{ "best_metric": null, "best_model_checkpoint": null, "epoch": 12.0, "eval_steps": 500, "global_step": 492, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 1.0, "eval_Claim": { "f1-score": 0.3926645091693635, "precision": 0.4813399941228328, "recall": 0.33157894736842103, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.5305101700566855, "precision": 0.41758530183727033, "recall": 0.7271480804387569, "support": 2188.0 }, "eval_O": { "f1-score": 0.8568685802054334, "precision": 0.8614998552263295, "recall": 0.8522868328081734, "support": 10473.0 }, "eval_Premise": { "f1-score": 0.8535428122545169, "precision": 0.8528192892126083, "recall": 0.8542675639977357, "support": 15899.0 }, "eval_accuracy": 0.7682686567164179, "eval_loss": 0.6236585974693298, "eval_macro avg": { "f1-score": 0.6583965179214999, "precision": 0.6533111100997602, "recall": 0.6913203561532717, "support": 33500.0 }, "eval_runtime": 1.4247, "eval_samples_per_second": 56.854, "eval_steps_per_second": 7.721, "eval_weighted avg": { "f1-score": 0.7655218131315448, "precision": 0.7723271066974134, "recall": 0.7682686567164179, "support": 33500.0 }, "step": 41 }, { "epoch": 2.0, "eval_Claim": { "f1-score": 0.5235859602056785, "precision": 0.5846230654018971, "recall": 0.47408906882591095, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.7206085753803596, "precision": 0.7269767441860465, "recall": 0.7143510054844607, "support": 2188.0 }, "eval_O": { "f1-score": 0.8884743765953269, "precision": 0.9142337609859582, "recall": 0.8641268022534135, "support": 10473.0 }, "eval_Premise": { "f1-score": 0.8745201535508637, "precision": 0.8357695614789338, "recall": 0.917038807472168, "support": 15899.0 }, "eval_accuracy": 0.8219402985074626, "eval_loss": 0.47505903244018555, "eval_macro avg": { "f1-score": 0.7517972664330571, "precision": 0.7654007830132088, "recall": 0.7424014210089883, "support": 33500.0 }, "eval_runtime": 1.4322, "eval_samples_per_second": 56.557, "eval_steps_per_second": 7.681, "eval_weighted avg": { "f1-score": 0.8170804260816812, "precision": 0.816159208839521, "recall": 0.8219402985074626, "support": 33500.0 }, "step": 82 }, { "epoch": 3.0, "eval_Claim": { "f1-score": 0.5033996474439688, "precision": 0.6658894070619586, "recall": 0.4046558704453441, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.7934708682838358, "precision": 0.7872244714349977, "recall": 0.79981718464351, "support": 2188.0 }, "eval_O": { "f1-score": 0.9017883608339096, "precision": 0.9342819121711536, "recall": 0.8714790413444095, "support": 10473.0 }, "eval_Premise": { "f1-score": 0.8787676209853219, "precision": 0.8168702042580784, "recall": 0.9508145166362665, "support": 15899.0 }, "eval_accuracy": 0.8356119402985075, "eval_loss": 0.458564430475235, "eval_macro avg": { "f1-score": 0.7693566243867591, "precision": 0.8010664987315472, "recall": 0.7566916532673825, "support": 33500.0 }, "eval_runtime": 1.4302, "eval_samples_per_second": 56.636, "eval_steps_per_second": 7.691, "eval_weighted avg": { "f1-score": 0.8250407291712659, "precision": 0.8293759599418965, "recall": 0.8356119402985075, "support": 33500.0 }, "step": 123 }, { "epoch": 4.0, "eval_Claim": { "f1-score": 0.6119839240043845, "precision": 0.5575898801597869, "recall": 0.6781376518218624, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.8013553578991952, "precision": 0.7466456195737964, "recall": 0.8647166361974405, "support": 2188.0 }, "eval_O": { "f1-score": 0.9009650063359004, "precision": 0.9201592832254853, "recall": 0.8825551417931825, "support": 10473.0 }, "eval_Premise": { "f1-score": 0.8636894716344281, "precision": 0.8922416683430564, "recall": 0.836907981634065, "support": 15899.0 }, "eval_accuracy": 0.8295820895522388, "eval_loss": 0.45249882340431213, "eval_macro avg": { "f1-score": 0.7944984399684771, "precision": 0.7791591128255312, "recall": 0.8155793528616375, "support": 33500.0 }, "eval_runtime": 1.4313, "eval_samples_per_second": 56.59, "eval_steps_per_second": 7.685, "eval_weighted avg": { "f1-score": 0.8341543739861718, "precision": 0.8421114352783158, "recall": 0.8295820895522388, "support": 33500.0 }, "step": 164 }, { "epoch": 5.0, "eval_Claim": { "f1-score": 0.6177297297297297, "precision": 0.662877030162413, "recall": 0.5783400809716599, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.8327169893408746, "precision": 0.7945205479452054, "recall": 0.8747714808043876, "support": 2188.0 }, "eval_O": { "f1-score": 0.9074411905904946, "precision": 0.9125229313507772, "recall": 0.9024157357013273, "support": 10473.0 }, "eval_Premise": { "f1-score": 0.8867988738669058, "precision": 0.8726254262055528, "recall": 0.9014403421598842, "support": 15899.0 }, "eval_accuracy": 0.8523582089552239, "eval_loss": 0.47209030389785767, "eval_macro avg": { "f1-score": 0.8111716958820011, "precision": 0.8106364839159872, "recall": 0.8142419099093148, "support": 33500.0 }, "eval_runtime": 1.4276, "eval_samples_per_second": 56.738, "eval_steps_per_second": 7.705, "eval_weighted avg": { "f1-score": 0.8500422842449816, "precision": 0.8490670984831405, "recall": 0.8523582089552239, "support": 33500.0 }, "step": 205 }, { "epoch": 6.0, "eval_Claim": { "f1-score": 0.6278612118073537, "precision": 0.6428419936373276, "recall": 0.6135627530364373, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.83248730964467, "precision": 0.804950917626974, "recall": 0.8619744058500914, "support": 2188.0 }, "eval_O": { "f1-score": 0.9114601059950406, "precision": 0.9285714285714286, "recall": 0.8949680129857729, "support": 10473.0 }, "eval_Premise": { "f1-score": 0.8842993146649301, "precision": 0.872155615365794, "recall": 0.8967859613812189, "support": 15899.0 }, "eval_accuracy": 0.8521791044776119, "eval_loss": 0.47916167974472046, "eval_macro avg": { "f1-score": 0.8140269855279986, "precision": 0.812129988800381, "recall": 0.8168227833133802, "support": 33500.0 }, "eval_runtime": 1.4308, "eval_samples_per_second": 56.613, "eval_steps_per_second": 7.688, "eval_weighted avg": { "f1-score": 0.8515914362320791, "precision": 0.8515881419840463, "recall": 0.8521791044776119, "support": 33500.0 }, "step": 246 }, { "epoch": 7.0, "eval_Claim": { "f1-score": 0.5961820851688694, "precision": 0.6744186046511628, "recall": 0.5342105263157895, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.8332962385933673, "precision": 0.8121475054229935, "recall": 0.8555758683729433, "support": 2188.0 }, "eval_O": { "f1-score": 0.9087219135056778, "precision": 0.9198786930150655, "recall": 0.8978325217225246, "support": 10473.0 }, "eval_Premise": { "f1-score": 0.8884371491853515, "precision": 0.8582063305978898, "recall": 0.9208755267626895, "support": 15899.0 }, "eval_accuracy": 0.8523880597014926, "eval_loss": 0.5201511383056641, "eval_macro avg": { "f1-score": 0.8066593466133165, "precision": 0.816162783421778, "recall": 0.8021236107934867, "support": 33500.0 }, "eval_runtime": 1.436, "eval_samples_per_second": 56.406, "eval_steps_per_second": 7.66, "eval_weighted avg": { "f1-score": 0.8480805524125185, "precision": 0.8473766761482054, "recall": 0.8523880597014926, "support": 33500.0 }, "step": 287 }, { "epoch": 8.0, "eval_Claim": { "f1-score": 0.6418997361477573, "precision": 0.6705622932745314, "recall": 0.6155870445344129, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.8422907488986784, "precision": 0.8129251700680272, "recall": 0.8738574040219378, "support": 2188.0 }, "eval_O": { "f1-score": 0.909046716251033, "precision": 0.9259259259259259, "recall": 0.89277188962093, "support": 10473.0 }, "eval_Premise": { "f1-score": 0.8894304929968533, "precision": 0.8728428701180745, "recall": 0.9066607962764954, "support": 15899.0 }, "eval_accuracy": 0.8572537313432835, "eval_loss": 0.5458493232727051, "eval_macro avg": { "f1-score": 0.8206669235735804, "precision": 0.8205640648466398, "recall": 0.822219283613444, "support": 33500.0 }, "eval_runtime": 1.4343, "eval_samples_per_second": 56.474, "eval_steps_per_second": 7.669, "eval_weighted avg": { "f1-score": 0.8559826424660975, "precision": 0.8556957914959558, "recall": 0.8572537313432835, "support": 33500.0 }, "step": 328 }, { "epoch": 9.0, "eval_Claim": { "f1-score": 0.6331998768093625, "precision": 0.6423661737138097, "recall": 0.6242914979757085, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.8381555153707052, "precision": 0.8291592128801432, "recall": 0.8473491773308958, "support": 2188.0 }, "eval_O": { "f1-score": 0.9061017111633034, "precision": 0.909720885466795, "recall": 0.9025112193258856, "support": 10473.0 }, "eval_Premise": { "f1-score": 0.8844614037282621, "precision": 0.8796739874323399, "recall": 0.8893012139128247, "support": 15899.0 }, "eval_accuracy": 0.8516119402985075, "eval_loss": 0.5549963116645813, "eval_macro avg": { "f1-score": 0.8154796267679083, "precision": 0.8152300648732719, "recall": 0.8158632771363286, "support": 33500.0 }, "eval_runtime": 1.4235, "eval_samples_per_second": 56.902, "eval_steps_per_second": 7.727, "eval_weighted avg": { "f1-score": 0.8511506488942767, "precision": 0.8507741138987609, "recall": 0.8516119402985075, "support": 33500.0 }, "step": 369 }, { "epoch": 10.0, "eval_Claim": { "f1-score": 0.6263982102908278, "precision": 0.6611198560827524, "recall": 0.5951417004048583, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.8409399005874378, "precision": 0.8315460232350312, "recall": 0.8505484460694699, "support": 2188.0 }, "eval_O": { "f1-score": 0.9098583349505143, "precision": 0.9248446592366111, "recall": 0.8953499474840065, "support": 10473.0 }, "eval_Premise": { "f1-score": 0.8851573292402148, "precision": 0.8645358599184456, "recall": 0.9067865903515945, "support": 15899.0 }, "eval_accuracy": 0.8535820895522388, "eval_loss": 0.5788276791572571, "eval_macro avg": { "f1-score": 0.8155884437672487, "precision": 0.8205115996182102, "recall": 0.8119566710774824, "support": 33500.0 }, "eval_runtime": 1.4232, "eval_samples_per_second": 56.912, "eval_steps_per_second": 7.729, "eval_weighted avg": { "f1-score": 0.8518342203238483, "precision": 0.851239060922849, "recall": 0.8535820895522388, "support": 33500.0 }, "step": 410 }, { "epoch": 11.0, "eval_Claim": { "f1-score": 0.6328578975171685, "precision": 0.661878453038674, "recall": 0.6062753036437247, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.8469798657718122, "precision": 0.829535495179667, "recall": 0.8651736745886655, "support": 2188.0 }, "eval_O": { "f1-score": 0.9110819097678493, "precision": 0.9291244788564622, "recall": 0.8937267258665139, "support": 10473.0 }, "eval_Premise": { "f1-score": 0.88966111076942, "precision": 0.8703893134364282, "recall": 0.9098056481539719, "support": 15899.0 }, "eval_accuracy": 0.8571044776119403, "eval_loss": 0.5865030288696289, "eval_macro avg": { "f1-score": 0.8201451959565625, "precision": 0.8227319351278078, "recall": 0.818745338063219, "support": 33500.0 }, "eval_runtime": 1.4299, "eval_samples_per_second": 56.648, "eval_steps_per_second": 7.693, "eval_weighted avg": { "f1-score": 0.8557012776467233, "precision": 0.8553356293389153, "recall": 0.8571044776119403, "support": 33500.0 }, "step": 451 }, { "epoch": 12.0, "eval_Claim": { "f1-score": 0.6447044940505456, "precision": 0.6268885064065787, "recall": 0.6635627530364372, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.8387389430709912, "precision": 0.8325078793336335, "recall": 0.8450639853747715, "support": 2188.0 }, "eval_O": { "f1-score": 0.9096724171351037, "precision": 0.923546196989078, "recall": 0.896209300105032, "support": 10473.0 }, "eval_Premise": { "f1-score": 0.885106650726735, "precision": 0.885440926543715, "recall": 0.8847726272092584, "support": 15899.0 }, "eval_accuracy": 0.8531343283582089, "eval_loss": 0.6139675378799438, "eval_macro avg": { "f1-score": 0.8195556262458439, "precision": 0.8170958773182513, "recall": 0.8224021664313748, "support": 33500.0 }, "eval_runtime": 1.4276, "eval_samples_per_second": 56.739, "eval_steps_per_second": 7.705, "eval_weighted avg": { "f1-score": 0.8543077872420695, "precision": 0.8557695842930038, "recall": 0.8531343283582089, "support": 33500.0 }, "step": 492 } ], "logging_steps": 500, "max_steps": 656, "num_input_tokens_seen": 0, "num_train_epochs": 16, "save_steps": 500, "total_flos": 1720106206408800.0, "train_batch_size": 8, "trial_name": null, "trial_params": null }