + "eval_allNLI-dev_euclidean_recall": 0.838150289017341, + "eval_allNLI-dev_manhattan_accuracy": 0.724609375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 594.1370849609375, + "eval_allNLI-dev_manhattan_ap": 0.6127360214886293, + "eval_allNLI-dev_manhattan_f1": 0.6361556064073227, + "eval_allNLI-dev_manhattan_f1_threshold": 810.3226318359375, + "eval_allNLI-dev_manhattan_precision": 0.5265151515151515, + "eval_allNLI-dev_manhattan_recall": 0.8034682080924855, + "eval_allNLI-dev_max_accuracy": 0.73046875, + "eval_allNLI-dev_max_accuracy_threshold": 802.7283935546875, + "eval_allNLI-dev_max_ap": 0.6270154596584896, + "eval_allNLI-dev_max_f1": 0.6387665198237885, + "eval_allNLI-dev_max_f1_threshold": 810.3226318359375, + "eval_allNLI-dev_max_precision": 0.5285171102661597, + "eval_allNLI-dev_max_recall": 0.8439306358381503, + "eval_sequential_score": 0.7695553886748087, + "eval_sts-test_pearson_cosine": 0.9172526996921664, + "eval_sts-test_pearson_dot": 0.9025961634730928, + "eval_sts-test_pearson_euclidean": 0.9302608980056621, + "eval_sts-test_pearson_manhattan": 0.9294578977730011, + "eval_sts-test_pearson_max": 0.9302608980056621, + "eval_sts-test_spearman_cosine": 0.9246459534891734, + "eval_sts-test_spearman_dot": 0.9017780589181682, + "eval_sts-test_spearman_euclidean": 0.9252701888343221, + "eval_sts-test_spearman_manhattan": 0.924371628256844, + "eval_sts-test_spearman_max": 0.9252701888343221, + "eval_vitaminc-pairs_loss": 2.5838048458099365, + "eval_vitaminc-pairs_runtime": 4.4571, + "eval_vitaminc-pairs_samples_per_second": 28.718, + "eval_vitaminc-pairs_steps_per_second": 0.224, + "step": 572 + }, + { + "epoch": 1.1028277634961439, + "eval_negation-triplets_loss": 0.4972306787967682, + "eval_negation-triplets_runtime": 3.2687, + "eval_negation-triplets_samples_per_second": 39.159, + "eval_negation-triplets_steps_per_second": 0.306, + "step": 572 + }, + { + "epoch": 1.1028277634961439, + "eval_scitail-pairs-pos_loss": 0.015652282163500786, + "eval_scitail-pairs-pos_runtime": 2.6482, + "eval_scitail-pairs-pos_samples_per_second": 48.334, + "eval_scitail-pairs-pos_steps_per_second": 0.378, + "step": 572 + }, + { + "epoch": 1.1028277634961439, + "eval_scitail-pairs-qa_loss": 0.019644802436232567, + "eval_scitail-pairs-qa_runtime": 2.1941, + "eval_scitail-pairs-qa_samples_per_second": 58.339, + "eval_scitail-pairs-qa_steps_per_second": 0.456, + "step": 572 + }, + { + "epoch": 1.1028277634961439, + "eval_xsum-pairs_loss": 0.0077519528567790985, + "eval_xsum-pairs_runtime": 3.0974, + "eval_xsum-pairs_samples_per_second": 41.325, + "eval_xsum-pairs_steps_per_second": 0.323, + "step": 572 + }, + { + "epoch": 1.1028277634961439, + "eval_sciq_pairs_loss": 0.07203073799610138, + "eval_sciq_pairs_runtime": 4.7359, + "eval_sciq_pairs_samples_per_second": 27.028, + "eval_sciq_pairs_steps_per_second": 0.211, + "step": 572 + }, + { + "epoch": 1.1028277634961439, + "eval_qasc_pairs_loss": 0.027775608003139496, + "eval_qasc_pairs_runtime": 2.0931, + "eval_qasc_pairs_samples_per_second": 61.153, + "eval_qasc_pairs_steps_per_second": 0.478, + "step": 572 + }, + { + "epoch": 1.1028277634961439, + "eval_openbookqa_pairs_loss": 0.3412659466266632, + "eval_openbookqa_pairs_runtime": 2.1842, + "eval_openbookqa_pairs_samples_per_second": 58.602, + "eval_openbookqa_pairs_steps_per_second": 0.458, + "step": 572 + }, + { + "epoch": 1.1028277634961439, + "eval_msmarco_pairs_loss": 0.01285499520599842, + "eval_msmarco_pairs_runtime": 2.2378, + "eval_msmarco_pairs_samples_per_second": 57.198, + "eval_msmarco_pairs_steps_per_second": 0.447, + "step": 572 + }, + { + "epoch": 1.1028277634961439, + "eval_nq_pairs_loss": 0.0469130277633667, + "eval_nq_pairs_runtime": 3.6196, + "eval_nq_pairs_samples_per_second": 35.363, + "eval_nq_pairs_steps_per_second": 0.276, + "step": 572 + }, + { + "epoch": 1.1028277634961439, + "eval_trivia_pairs_loss": 0.013011205941438675, + "eval_trivia_pairs_runtime": 3.056, + "eval_trivia_pairs_samples_per_second": 41.885, + "eval_trivia_pairs_steps_per_second": 0.327, + "step": 572 + }, + { + "epoch": 1.1028277634961439, + "eval_gooaq_pairs_loss": 0.06344185769557953, + "eval_gooaq_pairs_runtime": 2.1206, + "eval_gooaq_pairs_samples_per_second": 60.36, + "eval_gooaq_pairs_steps_per_second": 0.472, + "step": 572 + }, + { + "epoch": 1.1028277634961439, + "eval_paws-pos_loss": 0.028861284255981445, + "eval_paws-pos_runtime": 2.254, + "eval_paws-pos_samples_per_second": 56.788, + "eval_paws-pos_steps_per_second": 0.444, + "step": 572 + }, + { + "epoch": 1.1028277634961439, + "eval_global_dataset_loss": 0.1511072963476181, + "eval_global_dataset_runtime": 8.3415, + "eval_global_dataset_samples_per_second": 38.962, + "eval_global_dataset_steps_per_second": 0.36, + "step": 572 + }, + { + "epoch": 1.1047557840616966, + "grad_norm": 2.5909698009490967, + "learning_rate": 1.0511555435181364e-05, + "loss": 0.0428, + "step": 573 + }, + { + "epoch": 1.1105398457583548, + "grad_norm": 1.4424036741256714, + "learning_rate": 1.043424770984659e-05, + "loss": 0.2265, + "step": 576 + }, + { + "epoch": 1.1163239074550129, + "grad_norm": 2.903937339782715, + "learning_rate": 1.0356618620839581e-05, + "loss": 0.0881, + "step": 579 + }, + { + "epoch": 1.1221079691516709, + "grad_norm": 3.5707309246063232, + "learning_rate": 1.0278679589561402e-05, + "loss": 0.0932, + "step": 582 + }, + { + "epoch": 1.1278920308483291, + "grad_norm": 2.6292388439178467, + "learning_rate": 1.0200442083014253e-05, + "loss": 0.0474, + "step": 585 + }, + { + "epoch": 1.1336760925449871, + "grad_norm": 3.0157814025878906, + "learning_rate": 1.0121917612114365e-05, + "loss": 0.0512, + "step": 588 + }, + { + "epoch": 1.1394601542416452, + "grad_norm": 2.6974472999572754, + "learning_rate": 1.0043117729998415e-05, + "loss": 0.0316, + "step": 591 + }, + { + "epoch": 1.1452442159383034, + "grad_norm": 2.424243688583374, + "learning_rate": 9.96405403032374e-06, + "loss": 0.1243, + "step": 594 + }, + { + "epoch": 1.1510282776349614, + "grad_norm": 3.2901053428649902, + "learning_rate": 9.88473814556259e-06, + "loss": 0.0416, + "step": 597 + }, + { + "epoch": 1.1529562982005142, + "eval_Qnli-dev_cosine_accuracy": 0.697265625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.73479163646698, + "eval_Qnli-dev_cosine_ap": 0.7392549146749757, + "eval_Qnli-dev_cosine_f1": 0.7017543859649122, + "eval_Qnli-dev_cosine_f1_threshold": 0.5977563858032227, + "eval_Qnli-dev_cosine_precision": 0.5988023952095808, + "eval_Qnli-dev_cosine_recall": 0.847457627118644, + "eval_Qnli-dev_dot_accuracy": 0.6796875, + "eval_Qnli-dev_dot_accuracy_threshold": 656.5439453125, + "eval_Qnli-dev_dot_ap": 0.7093180934545957, + "eval_Qnli-dev_dot_f1": 0.6918918918918919, + "eval_Qnli-dev_dot_f1_threshold": 559.746337890625, + "eval_Qnli-dev_dot_precision": 0.6018808777429467, + "eval_Qnli-dev_dot_recall": 0.8135593220338984, + "eval_Qnli-dev_euclidean_accuracy": 0.71484375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 23.493846893310547, + "eval_Qnli-dev_euclidean_ap": 0.7455155904183673, + "eval_Qnli-dev_euclidean_f1": 0.7012089810017271, + "eval_Qnli-dev_euclidean_f1_threshold": 27.43291473388672, + "eval_Qnli-dev_euclidean_precision": 0.5918367346938775, + "eval_Qnli-dev_euclidean_recall": 0.8601694915254238, + "eval_Qnli-dev_manhattan_accuracy": 0.708984375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 710.3681640625, + "eval_Qnli-dev_manhattan_ap": 0.7437923697960528, + "eval_Qnli-dev_manhattan_f1": 0.7045009784735812, + "eval_Qnli-dev_manhattan_f1_threshold": 796.3446044921875, + "eval_Qnli-dev_manhattan_precision": 0.6545454545454545, + "eval_Qnli-dev_manhattan_recall": 0.7627118644067796, + "eval_Qnli-dev_max_accuracy": 0.71484375, + "eval_Qnli-dev_max_accuracy_threshold": 710.3681640625, + "eval_Qnli-dev_max_ap": 0.7455155904183673, + "eval_Qnli-dev_max_f1": 0.7045009784735812, + "eval_Qnli-dev_max_f1_threshold": 796.3446044921875, + "eval_Qnli-dev_max_precision": 0.6545454545454545, + "eval_Qnli-dev_max_recall": 0.8601694915254238, + "eval_allNLI-dev_cosine_accuracy": 0.724609375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.8415929675102234, + "eval_allNLI-dev_cosine_ap": 0.6082218915414402, + "eval_allNLI-dev_cosine_f1": 0.6330935251798562, + "eval_allNLI-dev_cosine_f1_threshold": 0.680069088935852, + "eval_allNLI-dev_cosine_precision": 0.5409836065573771, + "eval_allNLI-dev_cosine_recall": 0.7630057803468208, + "eval_allNLI-dev_dot_accuracy": 0.724609375, + "eval_allNLI-dev_dot_accuracy_threshold": 834.5533447265625, + "eval_allNLI-dev_dot_ap": 0.6175956248571282, + "eval_allNLI-dev_dot_f1": 0.6261682242990655, + "eval_allNLI-dev_dot_f1_threshold": 644.3388671875, + "eval_allNLI-dev_dot_precision": 0.5254901960784314, + "eval_allNLI-dev_dot_recall": 0.7745664739884393, + "eval_allNLI-dev_euclidean_accuracy": 0.72265625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 17.67273712158203, + "eval_allNLI-dev_euclidean_ap": 0.6031971080153227, + "eval_allNLI-dev_euclidean_f1": 0.6229508196721311, + "eval_allNLI-dev_euclidean_f1_threshold": 25.2102108001709, + "eval_allNLI-dev_euclidean_precision": 0.5236220472440944, + "eval_allNLI-dev_euclidean_recall": 0.7687861271676301, + "eval_allNLI-dev_manhattan_accuracy": 0.72265625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 547.6099243164062, + "eval_allNLI-dev_manhattan_ap": 0.6040106690581193, + "eval_allNLI-dev_manhattan_f1": 0.6278026905829597, + "eval_allNLI-dev_manhattan_f1_threshold": 798.2863159179688, + "eval_allNLI-dev_manhattan_precision": 0.5128205128205128, + "eval_allNLI-dev_manhattan_recall": 0.8092485549132948, + "eval_allNLI-dev_max_accuracy": 0.724609375, + "eval_allNLI-dev_max_accuracy_threshold": 834.5533447265625, + "eval_allNLI-dev_max_ap": 0.6175956248571282, + "eval_allNLI-dev_max_f1": 0.6330935251798562, + "eval_allNLI-dev_max_f1_threshold": 798.2863159179688, + "eval_allNLI-dev_max_precision": 0.5409836065573771, + "eval_allNLI-dev_max_recall": 0.8092485549132948, + "eval_sequential_score": 0.7455155904183673, + "eval_sts-test_pearson_cosine": 0.9167693090648592, + "eval_sts-test_pearson_dot": 0.9069501193890914, + "eval_sts-test_pearson_euclidean": 0.9337570508793084, + "eval_sts-test_pearson_manhattan": 0.9325869075370774, + "eval_sts-test_pearson_max": 0.9337570508793084, + "eval_sts-test_spearman_cosine": 0.9281000795996684, + "eval_sts-test_spearman_dot": 0.9121680487779323, + "eval_sts-test_spearman_euclidean": 0.9278352416347543, + "eval_sts-test_spearman_manhattan": 0.926563903049886, + "eval_sts-test_spearman_max": 0.9281000795996684, + "eval_vitaminc-pairs_loss": 3.1178977489471436, + "eval_vitaminc-pairs_runtime": 4.5495, + "eval_vitaminc-pairs_samples_per_second": 28.135, + "eval_vitaminc-pairs_steps_per_second": 0.22, + "step": 598 + }, + { + "epoch": 1.1529562982005142, + "eval_negation-triplets_loss": 0.540117084980011, + "eval_negation-triplets_runtime": 3.3419, + "eval_negation-triplets_samples_per_second": 38.301, + "eval_negation-triplets_steps_per_second": 0.299, + "step": 598 + }, + { + "epoch": 1.1529562982005142, + "eval_scitail-pairs-pos_loss": 0.011858026497066021, + "eval_scitail-pairs-pos_runtime": 2.7673, + "eval_scitail-pairs-pos_samples_per_second": 46.255, + "eval_scitail-pairs-pos_steps_per_second": 0.361, + "step": 598 + }, + { + "epoch": 1.1529562982005142, + "eval_scitail-pairs-qa_loss": 0.034890156239271164, + "eval_scitail-pairs-qa_runtime": 2.3236, + "eval_scitail-pairs-qa_samples_per_second": 55.086, + "eval_scitail-pairs-qa_steps_per_second": 0.43, + "step": 598 + }, + { + "epoch": 1.1529562982005142, + "eval_xsum-pairs_loss": 0.005378103815019131, + "eval_xsum-pairs_runtime": 3.1348, + "eval_xsum-pairs_samples_per_second": 40.832, + "eval_xsum-pairs_steps_per_second": 0.319, + "step": 598 + }, + { + "epoch": 1.1529562982005142, + "eval_sciq_pairs_loss": 0.07895553112030029, + "eval_sciq_pairs_runtime": 4.8335, + "eval_sciq_pairs_samples_per_second": 26.482, + "eval_sciq_pairs_steps_per_second": 0.207, + "step": 598 + }, + { + "epoch": 1.1529562982005142, + "eval_qasc_pairs_loss": 0.02402154728770256, + "eval_qasc_pairs_runtime": 2.1497, + "eval_qasc_pairs_samples_per_second": 59.544, + "eval_qasc_pairs_steps_per_second": 0.465, + "step": 598 + }, + { + "epoch": 1.1529562982005142, + "eval_openbookqa_pairs_loss": 0.3324708640575409, + "eval_openbookqa_pairs_runtime": 2.4467, + "eval_openbookqa_pairs_samples_per_second": 52.315, + "eval_openbookqa_pairs_steps_per_second": 0.409, + "step": 598 + }, + { + "epoch": 1.1529562982005142, + "eval_msmarco_pairs_loss": 0.007721642032265663, + "eval_msmarco_pairs_runtime": 2.2839, + "eval_msmarco_pairs_samples_per_second": 56.046, + "eval_msmarco_pairs_steps_per_second": 0.438, + "step": 598 + }, + { + "epoch": 1.1529562982005142, + "eval_nq_pairs_loss": 0.04540184885263443, + "eval_nq_pairs_runtime": 3.6668, + "eval_nq_pairs_samples_per_second": 34.908, + "eval_nq_pairs_steps_per_second": 0.273, + "step": 598 + }, + { + "epoch": 1.1529562982005142, + "eval_trivia_pairs_loss": 0.01642066240310669, + "eval_trivia_pairs_runtime": 3.132, + "eval_trivia_pairs_samples_per_second": 40.868, + "eval_trivia_pairs_steps_per_second": 0.319, + "step": 598 + }, + { + "epoch": 1.1529562982005142, + "eval_gooaq_pairs_loss": 0.06850580871105194, + "eval_gooaq_pairs_runtime": 2.1873, + "eval_gooaq_pairs_samples_per_second": 58.519, + "eval_gooaq_pairs_steps_per_second": 0.457, + "step": 598 + }, + { + "epoch": 1.1529562982005142, + "eval_paws-pos_loss": 0.02607266791164875, + "eval_paws-pos_runtime": 2.3007, + "eval_paws-pos_samples_per_second": 55.634, + "eval_paws-pos_steps_per_second": 0.435, + "step": 598 + }, + { + "epoch": 1.1529562982005142, + "eval_global_dataset_loss": 0.17176896333694458, + "eval_global_dataset_runtime": 8.3995, + "eval_global_dataset_samples_per_second": 38.693, + "eval_global_dataset_steps_per_second": 0.357, + "step": 598 + }, + { + "epoch": 1.1568123393316196, + "grad_norm": 3.0405514240264893, + "learning_rate": 9.805181745290658e-06, + "loss": 0.0591, + "step": 600 + }, + { + "epoch": 1.1625964010282777, + "grad_norm": 5.0244550704956055, + "learning_rate": 9.725396534470176e-06, + "loss": 0.1668, + "step": 603 + }, + { + "epoch": 1.1683804627249357, + "grad_norm": 3.8621885776519775, + "learning_rate": 9.645394251727778e-06, + "loss": 0.1302, + "step": 606 + }, + { + "epoch": 1.174164524421594, + "grad_norm": 13.87588882446289, + "learning_rate": 9.56518666762742e-06, + "loss": 0.0912, + "step": 609 + }, + { + "epoch": 1.179948586118252, + "grad_norm": 1.3986790180206299, + "learning_rate": 9.48478558293861e-06, + "loss": 0.0357, + "step": 612 + }, + { + "epoch": 1.18573264781491, + "grad_norm": 1.9629465341567993, + "learning_rate": 9.40420282690018e-06, + "loss": 0.0239, + "step": 615 + }, + { + "epoch": 1.1915167095115682, + "grad_norm": 0.35275983810424805, + "learning_rate": 9.323450255479868e-06, + "loss": 0.0682, + "step": 618 + }, + { + "epoch": 1.1973007712082262, + "grad_norm": 3.304192066192627, + "learning_rate": 9.242539749630008e-06, + "loss": 0.0634, + "step": 621 + }, + { + "epoch": 1.2030848329048842, + "grad_norm": 3.8888630867004395, + "learning_rate": 9.161483213539461e-06, + "loss": 0.1228, + "step": 624 + }, + { + "epoch": 1.2030848329048842, + "eval_Qnli-dev_cosine_accuracy": 0.71484375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.7114617824554443, + "eval_Qnli-dev_cosine_ap": 0.74748968364284, + "eval_Qnli-dev_cosine_f1": 0.699410609037328, + "eval_Qnli-dev_cosine_f1_threshold": 0.644582986831665, + "eval_Qnli-dev_cosine_precision": 0.652014652014652, + "eval_Qnli-dev_cosine_recall": 0.7542372881355932, + "eval_Qnli-dev_dot_accuracy": 0.693359375, + "eval_Qnli-dev_dot_accuracy_threshold": 630.5953369140625, + "eval_Qnli-dev_dot_ap": 0.7222356207944656, + "eval_Qnli-dev_dot_f1": 0.6863711001642037, + "eval_Qnli-dev_dot_f1_threshold": 513.5455932617188, + "eval_Qnli-dev_dot_precision": 0.5603217158176944, + "eval_Qnli-dev_dot_recall": 0.885593220338983, + "eval_Qnli-dev_euclidean_accuracy": 0.720703125, + "eval_Qnli-dev_euclidean_accuracy_threshold": 23.42608070373535, + "eval_Qnli-dev_euclidean_ap": 0.7506295294785469, + "eval_Qnli-dev_euclidean_f1": 0.700507614213198, + "eval_Qnli-dev_euclidean_f1_threshold": 27.891647338867188, + "eval_Qnli-dev_euclidean_precision": 0.5830985915492958, + "eval_Qnli-dev_euclidean_recall": 0.8771186440677966, + "eval_Qnli-dev_manhattan_accuracy": 0.71875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 717.260986328125, + "eval_Qnli-dev_manhattan_ap": 0.7508208159827998, + "eval_Qnli-dev_manhattan_f1": 0.7033333333333334, + "eval_Qnli-dev_manhattan_f1_threshold": 874.6651000976562, + "eval_Qnli-dev_manhattan_precision": 0.5796703296703297, + "eval_Qnli-dev_manhattan_recall": 0.8940677966101694, + "eval_Qnli-dev_max_accuracy": 0.720703125, + "eval_Qnli-dev_max_accuracy_threshold": 717.260986328125, + "eval_Qnli-dev_max_ap": 0.7508208159827998, + "eval_Qnli-dev_max_f1": 0.7033333333333334, + "eval_Qnli-dev_max_f1_threshold": 874.6651000976562, + "eval_Qnli-dev_max_precision": 0.652014652014652, + "eval_Qnli-dev_max_recall": 0.8940677966101694, + "eval_allNLI-dev_cosine_accuracy": 0.72265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7328661680221558, + "eval_allNLI-dev_cosine_ap": 0.6163368223972276, + "eval_allNLI-dev_cosine_f1": 0.6232558139534884, + "eval_allNLI-dev_cosine_f1_threshold": 0.6505966782569885, + "eval_allNLI-dev_cosine_precision": 0.5214007782101168, + "eval_allNLI-dev_cosine_recall": 0.7745664739884393, + "eval_allNLI-dev_dot_accuracy": 0.72265625, + "eval_allNLI-dev_dot_accuracy_threshold": 828.3904418945312, + "eval_allNLI-dev_dot_ap": 0.6278874973841797, + "eval_allNLI-dev_dot_f1": 0.624390243902439, + "eval_allNLI-dev_dot_f1_threshold": 653.0433959960938, + "eval_allNLI-dev_dot_precision": 0.540084388185654, + "eval_allNLI-dev_dot_recall": 0.7398843930635838, + "eval_allNLI-dev_euclidean_accuracy": 0.728515625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 22.045154571533203, + "eval_allNLI-dev_euclidean_ap": 0.6119417414575073, + "eval_allNLI-dev_euclidean_f1": 0.6208530805687204, + "eval_allNLI-dev_euclidean_f1_threshold": 25.78974151611328, + "eval_allNLI-dev_euclidean_precision": 0.5261044176706827, + "eval_allNLI-dev_euclidean_recall": 0.7572254335260116, + "eval_allNLI-dev_manhattan_accuracy": 0.728515625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 688.5887451171875, + "eval_allNLI-dev_manhattan_ap": 0.6134324305699217, + "eval_allNLI-dev_manhattan_f1": 0.620525059665871, + "eval_allNLI-dev_manhattan_f1_threshold": 801.7684326171875, + "eval_allNLI-dev_manhattan_precision": 0.5284552845528455, + "eval_allNLI-dev_manhattan_recall": 0.7514450867052023, + "eval_allNLI-dev_max_accuracy": 0.728515625, + "eval_allNLI-dev_max_accuracy_threshold": 828.3904418945312, + "eval_allNLI-dev_max_ap": 0.6278874973841797, + "eval_allNLI-dev_max_f1": 0.624390243902439, + "eval_allNLI-dev_max_f1_threshold": 801.7684326171875, + "eval_allNLI-dev_max_precision": 0.540084388185654, + "eval_allNLI-dev_max_recall": 0.7745664739884393, + "eval_sequential_score": 0.7508208159827998, + "eval_sts-test_pearson_cosine": 0.9187156498875052, + "eval_sts-test_pearson_dot": 0.9077884045200496, + "eval_sts-test_pearson_euclidean": 0.9338738176759427, + "eval_sts-test_pearson_manhattan": 0.9328491962335101, + "eval_sts-test_pearson_max": 0.9338738176759427, + "eval_sts-test_spearman_cosine": 0.9260453254328825, + "eval_sts-test_spearman_dot": 0.9055982503468449, + "eval_sts-test_spearman_euclidean": 0.9263229175072933, + "eval_sts-test_spearman_manhattan": 0.9251408129376393, + "eval_sts-test_spearman_max": 0.9263229175072933, + "eval_vitaminc-pairs_loss": 3.139256715774536, + "eval_vitaminc-pairs_runtime": 4.4256, + "eval_vitaminc-pairs_samples_per_second": 28.923, + "eval_vitaminc-pairs_steps_per_second": 0.226, + "step": 624 + }, + { + "epoch": 1.2030848329048842, + "eval_negation-triplets_loss": 0.529778778553009, + "eval_negation-triplets_runtime": 3.2568, + "eval_negation-triplets_samples_per_second": 39.302, + "eval_negation-triplets_steps_per_second": 0.307, + "step": 624 + }, + { + "epoch": 1.2030848329048842, + "eval_scitail-pairs-pos_loss": 0.01055553462356329, + "eval_scitail-pairs-pos_runtime": 2.6293, + "eval_scitail-pairs-pos_samples_per_second": 48.683, + "eval_scitail-pairs-pos_steps_per_second": 0.38, + "step": 624 + }, + { + "epoch": 1.2030848329048842, + "eval_scitail-pairs-qa_loss": 0.002362636849284172, + "eval_scitail-pairs-qa_runtime": 2.2564, + "eval_scitail-pairs-qa_samples_per_second": 56.728, + "eval_scitail-pairs-qa_steps_per_second": 0.443, + "step": 624 + }, + { + "epoch": 1.2030848329048842, + "eval_xsum-pairs_loss": 0.004232893232256174, + "eval_xsum-pairs_runtime": 3.058, + "eval_xsum-pairs_samples_per_second": 41.858, + "eval_xsum-pairs_steps_per_second": 0.327, + "step": 624 + }, + { + "epoch": 1.2030848329048842, + "eval_sciq_pairs_loss": 0.07681445777416229, + "eval_sciq_pairs_runtime": 4.742, + "eval_sciq_pairs_samples_per_second": 26.993, + "eval_sciq_pairs_steps_per_second": 0.211, + "step": 624 + }, + { + "epoch": 1.2030848329048842, + "eval_qasc_pairs_loss": 0.04607044905424118, + "eval_qasc_pairs_runtime": 2.0742, + "eval_qasc_pairs_samples_per_second": 61.71, + "eval_qasc_pairs_steps_per_second": 0.482, + "step": 624 + }, + { + "epoch": 1.2030848329048842, + "eval_openbookqa_pairs_loss": 0.35425248742103577, + "eval_openbookqa_pairs_runtime": 2.2649, + "eval_openbookqa_pairs_samples_per_second": 56.515, + "eval_openbookqa_pairs_steps_per_second": 0.442, + "step": 624 + }, + { + "epoch": 1.2030848329048842, + "eval_msmarco_pairs_loss": 0.009293444454669952, + "eval_msmarco_pairs_runtime": 2.2209, + "eval_msmarco_pairs_samples_per_second": 57.634, + "eval_msmarco_pairs_steps_per_second": 0.45, + "step": 624 + }, + { + "epoch": 1.2030848329048842, + "eval_nq_pairs_loss": 0.021805401891469955, + "eval_nq_pairs_runtime": 3.6205, + "eval_nq_pairs_samples_per_second": 35.354, + "eval_nq_pairs_steps_per_second": 0.276, + "step": 624 + }, + { + "epoch": 1.2030848329048842, + "eval_trivia_pairs_loss": 0.009247110225260258, + "eval_trivia_pairs_runtime": 3.0648, + "eval_trivia_pairs_samples_per_second": 41.764, + "eval_trivia_pairs_steps_per_second": 0.326, + "step": 624 + }, + { + "epoch": 1.2030848329048842, + "eval_gooaq_pairs_loss": 0.05935032293200493, + "eval_gooaq_pairs_runtime": 2.1541, + "eval_gooaq_pairs_samples_per_second": 59.421, + "eval_gooaq_pairs_steps_per_second": 0.464, + "step": 624 + }, + { + "epoch": 1.2030848329048842, + "eval_paws-pos_loss": 0.027399806305766106, + "eval_paws-pos_runtime": 2.1806, + "eval_paws-pos_samples_per_second": 58.699, + "eval_paws-pos_steps_per_second": 0.459, + "step": 624 + }, + { + "epoch": 1.2030848329048842, + "eval_global_dataset_loss": 0.17402325570583344, + "eval_global_dataset_runtime": 8.3332, + "eval_global_dataset_samples_per_second": 39.001, + "eval_global_dataset_steps_per_second": 0.36, + "step": 624 + }, + { + "epoch": 1.2088688946015425, + "grad_norm": 4.095309734344482, + "learning_rate": 9.080292572882221e-06, + "loss": 0.1741, + "step": 627 + }, + { + "epoch": 1.2146529562982005, + "grad_norm": 2.805889129638672, + "learning_rate": 8.998979773062789e-06, + "loss": 0.0413, + "step": 630 + }, + { + "epoch": 1.2204370179948587, + "grad_norm": 2.498549699783325, + "learning_rate": 8.917556777458686e-06, + "loss": 0.1496, + "step": 633 + }, + { + "epoch": 1.2262210796915167, + "grad_norm": 1.6500836610794067, + "learning_rate": 8.836035565660305e-06, + "loss": 0.0229, + "step": 636 + }, + { + "epoch": 1.2320051413881747, + "grad_norm": 2.163297653198242, + "learning_rate": 8.754428131708364e-06, + "loss": 0.0871, + "step": 639 + }, + { + "epoch": 1.237789203084833, + "grad_norm": 4.4601731300354, + "learning_rate": 8.672746482329276e-06, + "loss": 0.0785, + "step": 642 + }, + { + "epoch": 1.243573264781491, + "grad_norm": 5.807482719421387, + "learning_rate": 8.591002635168603e-06, + "loss": 0.2074, + "step": 645 + }, + { + "epoch": 1.249357326478149, + "grad_norm": 5.01522970199585, + "learning_rate": 8.509208617022934e-06, + "loss": 0.0647, + "step": 648 + }, + { + "epoch": 1.2532133676092545, + "eval_Qnli-dev_cosine_accuracy": 0.7109375, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6890871524810791, + "eval_Qnli-dev_cosine_ap": 0.7543642890383933, + "eval_Qnli-dev_cosine_f1": 0.7012987012987013, + "eval_Qnli-dev_cosine_f1_threshold": 0.6200203895568848, + "eval_Qnli-dev_cosine_precision": 0.6237623762376238, + "eval_Qnli-dev_cosine_recall": 0.8008474576271186, + "eval_Qnli-dev_dot_accuracy": 0.6875, + "eval_Qnli-dev_dot_accuracy_threshold": 681.445556640625, + "eval_Qnli-dev_dot_ap": 0.7317749317318317, + "eval_Qnli-dev_dot_f1": 0.6935201401050788, + "eval_Qnli-dev_dot_f1_threshold": 555.4764404296875, + "eval_Qnli-dev_dot_precision": 0.591044776119403, + "eval_Qnli-dev_dot_recall": 0.8389830508474576, + "eval_Qnli-dev_euclidean_accuracy": 0.71484375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 23.2615966796875, + "eval_Qnli-dev_euclidean_ap": 0.7573423585511763, + "eval_Qnli-dev_euclidean_f1": 0.7005444646098005, + "eval_Qnli-dev_euclidean_f1_threshold": 27.225067138671875, + "eval_Qnli-dev_euclidean_precision": 0.6126984126984127, + "eval_Qnli-dev_euclidean_recall": 0.8177966101694916, + "eval_Qnli-dev_manhattan_accuracy": 0.71875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 749.3499755859375, + "eval_Qnli-dev_manhattan_ap": 0.7598790620995209, + "eval_Qnli-dev_manhattan_f1": 0.704950495049505, + "eval_Qnli-dev_manhattan_f1_threshold": 805.7199096679688, + "eval_Qnli-dev_manhattan_precision": 0.6617100371747212, + "eval_Qnli-dev_manhattan_recall": 0.7542372881355932, + "eval_Qnli-dev_max_accuracy": 0.71875, + "eval_Qnli-dev_max_accuracy_threshold": 749.3499755859375, + "eval_Qnli-dev_max_ap": 0.7598790620995209, + "eval_Qnli-dev_max_f1": 0.704950495049505, + "eval_Qnli-dev_max_f1_threshold": 805.7199096679688, + "eval_Qnli-dev_max_precision": 0.6617100371747212, + "eval_Qnli-dev_max_recall": 0.8389830508474576, + "eval_allNLI-dev_cosine_accuracy": 0.7265625, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.769247829914093, + "eval_allNLI-dev_cosine_ap": 0.6177014050456715, + "eval_allNLI-dev_cosine_f1": 0.6238095238095238, + "eval_allNLI-dev_cosine_f1_threshold": 0.6566067337989807, + "eval_allNLI-dev_cosine_precision": 0.5303643724696356, + "eval_allNLI-dev_cosine_recall": 0.7572254335260116, + "eval_allNLI-dev_dot_accuracy": 0.732421875, + "eval_allNLI-dev_dot_accuracy_threshold": 775.36865234375, + "eval_allNLI-dev_dot_ap": 0.6286796748413648, + "eval_allNLI-dev_dot_f1": 0.6244897959183673, + "eval_allNLI-dev_dot_f1_threshold": 600.73828125, + "eval_allNLI-dev_dot_precision": 0.48264984227129337, + "eval_allNLI-dev_dot_recall": 0.884393063583815, + "eval_allNLI-dev_euclidean_accuracy": 0.724609375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 22.436595916748047, + "eval_allNLI-dev_euclidean_ap": 0.6116063370357722, + "eval_allNLI-dev_euclidean_f1": 0.6279069767441859, + "eval_allNLI-dev_euclidean_f1_threshold": 26.56549072265625, + "eval_allNLI-dev_euclidean_precision": 0.5252918287937743, + "eval_allNLI-dev_euclidean_recall": 0.7803468208092486, + "eval_allNLI-dev_manhattan_accuracy": 0.7265625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 712.5651245117188, + "eval_allNLI-dev_manhattan_ap": 0.6131873293605454, + "eval_allNLI-dev_manhattan_f1": 0.6295399515738498, + "eval_allNLI-dev_manhattan_f1_threshold": 808.766357421875, + "eval_allNLI-dev_manhattan_precision": 0.5416666666666666, + "eval_allNLI-dev_manhattan_recall": 0.7514450867052023, + "eval_allNLI-dev_max_accuracy": 0.732421875, + "eval_allNLI-dev_max_accuracy_threshold": 775.36865234375, + "eval_allNLI-dev_max_ap": 0.6286796748413648, + "eval_allNLI-dev_max_f1": 0.6295399515738498, + "eval_allNLI-dev_max_f1_threshold": 808.766357421875, + "eval_allNLI-dev_max_precision": 0.5416666666666666, + "eval_allNLI-dev_max_recall": 0.884393063583815, + "eval_sequential_score": 0.7598790620995209, + "eval_sts-test_pearson_cosine": 0.9140268771415342, + "eval_sts-test_pearson_dot": 0.901979495096235, + "eval_sts-test_pearson_euclidean": 0.9295051825009135, + "eval_sts-test_pearson_manhattan": 0.928212669281842, + "eval_sts-test_pearson_max": 0.9295051825009135, + "eval_sts-test_spearman_cosine": 0.9222207931318507, + "eval_sts-test_spearman_dot": 0.9016364211758645, + "eval_sts-test_spearman_euclidean": 0.923304042165091, + "eval_sts-test_spearman_manhattan": 0.9218368720603738, + "eval_sts-test_spearman_max": 0.923304042165091, + "eval_vitaminc-pairs_loss": 3.371332883834839, + "eval_vitaminc-pairs_runtime": 4.5289, + "eval_vitaminc-pairs_samples_per_second": 28.263, + "eval_vitaminc-pairs_steps_per_second": 0.221, + "step": 650 + }, + { + "epoch": 1.2532133676092545, + "eval_negation-triplets_loss": 0.5323299169540405, + "eval_negation-triplets_runtime": 3.4511, + "eval_negation-triplets_samples_per_second": 37.09, + "eval_negation-triplets_steps_per_second": 0.29, + "step": 650 + }, + { + "epoch": 1.2532133676092545, + "eval_scitail-pairs-pos_loss": 0.010347447358071804, + "eval_scitail-pairs-pos_runtime": 2.7904, + "eval_scitail-pairs-pos_samples_per_second": 45.871, + "eval_scitail-pairs-pos_steps_per_second": 0.358, + "step": 650 + }, + { + "epoch": 1.2532133676092545, + "eval_scitail-pairs-qa_loss": 0.000511456688400358, + "eval_scitail-pairs-qa_runtime": 2.2825, + "eval_scitail-pairs-qa_samples_per_second": 56.08, + "eval_scitail-pairs-qa_steps_per_second": 0.438, + "step": 650 + }, + { + "epoch": 1.2532133676092545, + "eval_xsum-pairs_loss": 0.002647328656166792, + "eval_xsum-pairs_runtime": 3.0778, + "eval_xsum-pairs_samples_per_second": 41.588, + "eval_xsum-pairs_steps_per_second": 0.325, + "step": 650 + }, + { + "epoch": 1.2532133676092545, + "eval_sciq_pairs_loss": 0.07038552314043045, + "eval_sciq_pairs_runtime": 4.834, + "eval_sciq_pairs_samples_per_second": 26.479, + "eval_sciq_pairs_steps_per_second": 0.207, + "step": 650 + }, + { + "epoch": 1.2532133676092545, + "eval_qasc_pairs_loss": 0.047020211815834045, + "eval_qasc_pairs_runtime": 2.1495, + "eval_qasc_pairs_samples_per_second": 59.549, + "eval_qasc_pairs_steps_per_second": 0.465, + "step": 650 + }, + { + "epoch": 1.2532133676092545, + "eval_openbookqa_pairs_loss": 0.34217798709869385, + "eval_openbookqa_pairs_runtime": 2.2685, + "eval_openbookqa_pairs_samples_per_second": 56.426, + "eval_openbookqa_pairs_steps_per_second": 0.441, + "step": 650 + }, + { + "epoch": 1.2532133676092545, + "eval_msmarco_pairs_loss": 0.012268541380763054, + "eval_msmarco_pairs_runtime": 2.2172, + "eval_msmarco_pairs_samples_per_second": 57.73, + "eval_msmarco_pairs_steps_per_second": 0.451, + "step": 650 + }, + { + "epoch": 1.2532133676092545, + "eval_nq_pairs_loss": 0.019458645954728127, + "eval_nq_pairs_runtime": 3.7007, + "eval_nq_pairs_samples_per_second": 34.588, + "eval_nq_pairs_steps_per_second": 0.27, + "step": 650 + }, + { + "epoch": 1.2532133676092545, + "eval_trivia_pairs_loss": 0.008974222466349602, + "eval_trivia_pairs_runtime": 3.1201, + "eval_trivia_pairs_samples_per_second": 41.024, + "eval_trivia_pairs_steps_per_second": 0.321, + "step": 650 + }, + { + "epoch": 1.2532133676092545, + "eval_gooaq_pairs_loss": 0.05843409150838852, + "eval_gooaq_pairs_runtime": 2.2465, + "eval_gooaq_pairs_samples_per_second": 56.978, + "eval_gooaq_pairs_steps_per_second": 0.445, + "step": 650 + }, + { + "epoch": 1.2532133676092545, + "eval_paws-pos_loss": 0.027927419170737267, + "eval_paws-pos_runtime": 2.3507, + "eval_paws-pos_samples_per_second": 54.452, + "eval_paws-pos_steps_per_second": 0.425, + "step": 650 + }, + { + "epoch": 1.2532133676092545, + "eval_global_dataset_loss": 0.1904364973306656, + "eval_global_dataset_runtime": 8.4532, + "eval_global_dataset_samples_per_second": 38.447, + "eval_global_dataset_steps_per_second": 0.355, + "step": 650 + }, + { + "epoch": 1.2551413881748072, + "grad_norm": 1.4944630861282349, + "learning_rate": 8.427376462070407e-06, + "loss": 0.0664, + "step": 651 + }, + { + "epoch": 1.2609254498714653, + "grad_norm": 1.80479097366333, + "learning_rate": 8.345518210100145e-06, + "loss": 0.012, + "step": 654 + }, + { + "epoch": 1.2667095115681235, + "grad_norm": 2.6618971824645996, + "learning_rate": 8.263645904740871e-06, + "loss": 0.1048, + "step": 657 + }, + { + "epoch": 1.2724935732647815, + "grad_norm": 1.700681447982788, + "learning_rate": 8.181771591688956e-06, + "loss": 0.0221, + "step": 660 + }, + { + "epoch": 1.2782776349614395, + "grad_norm": 1.8328262567520142, + "learning_rate": 8.09990731693615e-06, + "loss": 0.1205, + "step": 663 + }, + { + "epoch": 1.2840616966580978, + "grad_norm": 3.5405755043029785, + "learning_rate": 8.0180651249973e-06, + "loss": 0.1324, + "step": 666 + }, + { + "epoch": 1.2898457583547558, + "grad_norm": 1.9436181783676147, + "learning_rate": 7.936257057138247e-06, + "loss": 0.0897, + "step": 669 + }, + { + "epoch": 1.2956298200514138, + "grad_norm": 2.8245396614074707, + "learning_rate": 7.854495149604237e-06, + "loss": 0.0388, + "step": 672 + }, + { + "epoch": 1.301413881748072, + "grad_norm": 1.1756666898727417, + "learning_rate": 7.772791431849036e-06, + "loss": 0.0425, + "step": 675 + }, + { + "epoch": 1.3033419023136248, + "eval_Qnli-dev_cosine_accuracy": 0.712890625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6512588262557983, + "eval_Qnli-dev_cosine_ap": 0.7569407635371652, + "eval_Qnli-dev_cosine_f1": 0.7045454545454546, + "eval_Qnli-dev_cosine_f1_threshold": 0.6241868734359741, + "eval_Qnli-dev_cosine_precision": 0.636986301369863, + "eval_Qnli-dev_cosine_recall": 0.788135593220339, + "eval_Qnli-dev_dot_accuracy": 0.6953125, + "eval_Qnli-dev_dot_accuracy_threshold": 650.5750732421875, + "eval_Qnli-dev_dot_ap": 0.7340779693577646, + "eval_Qnli-dev_dot_f1": 0.6950596252129473, + "eval_Qnli-dev_dot_f1_threshold": 543.4408569335938, + "eval_Qnli-dev_dot_precision": 0.5811965811965812, + "eval_Qnli-dev_dot_recall": 0.864406779661017, + "eval_Qnli-dev_euclidean_accuracy": 0.716796875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 24.167133331298828, + "eval_Qnli-dev_euclidean_ap": 0.7606793766515915, + "eval_Qnli-dev_euclidean_f1": 0.7017543859649122, + "eval_Qnli-dev_euclidean_f1_threshold": 26.369770050048828, + "eval_Qnli-dev_euclidean_precision": 0.6498194945848376, + "eval_Qnli-dev_euclidean_recall": 0.7627118644067796, + "eval_Qnli-dev_manhattan_accuracy": 0.720703125, + "eval_Qnli-dev_manhattan_accuracy_threshold": 749.0418701171875, + "eval_Qnli-dev_manhattan_ap": 0.7620762177824322, + "eval_Qnli-dev_manhattan_f1": 0.7132352941176471, + "eval_Qnli-dev_manhattan_f1_threshold": 847.3385009765625, + "eval_Qnli-dev_manhattan_precision": 0.6298701298701299, + "eval_Qnli-dev_manhattan_recall": 0.8220338983050848, + "eval_Qnli-dev_max_accuracy": 0.720703125, + "eval_Qnli-dev_max_accuracy_threshold": 749.0418701171875, + "eval_Qnli-dev_max_ap": 0.7620762177824322, + "eval_Qnli-dev_max_f1": 0.7132352941176471, + "eval_Qnli-dev_max_f1_threshold": 847.3385009765625, + "eval_Qnli-dev_max_precision": 0.6498194945848376, + "eval_Qnli-dev_max_recall": 0.864406779661017, + "eval_allNLI-dev_cosine_accuracy": 0.724609375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7729952335357666, + "eval_allNLI-dev_cosine_ap": 0.6070093722956825, + "eval_allNLI-dev_cosine_f1": 0.630841121495327, + "eval_allNLI-dev_cosine_f1_threshold": 0.643187940120697, + "eval_allNLI-dev_cosine_precision": 0.5294117647058824, + "eval_allNLI-dev_cosine_recall": 0.7803468208092486, + "eval_allNLI-dev_dot_accuracy": 0.732421875, + "eval_allNLI-dev_dot_accuracy_threshold": 768.6075439453125, + "eval_allNLI-dev_dot_ap": 0.6191366112486252, + "eval_allNLI-dev_dot_f1": 0.6255707762557078, + "eval_allNLI-dev_dot_f1_threshold": 628.1797485351562, + "eval_allNLI-dev_dot_precision": 0.5169811320754717, + "eval_allNLI-dev_dot_recall": 0.791907514450867, + "eval_allNLI-dev_euclidean_accuracy": 0.724609375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 20.34807586669922, + "eval_allNLI-dev_euclidean_ap": 0.6033144000557898, + "eval_allNLI-dev_euclidean_f1": 0.6298076923076924, + "eval_allNLI-dev_euclidean_f1_threshold": 26.242013931274414, + "eval_allNLI-dev_euclidean_precision": 0.5390946502057613, + "eval_allNLI-dev_euclidean_recall": 0.7572254335260116, + "eval_allNLI-dev_manhattan_accuracy": 0.72265625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 693.0238037109375, + "eval_allNLI-dev_manhattan_ap": 0.6056066138614866, + "eval_allNLI-dev_manhattan_f1": 0.6388206388206389, + "eval_allNLI-dev_manhattan_f1_threshold": 811.5244140625, + "eval_allNLI-dev_manhattan_precision": 0.5555555555555556, + "eval_allNLI-dev_manhattan_recall": 0.7514450867052023, + "eval_allNLI-dev_max_accuracy": 0.732421875, + "eval_allNLI-dev_max_accuracy_threshold": 768.6075439453125, + "eval_allNLI-dev_max_ap": 0.6191366112486252, + "eval_allNLI-dev_max_f1": 0.6388206388206389, + "eval_allNLI-dev_max_f1_threshold": 811.5244140625, + "eval_allNLI-dev_max_precision": 0.5555555555555556, + "eval_allNLI-dev_max_recall": 0.791907514450867, + "eval_sequential_score": 0.7620762177824322, + "eval_sts-test_pearson_cosine": 0.9165898359667968, + "eval_sts-test_pearson_dot": 0.9069315524505921, + "eval_sts-test_pearson_euclidean": 0.9302573136297625, + "eval_sts-test_pearson_manhattan": 0.9290448014021928, + "eval_sts-test_pearson_max": 0.9302573136297625, + "eval_sts-test_spearman_cosine": 0.9232708814803999, + "eval_sts-test_spearman_dot": 0.904779928781042, + "eval_sts-test_spearman_euclidean": 0.9229526552606455, + "eval_sts-test_spearman_manhattan": 0.9218775509569787, + "eval_sts-test_spearman_max": 0.9232708814803999, + "eval_vitaminc-pairs_loss": 3.6177520751953125, + "eval_vitaminc-pairs_runtime": 4.4446, + "eval_vitaminc-pairs_samples_per_second": 28.799, + "eval_vitaminc-pairs_steps_per_second": 0.225, + "step": 676 + }, + { + "epoch": 1.3033419023136248, + "eval_negation-triplets_loss": 0.5532333254814148, + "eval_negation-triplets_runtime": 3.2635, + "eval_negation-triplets_samples_per_second": 39.221, + "eval_negation-triplets_steps_per_second": 0.306, + "step": 676 + }, + { + "epoch": 1.3033419023136248, + "eval_scitail-pairs-pos_loss": 0.008254562504589558, + "eval_scitail-pairs-pos_runtime": 2.6485, + "eval_scitail-pairs-pos_samples_per_second": 48.33, + "eval_scitail-pairs-pos_steps_per_second": 0.378, + "step": 676 + }, + { + "epoch": 1.3033419023136248, + "eval_scitail-pairs-qa_loss": 0.0011200320441275835, + "eval_scitail-pairs-qa_runtime": 2.1814, + "eval_scitail-pairs-qa_samples_per_second": 58.678, + "eval_scitail-pairs-qa_steps_per_second": 0.458, + "step": 676 + }, + { + "epoch": 1.3033419023136248, + "eval_xsum-pairs_loss": 0.0031933763530105352, + "eval_xsum-pairs_runtime": 3.0641, + "eval_xsum-pairs_samples_per_second": 41.775, + "eval_xsum-pairs_steps_per_second": 0.326, + "step": 676 + }, + { + "epoch": 1.3033419023136248, + "eval_sciq_pairs_loss": 0.07220430672168732, + "eval_sciq_pairs_runtime": 4.8122, + "eval_sciq_pairs_samples_per_second": 26.599, + "eval_sciq_pairs_steps_per_second": 0.208, + "step": 676 + }, + { + "epoch": 1.3033419023136248, + "eval_qasc_pairs_loss": 0.04014963656663895, + "eval_qasc_pairs_runtime": 2.0909, + "eval_qasc_pairs_samples_per_second": 61.218, + "eval_qasc_pairs_steps_per_second": 0.478, + "step": 676 + }, + { + "epoch": 1.3033419023136248, + "eval_openbookqa_pairs_loss": 0.36179760098457336, + "eval_openbookqa_pairs_runtime": 2.2247, + "eval_openbookqa_pairs_samples_per_second": 57.537, + "eval_openbookqa_pairs_steps_per_second": 0.45, + "step": 676 + }, + { + "epoch": 1.3033419023136248, + "eval_msmarco_pairs_loss": 0.0059233661741018295, + "eval_msmarco_pairs_runtime": 2.3146, + "eval_msmarco_pairs_samples_per_second": 55.302, + "eval_msmarco_pairs_steps_per_second": 0.432, + "step": 676 + }, + { + "epoch": 1.3033419023136248, + "eval_nq_pairs_loss": 0.01814723014831543, + "eval_nq_pairs_runtime": 3.655, + "eval_nq_pairs_samples_per_second": 35.021, + "eval_nq_pairs_steps_per_second": 0.274, + "step": 676 + }, + { + "epoch": 1.3033419023136248, + "eval_trivia_pairs_loss": 0.00952482782304287, + "eval_trivia_pairs_runtime": 3.0846, + "eval_trivia_pairs_samples_per_second": 41.497, + "eval_trivia_pairs_steps_per_second": 0.324, + "step": 676 + }, + { + "epoch": 1.3033419023136248, + "eval_gooaq_pairs_loss": 0.04494225233793259, + "eval_gooaq_pairs_runtime": 2.1152, + "eval_gooaq_pairs_samples_per_second": 60.514, + "eval_gooaq_pairs_steps_per_second": 0.473, + "step": 676 + }, + { + "epoch": 1.3033419023136248, + "eval_paws-pos_loss": 0.02739187330007553, + "eval_paws-pos_runtime": 2.2405, + "eval_paws-pos_samples_per_second": 57.129, + "eval_paws-pos_steps_per_second": 0.446, + "step": 676 + }, + { + "epoch": 1.3033419023136248, + "eval_global_dataset_loss": 0.21470122039318085, + "eval_global_dataset_runtime": 8.3659, + "eval_global_dataset_samples_per_second": 38.848, + "eval_global_dataset_steps_per_second": 0.359, + "step": 676 + }, + { + "epoch": 1.30719794344473, + "grad_norm": 4.4491286277771, + "learning_rate": 7.691157924765089e-06, + "loss": 0.0356, + "step": 678 + }, + { + "epoch": 1.312982005141388, + "grad_norm": 3.6744799613952637, + "learning_rate": 7.60960663891488e-06, + "loss": 0.0997, + "step": 681 + }, + { + "epoch": 1.3187660668380463, + "grad_norm": 2.078890085220337, + "learning_rate": 7.528149572763867e-06, + "loss": 0.0517, + "step": 684 + }, + { + "epoch": 1.3245501285347043, + "grad_norm": 1.3222216367721558, + "learning_rate": 7.446798710915168e-06, + "loss": 0.0348, + "step": 687 + }, + { + "epoch": 1.3303341902313623, + "grad_norm": 7.282771110534668, + "learning_rate": 7.365566022346283e-06, + "loss": 0.2454, + "step": 690 + }, + { + "epoch": 1.3361182519280206, + "grad_norm": 3.041977643966675, + "learning_rate": 7.28446345864814e-06, + "loss": 0.0975, + "step": 693 + }, + { + "epoch": 1.3419023136246786, + "grad_norm": 6.863183498382568, + "learning_rate": 7.203502952266664e-06, + "loss": 0.1477, + "step": 696 + }, + { + "epoch": 1.3476863753213368, + "grad_norm": 3.121077537536621, + "learning_rate": 7.122696414747207e-06, + "loss": 0.0737, + "step": 699 + }, + { + "epoch": 1.3534704370179949, + "grad_norm": 5.715446949005127, + "learning_rate": 7.042055734981999e-06, + "loss": 0.1786, + "step": 702 + }, + { + "epoch": 1.3534704370179949, + "eval_Qnli-dev_cosine_accuracy": 0.73046875, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6626905798912048, + "eval_Qnli-dev_cosine_ap": 0.7709539394033288, + "eval_Qnli-dev_cosine_f1": 0.7109375, + "eval_Qnli-dev_cosine_f1_threshold": 0.6342096328735352, + "eval_Qnli-dev_cosine_precision": 0.6594202898550725, + "eval_Qnli-dev_cosine_recall": 0.7711864406779662, + "eval_Qnli-dev_dot_accuracy": 0.705078125, + "eval_Qnli-dev_dot_accuracy_threshold": 648.3182373046875, + "eval_Qnli-dev_dot_ap": 0.744965122332887, + "eval_Qnli-dev_dot_f1": 0.704424778761062, + "eval_Qnli-dev_dot_f1_threshold": 544.717529296875, + "eval_Qnli-dev_dot_precision": 0.6048632218844985, + "eval_Qnli-dev_dot_recall": 0.8432203389830508, + "eval_Qnli-dev_euclidean_accuracy": 0.732421875, + "eval_Qnli-dev_euclidean_accuracy_threshold": 24.26491355895996, + "eval_Qnli-dev_euclidean_ap": 0.7731490281194371, + "eval_Qnli-dev_euclidean_f1": 0.7164179104477612, + "eval_Qnli-dev_euclidean_f1_threshold": 26.51512908935547, + "eval_Qnli-dev_euclidean_precision": 0.64, + "eval_Qnli-dev_euclidean_recall": 0.8135593220338984, + "eval_Qnli-dev_manhattan_accuracy": 0.734375, + "eval_Qnli-dev_manhattan_accuracy_threshold": 774.825439453125, + "eval_Qnli-dev_manhattan_ap": 0.7760404255129649, + "eval_Qnli-dev_manhattan_f1": 0.7220216606498194, + "eval_Qnli-dev_manhattan_f1_threshold": 841.4962158203125, + "eval_Qnli-dev_manhattan_precision": 0.6289308176100629, + "eval_Qnli-dev_manhattan_recall": 0.847457627118644, + "eval_Qnli-dev_max_accuracy": 0.734375, + "eval_Qnli-dev_max_accuracy_threshold": 774.825439453125, + "eval_Qnli-dev_max_ap": 0.7760404255129649, + "eval_Qnli-dev_max_f1": 0.7220216606498194, + "eval_Qnli-dev_max_f1_threshold": 841.4962158203125, + "eval_Qnli-dev_max_precision": 0.6594202898550725, + "eval_Qnli-dev_max_recall": 0.847457627118644, + "eval_allNLI-dev_cosine_accuracy": 0.734375, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7453016042709351, + "eval_allNLI-dev_cosine_ap": 0.6099662458742457, + "eval_allNLI-dev_cosine_f1": 0.6206896551724138, + "eval_allNLI-dev_cosine_f1_threshold": 0.6737011075019836, + "eval_allNLI-dev_cosine_precision": 0.5407725321888412, + "eval_allNLI-dev_cosine_recall": 0.7283236994219653, + "eval_allNLI-dev_dot_accuracy": 0.740234375, + "eval_allNLI-dev_dot_accuracy_threshold": 758.638916015625, + "eval_allNLI-dev_dot_ap": 0.6212065635838847, + "eval_allNLI-dev_dot_f1": 0.6192660550458715, + "eval_allNLI-dev_dot_f1_threshold": 631.0904541015625, + "eval_allNLI-dev_dot_precision": 0.5133079847908745, + "eval_allNLI-dev_dot_recall": 0.7803468208092486, + "eval_allNLI-dev_euclidean_accuracy": 0.7265625, + "eval_allNLI-dev_euclidean_accuracy_threshold": 21.186803817749023, + "eval_allNLI-dev_euclidean_ap": 0.6038175128262446, + "eval_allNLI-dev_euclidean_f1": 0.6272727272727273, + "eval_allNLI-dev_euclidean_f1_threshold": 26.413116455078125, + "eval_allNLI-dev_euclidean_precision": 0.5168539325842697, + "eval_allNLI-dev_euclidean_recall": 0.7976878612716763, + "eval_allNLI-dev_manhattan_accuracy": 0.724609375, + "eval_allNLI-dev_manhattan_accuracy_threshold": 650.3385009765625, + "eval_allNLI-dev_manhattan_ap": 0.6060814227124063, + "eval_allNLI-dev_manhattan_f1": 0.6292134831460674, + "eval_allNLI-dev_manhattan_f1_threshold": 832.4495849609375, + "eval_allNLI-dev_manhattan_precision": 0.5147058823529411, + "eval_allNLI-dev_manhattan_recall": 0.8092485549132948, + "eval_allNLI-dev_max_accuracy": 0.740234375, + "eval_allNLI-dev_max_accuracy_threshold": 758.638916015625, + "eval_allNLI-dev_max_ap": 0.6212065635838847, + "eval_allNLI-dev_max_f1": 0.6292134831460674, + "eval_allNLI-dev_max_f1_threshold": 832.4495849609375, + "eval_allNLI-dev_max_precision": 0.5407725321888412, + "eval_allNLI-dev_max_recall": 0.8092485549132948, + "eval_sequential_score": 0.7760404255129649, + "eval_sts-test_pearson_cosine": 0.9186535830521056, + "eval_sts-test_pearson_dot": 0.9055455832051154, + "eval_sts-test_pearson_euclidean": 0.9328012897337526, + "eval_sts-test_pearson_manhattan": 0.9317657951283651, + "eval_sts-test_pearson_max": 0.9328012897337526, + "eval_sts-test_spearman_cosine": 0.92503886956419, + "eval_sts-test_spearman_dot": 0.902418538971202, + "eval_sts-test_spearman_euclidean": 0.9258251044755754, + "eval_sts-test_spearman_manhattan": 0.924931645541896, + "eval_sts-test_spearman_max": 0.9258251044755754, + "eval_vitaminc-pairs_loss": 3.024595022201538, + "eval_vitaminc-pairs_runtime": 4.5839, + "eval_vitaminc-pairs_samples_per_second": 27.924, + "eval_vitaminc-pairs_steps_per_second": 0.218, + "step": 702 + }, + { + "epoch": 1.3534704370179949, + "eval_negation-triplets_loss": 0.5546241998672485, + "eval_negation-triplets_runtime": 3.4431, + "eval_negation-triplets_samples_per_second": 37.176, + "eval_negation-triplets_steps_per_second": 0.29, + "step": 702 + }, + { + "epoch": 1.3534704370179949, + "eval_scitail-pairs-pos_loss": 0.014527450315654278, + "eval_scitail-pairs-pos_runtime": 2.8659, + "eval_scitail-pairs-pos_samples_per_second": 44.664, + "eval_scitail-pairs-pos_steps_per_second": 0.349, + "step": 702 + }, + { + "epoch": 1.3534704370179949, + "eval_scitail-pairs-qa_loss": 0.0013152618193998933, + "eval_scitail-pairs-qa_runtime": 2.3599, + "eval_scitail-pairs-qa_samples_per_second": 54.24, + "eval_scitail-pairs-qa_steps_per_second": 0.424, + "step": 702 + }, + { + "epoch": 1.3534704370179949, + "eval_xsum-pairs_loss": 0.004006410017609596, + "eval_xsum-pairs_runtime": 3.0974, + "eval_xsum-pairs_samples_per_second": 41.325, + "eval_xsum-pairs_steps_per_second": 0.323, + "step": 702 + }, + { + "epoch": 1.3534704370179949, + "eval_sciq_pairs_loss": 0.06901878863573074, + "eval_sciq_pairs_runtime": 4.9941, + "eval_sciq_pairs_samples_per_second": 25.63, + "eval_sciq_pairs_steps_per_second": 0.2, + "step": 702 + }, + { + "epoch": 1.3534704370179949, + "eval_qasc_pairs_loss": 0.03258923813700676, + "eval_qasc_pairs_runtime": 2.1764, + "eval_qasc_pairs_samples_per_second": 58.814, + "eval_qasc_pairs_steps_per_second": 0.459, + "step": 702 + }, + { + "epoch": 1.3534704370179949, + "eval_openbookqa_pairs_loss": 0.3399459421634674, + "eval_openbookqa_pairs_runtime": 2.3295, + "eval_openbookqa_pairs_samples_per_second": 54.949, + "eval_openbookqa_pairs_steps_per_second": 0.429, + "step": 702 + }, + { + "epoch": 1.3534704370179949, + "eval_msmarco_pairs_loss": 0.008152350783348083, + "eval_msmarco_pairs_runtime": 2.3449, + "eval_msmarco_pairs_samples_per_second": 54.586, + "eval_msmarco_pairs_steps_per_second": 0.426, + "step": 702 + }, + { + "epoch": 1.3534704370179949, + "eval_nq_pairs_loss": 0.01521628350019455, + "eval_nq_pairs_runtime": 3.6786, + "eval_nq_pairs_samples_per_second": 34.796, + "eval_nq_pairs_steps_per_second": 0.272, + "step": 702 + }, + { + "epoch": 1.3534704370179949, + "eval_trivia_pairs_loss": 0.00645024236291647, + "eval_trivia_pairs_runtime": 3.2376, + "eval_trivia_pairs_samples_per_second": 39.536, + "eval_trivia_pairs_steps_per_second": 0.309, + "step": 702 + }, + { + "epoch": 1.3534704370179949, + "eval_gooaq_pairs_loss": 0.0335497222840786, + "eval_gooaq_pairs_runtime": 2.199, + "eval_gooaq_pairs_samples_per_second": 58.209, + "eval_gooaq_pairs_steps_per_second": 0.455, + "step": 702 + }, + { + "epoch": 1.3534704370179949, + "eval_paws-pos_loss": 0.02810763008892536, + "eval_paws-pos_runtime": 2.2998, + "eval_paws-pos_samples_per_second": 55.657, + "eval_paws-pos_steps_per_second": 0.435, + "step": 702 + }, + { + "epoch": 1.3534704370179949, + "eval_global_dataset_loss": 0.1622796654701233, + "eval_global_dataset_runtime": 8.4923, + "eval_global_dataset_samples_per_second": 38.27, + "eval_global_dataset_steps_per_second": 0.353, + "step": 702 + }, + { + "epoch": 1.359254498714653, + "grad_norm": 11.614518165588379, + "learning_rate": 6.9615927774609815e-06, + "loss": 0.1141, + "step": 705 + }, + { + "epoch": 1.365038560411311, + "grad_norm": 5.832677364349365, + "learning_rate": 6.881319380526202e-06, + "loss": 0.072, + "step": 708 + }, + { + "epoch": 1.3708226221079691, + "grad_norm": 1.1572054624557495, + "learning_rate": 6.801247354630069e-06, + "loss": 0.0617, + "step": 711 + }, + { + "epoch": 1.3766066838046274, + "grad_norm": 2.936616897583008, + "learning_rate": 6.7213884805977e-06, + "loss": 0.0886, + "step": 714 + }, + { + "epoch": 1.3823907455012854, + "grad_norm": 1.7748816013336182, + "learning_rate": 6.641754507893638e-06, + "loss": 0.047, + "step": 717 + }, + { + "epoch": 1.3881748071979434, + "grad_norm": 3.0840072631835938, + "learning_rate": 6.5623571528931755e-06, + "loss": 0.1327, + "step": 720 + }, + { + "epoch": 1.3939588688946016, + "grad_norm": 4.223444938659668, + "learning_rate": 6.483208097158554e-06, + "loss": 0.0498, + "step": 723 + }, + { + "epoch": 1.3997429305912596, + "grad_norm": 3.0616002082824707, + "learning_rate": 6.404318985720274e-06, + "loss": 0.0671, + "step": 726 + }, + { + "epoch": 1.403598971722365, + "eval_Qnli-dev_cosine_accuracy": 0.72265625, + "eval_Qnli-dev_cosine_accuracy_threshold": 0.6773581504821777, + "eval_Qnli-dev_cosine_ap": 0.7602286178078475, + "eval_Qnli-dev_cosine_f1": 0.70703125, + "eval_Qnli-dev_cosine_f1_threshold": 0.6473040580749512, + "eval_Qnli-dev_cosine_precision": 0.6557971014492754, + "eval_Qnli-dev_cosine_recall": 0.7669491525423728, + "eval_Qnli-dev_dot_accuracy": 0.697265625, + "eval_Qnli-dev_dot_accuracy_threshold": 651.5748291015625, + "eval_Qnli-dev_dot_ap": 0.7285043251404486, + "eval_Qnli-dev_dot_f1": 0.6933333333333332, + "eval_Qnli-dev_dot_f1_threshold": 532.1196899414062, + "eval_Qnli-dev_dot_precision": 0.5714285714285714, + "eval_Qnli-dev_dot_recall": 0.8813559322033898, + "eval_Qnli-dev_euclidean_accuracy": 0.724609375, + "eval_Qnli-dev_euclidean_accuracy_threshold": 23.379541397094727, + "eval_Qnli-dev_euclidean_ap": 0.7644183411294448, + "eval_Qnli-dev_euclidean_f1": 0.7067961165048545, + "eval_Qnli-dev_euclidean_f1_threshold": 25.545307159423828, + "eval_Qnli-dev_euclidean_precision": 0.6523297491039427, + "eval_Qnli-dev_euclidean_recall": 0.7711864406779662, + "eval_Qnli-dev_manhattan_accuracy": 0.73046875, + "eval_Qnli-dev_manhattan_accuracy_threshold": 735.138916015625, + "eval_Qnli-dev_manhattan_ap": 0.7675072683854777, + "eval_Qnli-dev_manhattan_f1": 0.7129798903107861, + "eval_Qnli-dev_manhattan_f1_threshold": 823.470458984375, + "eval_Qnli-dev_manhattan_precision": 0.6270096463022508, + "eval_Qnli-dev_manhattan_recall": 0.826271186440678, + "eval_Qnli-dev_max_accuracy": 0.73046875, + "eval_Qnli-dev_max_accuracy_threshold": 735.138916015625, + "eval_Qnli-dev_max_ap": 0.7675072683854777, + "eval_Qnli-dev_max_f1": 0.7129798903107861, + "eval_Qnli-dev_max_f1_threshold": 823.470458984375, + "eval_Qnli-dev_max_precision": 0.6557971014492754, + "eval_Qnli-dev_max_recall": 0.8813559322033898, + "eval_allNLI-dev_cosine_accuracy": 0.73046875, + "eval_allNLI-dev_cosine_accuracy_threshold": 0.7678493857383728, + "eval_allNLI-dev_cosine_ap": 0.6081951857631663, + "eval_allNLI-dev_cosine_f1": 0.6274509803921569, + "eval_allNLI-dev_cosine_f1_threshold": 0.6708965301513672, + "eval_allNLI-dev_cosine_precision": 0.5446808510638298, + "eval_allNLI-dev_cosine_recall": 0.7398843930635838, + "eval_allNLI-dev_dot_accuracy": 0.7265625, + "eval_allNLI-dev_dot_accuracy_threshold": 765.60693359375, + "eval_allNLI-dev_dot_ap": 0.6126465375361672, + "eval_allNLI-dev_dot_f1": 0.62004662004662, + "eval_allNLI-dev_dot_f1_threshold": 630.0067138671875, + "eval_allNLI-dev_dot_precision": 0.51953125, + "eval_allNLI-dev_dot_recall": 0.7687861271676301, + "eval_allNLI-dev_euclidean_accuracy": 0.724609375, + "eval_allNLI-dev_euclidean_accuracy_threshold": 20.469280242919922, + "eval_allNLI-dev_euclidean_ap": 0.6024805519329677, + "eval_allNLI-dev_euclidean_f1": 0.6264501160092808, + "eval_allNLI-dev_euclidean_f1_threshold": 26.046119689941406, + "eval_allNLI-dev_euclidean_precision": 0.5232558139534884, + "eval_allNLI-dev_euclidean_recall": 0.7803468208092486, + "eval_allNLI-dev_manhattan_accuracy": 0.7265625, + "eval_allNLI-dev_manhattan_accuracy_threshold": 639.059326171875, + "eval_allNLI-dev_manhattan_ap": 0.6030976441964669, + "eval_allNLI-dev_manhattan_f1": 0.6275395033860045, + "eval_allNLI-dev_manhattan_f1_threshold": 820.0479736328125, + "eval_allNLI-dev_manhattan_precision": 0.5148148148148148, + "eval_allNLI-dev_manhattan_recall": 0.8034682080924855, + "eval_allNLI-dev_max_accuracy": 0.73046875, + "eval_allNLI-dev_max_accuracy_threshold": 765.60693359375, + "eval_allNLI-dev_max_ap": 0.6126465375361672, + "eval_allNLI-dev_max_f1": 0.6275395033860045, + "eval_allNLI-dev_max_f1_threshold": 820.0479736328125, + "eval_allNLI-dev_max_precision": 0.5446808510638298, + "eval_allNLI-dev_max_recall": 0.8034682080924855, + "eval_sequential_score": 0.7675072683854777, + "eval_sts-test_pearson_cosine": 0.9181624683007681, + "eval_sts-test_pearson_dot": 0.9026377641625429, + "eval_sts-test_pearson_euclidean": 0.9339740017356808, + "eval_sts-test_pearson_manhattan": 0.9331083687364159, + "eval_sts-test_pearson_max": 0.9339740017356808, + "eval_sts-test_spearman_cosine": 0.9263124457121277, + "eval_sts-test_spearman_dot": 0.8988715539933195, + "eval_sts-test_spearman_euclidean": 0.9273519727629886, + "eval_sts-test_spearman_manhattan": 0.9267412970500414, + "eval_sts-test_spearman_max": 0.9273519727629886, + "eval_vitaminc-pairs_loss": 3.137650489807129, + "eval_vitaminc-pairs_runtime": 4.4935, + "eval_vitaminc-pairs_samples_per_second": 28.485, + "eval_vitaminc-pairs_steps_per_second": 0.223, + "step": 728 + }, + { + "epoch": 1.403598971722365, + "eval_negation-triplets_loss": 0.5126662850379944, + "eval_negation-triplets_runtime": 3.3654, + "eval_negation-triplets_samples_per_second": 38.034, + "eval_negation-triplets_steps_per_second": 0.297, + "step": 728 + }, + { + "epoch": 1.403598971722365, + "eval_scitail-pairs-pos_loss": 0.01808050461113453, + "eval_scitail-pairs-pos_runtime": 2.7048, + "eval_scitail-pairs-pos_samples_per_second": 47.324, + "eval_scitail-pairs-pos_steps_per_second": 0.37, + "step": 728 + }, + { + "epoch": 1.403598971722365, + "eval_scitail-pairs-qa_loss": 0.0016650371253490448, + "eval_scitail-pairs-qa_runtime": 2.2416, + "eval_scitail-pairs-qa_samples_per_second": 57.103, + "eval_scitail-pairs-qa_steps_per_second": 0.446, + "step": 728 + }, + { + "epoch": 1.403598971722365, + "eval_xsum-pairs_loss": 0.00410380307585001, + "eval_xsum-pairs_runtime": 3.0873, + "eval_xsum-pairs_samples_per_second": 41.46, + "eval_xsum-pairs_steps_per_second": 0.324, + "step": 728 + }, + { + "epoch": 1.403598971722365, + "eval_sciq_pairs_loss": 0.0691823959350586, + "eval_sciq_pairs_runtime": 4.8252, + "eval_sciq_pairs_samples_per_second": 26.527, + "eval_sciq_pairs_steps_per_second": 0.207, + "step": 728 + }, + { + "epoch": 1.403598971722365, + "eval_qasc_pairs_loss": 0.039144549518823624, + "eval_qasc_pairs_runtime": 2.1037, + "eval_qasc_pairs_samples_per_second": 60.845, + "eval_qasc_pairs_steps_per_second": 0.475, + "step": 728 + }, + { + "epoch": 1.403598971722365, + "eval_openbookqa_pairs_loss": 0.34394749999046326, + "eval_openbookqa_pairs_runtime": 2.2343, + "eval_openbookqa_pairs_samples_per_second": 57.289, + "eval_openbookqa_pairs_steps_per_second": 0.448, + "step": 728 + }, + { + "epoch": 1.403598971722365, + "eval_msmarco_pairs_loss": 0.009959997609257698, + "eval_msmarco_pairs_runtime": 2.2375, + "eval_msmarco_pairs_samples_per_second": 57.207, + "eval_msmarco_pairs_steps_per_second": 0.447, + "step": 728 + }, + { + "epoch": 1.403598971722365, + "eval_nq_pairs_loss": 0.010519111528992653, + "eval_nq_pairs_runtime": 3.661, + "eval_nq_pairs_samples_per_second": 34.963, + "eval_nq_pairs_steps_per_second": 0.273, + "step": 728 + }, + { + "epoch": 1.403598971722365, + "eval_trivia_pairs_loss": 0.007627342827618122, + "eval_trivia_pairs_runtime": 3.0823, + "eval_trivia_pairs_samples_per_second": 41.527, + "eval_trivia_pairs_steps_per_second": 0.324, + "step": 728 + }, + { + "epoch": 1.403598971722365, + "eval_gooaq_pairs_loss": 0.044966116547584534, + "eval_gooaq_pairs_runtime": 2.131, + "eval_gooaq_pairs_samples_per_second": 60.066, + "eval_gooaq_pairs_steps_per_second": 0.469, + "step": 728 + }, + { + "epoch": 1.403598971722365, + "eval_paws-pos_loss": 0.028517380356788635, + "eval_paws-pos_runtime": 2.2361, + "eval_paws-pos_samples_per_second": 57.243, + "eval_paws-pos_steps_per_second": 0.447, + "step": 728 + }, + { + "epoch": 1.403598971722365, + "eval_global_dataset_loss": 0.16524119675159454, + "eval_global_dataset_runtime": 8.3808, + "eval_global_dataset_samples_per_second": 38.779, + "eval_global_dataset_steps_per_second": 0.358, + "step": 728 + } + ], + "logging_steps": 3, + "max_steps": 1036, + "num_input_tokens_seen": 0, + "num_train_epochs": 2, + "save_steps": 52, + "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": 64, + "trial_name": null, + "trial_params": null +}