{ "best_metric": null, "best_model_checkpoint": null, "epoch": 13.0, "eval_steps": 500, "global_step": 533, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 1.0, "eval_Claim": { "f1-score": 0.476290832455216, "precision": 0.564843099139128, "recall": 0.4117408906882591, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.6887323943661973, "precision": 0.5601374570446735, "recall": 0.8939670932358318, "support": 2188.0 }, "eval_O": { "f1-score": 0.9920864118816337, "precision": 0.9936889556724268, "recall": 0.9904890286826931, "support": 13353.0 }, "eval_Premise": { "f1-score": 0.8866859078930858, "precision": 0.8845214996557551, "recall": 0.888860934649978, "support": 15899.0 }, "eval_accuracy": 0.8616822429906542, "eval_loss": 0.3469788730144501, "eval_macro avg": { "f1-score": 0.7609488866490333, "precision": 0.7507977528779959, "recall": 0.7962644868141905, "support": 36380.0 }, "eval_runtime": 1.4365, "eval_samples_per_second": 56.386, "eval_steps_per_second": 7.657, "eval_weighted avg": { "f1-score": 0.8577397553229695, "precision": 0.8616723918054371, "recall": 0.8616822429906542, "support": 36380.0 }, "step": 41 }, { "epoch": 2.0, "eval_Claim": { "f1-score": 0.6444232602478551, "precision": 0.609009009009009, "recall": 0.6842105263157895, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.8019559902200488, "precision": 0.8622502628811777, "recall": 0.7495429616087751, "support": 2188.0 }, "eval_O": { "f1-score": 0.998463671450519, "precision": 0.9991750412479377, "recall": 0.9977533138620535, "support": 13353.0 }, "eval_Premise": { "f1-score": 0.9057885879401772, "precision": 0.9146466589713993, "recall": 0.8971004465689666, "support": 15899.0 }, "eval_accuracy": 0.8962616822429906, "eval_loss": 0.24824203550815582, "eval_macro avg": { "f1-score": 0.83765787746465, "precision": 0.846270243027381, "recall": 0.8321518120888962, "support": 36380.0 }, "eval_runtime": 1.4437, "eval_samples_per_second": 56.108, "eval_steps_per_second": 7.62, "eval_weighted avg": { "f1-score": 0.898068960328904, "precision": 0.901018681595891, "recall": 0.8962616822429906, "support": 36380.0 }, "step": 82 }, { "epoch": 3.0, "eval_Claim": { "f1-score": 0.5137480141757302, "precision": 0.6481652790625964, "recall": 0.42550607287449393, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.8317541613316262, "precision": 0.9458357600465929, "recall": 0.7422303473491774, "support": 2188.0 }, "eval_O": { "f1-score": 0.9982381827042021, "precision": 0.9993245271690183, "recall": 0.997154197558601, "support": 13353.0 }, "eval_Premise": { "f1-score": 0.8998382115016914, "precision": 0.8452144120247569, "recall": 0.962010189320083, "support": 15899.0 }, "eval_accuracy": 0.8888400219901045, "eval_loss": 0.2560241222381592, "eval_macro avg": { "f1-score": 0.8108946424283126, "precision": 0.8596349945757411, "recall": 0.7817252017755888, "support": 36380.0 }, "eval_runtime": 1.4504, "eval_samples_per_second": 55.846, "eval_steps_per_second": 7.584, "eval_weighted avg": { "f1-score": 0.8794336303830761, "precision": 0.8810739271473523, "recall": 0.8888400219901045, "support": 36380.0 }, "step": 123 }, { "epoch": 4.0, "eval_Claim": { "f1-score": 0.5716363636363636, "precision": 0.7123867069486405, "recall": 0.4773279352226721, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.8958780081010246, "precision": 0.9357889497262319, "recall": 0.8592321755027422, "support": 2188.0 }, "eval_O": { "f1-score": 0.9977489307421025, "precision": 0.999699270731524, "recall": 0.9958061858758331, "support": 13353.0 }, "eval_Premise": { "f1-score": 0.9059092664666211, "precision": 0.8584459459459459, "recall": 0.958928234480156, "support": 15899.0 }, "eval_accuracy": 0.9010720175920836, "eval_loss": 0.24002882838249207, "eval_macro avg": { "f1-score": 0.8427931422365279, "precision": 0.8765802183380855, "recall": 0.8228236327703509, "support": 36380.0 }, "eval_runtime": 1.4479, "eval_samples_per_second": 55.942, "eval_steps_per_second": 7.597, "eval_weighted avg": { "f1-score": 0.8936244534865525, "precision": 0.8951103081638239, "recall": 0.9010720175920836, "support": 36380.0 }, "step": 164 }, { "epoch": 5.0, "eval_Claim": { "f1-score": 0.6379673804504604, "precision": 0.7058679106309845, "recall": 0.5819838056680162, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.8943466172381834, "precision": 0.9069548872180451, "recall": 0.8820840950639853, "support": 2188.0 }, "eval_O": { "f1-score": 0.9964329966582812, "precision": 0.999171686746988, "recall": 0.9937092788137497, "support": 13353.0 }, "eval_Premise": { "f1-score": 0.9112140984206354, "precision": 0.8842535061246227, "recall": 0.939870432102648, "support": 15899.0 }, "eval_accuracy": 0.9075590984057175, "eval_loss": 0.2713705599308014, "eval_macro avg": { "f1-score": 0.8599902731918901, "precision": 0.8740619976801601, "recall": 0.8494119029120998, "support": 36380.0 }, "eval_runtime": 1.4528, "eval_samples_per_second": 55.756, "eval_steps_per_second": 7.572, "eval_weighted avg": { "f1-score": 0.9043747117402454, "precision": 0.9035758878163292, "recall": 0.9075590984057175, "support": 36380.0 }, "step": 205 }, { "epoch": 6.0, "eval_Claim": { "f1-score": 0.6678616617460986, "precision": 0.6969630281690141, "recall": 0.6410931174089068, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.8903225806451615, "precision": 0.8673602080624188, "recall": 0.9145338208409507, "support": 2188.0 }, "eval_O": { "f1-score": 0.9950327387672161, "precision": 1.0, "recall": 0.9901145809930353, "support": 13353.0 }, "eval_Premise": { "f1-score": 0.9114167727512653, "precision": 0.8999877360804514, "recall": 0.9231398201144726, "support": 15899.0 }, "eval_accuracy": 0.9089059923034635, "eval_loss": 0.265927255153656, "eval_macro avg": { "f1-score": 0.8661584384774353, "precision": 0.8660777430779711, "recall": 0.8672203348393414, "support": 36380.0 }, "eval_runtime": 1.4493, "eval_samples_per_second": 55.89, "eval_steps_per_second": 7.59, "eval_weighted avg": { "f1-score": 0.9077666257891245, "precision": 0.9071656544897909, "recall": 0.9089059923034635, "support": 36380.0 }, "step": 246 }, { "epoch": 7.0, "eval_Claim": { "f1-score": 0.5876388218252052, "precision": 0.7278708133971292, "recall": 0.49271255060728747, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.888888888888889, "precision": 0.8874715261958998, "recall": 0.8903107861060329, "support": 2188.0 }, "eval_O": { "f1-score": 0.9968070320423725, "precision": 1.0, "recall": 0.9936343892758182, "support": 13353.0 }, "eval_Premise": { "f1-score": 0.9057122370936902, "precision": 0.8625732658054971, "recall": 0.9533932951757972, "support": 15899.0 }, "eval_accuracy": 0.9018141836173722, "eval_loss": 0.3055996596813202, "eval_macro avg": { "f1-score": 0.8447617449625392, "precision": 0.8694789013496315, "recall": 0.832512755291234, "support": 36380.0 }, "eval_runtime": 1.4427, "eval_samples_per_second": 56.144, "eval_steps_per_second": 7.624, "eval_weighted avg": { "f1-score": 0.8949452398328692, "precision": 0.8962210519664664, "recall": 0.9018141836173722, "support": 36380.0 }, "step": 287 }, { "epoch": 8.0, "eval_Claim": { "f1-score": 0.6830078936435396, "precision": 0.7013651877133106, "recall": 0.665587044534413, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.9099371654642774, "precision": 0.9269796111901375, "recall": 0.893510054844607, "support": 2188.0 }, "eval_O": { "f1-score": 0.9968451889130925, "precision": 0.9998493181646952, "recall": 0.9938590578896128, "support": 13353.0 }, "eval_Premise": { "f1-score": 0.9144649011145954, "precision": 0.9029429797670141, "recall": 0.9262846719919492, "support": 15899.0 }, "eval_accuracy": 0.9137163276525564, "eval_loss": 0.31213706731796265, "eval_macro avg": { "f1-score": 0.8760637872838762, "precision": 0.8827842742087894, "recall": 0.8698102073151455, "support": 36380.0 }, "eval_runtime": 1.4431, "eval_samples_per_second": 56.13, "eval_steps_per_second": 7.623, "eval_weighted avg": { "f1-score": 0.9130003513747224, "precision": 0.9125853160350936, "recall": 0.9137163276525564, "support": 36380.0 }, "step": 328 }, { "epoch": 9.0, "eval_Claim": { "f1-score": 0.6714174150722373, "precision": 0.6483785822021116, "recall": 0.6961538461538461, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.8994732221246707, "precision": 0.8652871621621622, "recall": 0.9364716636197441, "support": 2188.0 }, "eval_O": { "f1-score": 0.9986506746626687, "precision": 0.9996248217903504, "recall": 0.9976784243241219, "support": 13353.0 }, "eval_Premise": { "f1-score": 0.9039641943734017, "precision": 0.9191860087120474, "recall": 0.8892383168752752, "support": 15899.0 }, "eval_accuracy": 0.9056624518966465, "eval_loss": 0.3160102665424347, "eval_macro avg": { "f1-score": 0.8683763765582445, "precision": 0.8581191437166679, "recall": 0.8798855627432468, "support": 36380.0 }, "eval_runtime": 1.4435, "eval_samples_per_second": 56.114, "eval_steps_per_second": 7.62, "eval_weighted avg": { "f1-score": 0.9068707703567609, "precision": 0.9086961820991927, "recall": 0.9056624518966465, "support": 36380.0 }, "step": 369 }, { "epoch": 10.0, "eval_Claim": { "f1-score": 0.6854596331226034, "precision": 0.7022722446379274, "recall": 0.6694331983805668, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.912037037037037, "precision": 0.924015009380863, "recall": 0.9003656307129799, "support": 2188.0 }, "eval_O": { "f1-score": 0.9982003599280145, "precision": 0.999474435017644, "recall": 0.9969295289448065, "support": 13353.0 }, "eval_Premise": { "f1-score": 0.9142252249447368, "precision": 0.9051787916152898, "recall": 0.9234543053022203, "support": 15899.0 }, "eval_accuracy": 0.9145409565695437, "eval_loss": 0.345079630613327, "eval_macro avg": { "f1-score": 0.8774805637580978, "precision": 0.882735120162931, "recall": 0.8725456658351434, "support": 36380.0 }, "eval_runtime": 1.4444, "eval_samples_per_second": 56.08, "eval_steps_per_second": 7.616, "eval_weighted avg": { "f1-score": 0.9138522232594241, "precision": 0.9133696939999666, "recall": 0.9145409565695437, "support": 36380.0 }, "step": 410 }, { "epoch": 11.0, "eval_Claim": { "f1-score": 0.6640950792326938, "precision": 0.6846517626827171, "recall": 0.6447368421052632, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.9124319419237751, "precision": 0.9058558558558558, "recall": 0.9191042047531993, "support": 2188.0 }, "eval_O": { "f1-score": 0.9980882408066872, "precision": 0.9991744220954668, "recall": 0.9970044184827379, "support": 13353.0 }, "eval_Premise": { "f1-score": 0.9098899728828351, "precision": 0.9018783984181908, "recall": 0.9180451600729606, "support": 15899.0 }, "eval_accuracy": 0.909978009895547, "eval_loss": 0.3666386902332306, "eval_macro avg": { "f1-score": 0.8711263087114978, "precision": 0.8728901097630577, "recall": 0.8697226563535402, "support": 36380.0 }, "eval_runtime": 1.4494, "eval_samples_per_second": 55.887, "eval_steps_per_second": 7.59, "eval_weighted avg": { "f1-score": 0.9090391352032605, "precision": 0.9083324088773178, "recall": 0.909978009895547, "support": 36380.0 }, "step": 451 }, { "epoch": 12.0, "eval_Claim": { "f1-score": 0.6594242259641498, "precision": 0.7116060961313013, "recall": 0.6143724696356275, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.897596882442087, "precision": 0.8527354997943233, "recall": 0.9474405850091407, "support": 2188.0 }, "eval_O": { "f1-score": 0.9980512666766601, "precision": 0.9988748030905409, "recall": 0.9972290870965326, "support": 13353.0 }, "eval_Premise": { "f1-score": 0.9140518417462483, "precision": 0.9013636641594814, "recall": 0.9271023334800931, "support": 15899.0 }, "eval_accuracy": 0.9115997800989555, "eval_loss": 0.4192342460155487, "eval_macro avg": { "f1-score": 0.8672810542072864, "precision": 0.8661450157939118, "recall": 0.8715361188053484, "support": 36380.0 }, "eval_runtime": 1.4505, "eval_samples_per_second": 55.842, "eval_steps_per_second": 7.583, "eval_weighted avg": { "f1-score": 0.9093179343293907, "precision": 0.9084627688449202, "recall": 0.9115997800989555, "support": 36380.0 }, "step": 492 }, { "epoch": 12.2, "learning_rate": 4.75609756097561e-06, "loss": 0.1714, "step": 500 }, { "epoch": 13.0, "eval_Claim": { "f1-score": 0.6607543540976599, "precision": 0.6997058157954288, "recall": 0.625910931174089, "support": 4940.0 }, "eval_MajorClaim": { "f1-score": 0.9072817729534148, "precision": 0.8979409131602507, "recall": 0.916819012797075, "support": 2188.0 }, "eval_O": { "f1-score": 0.9972236812485931, "precision": 0.9991729945116908, "recall": 0.9952819591103123, "support": 13353.0 }, "eval_Premise": { "f1-score": 0.9105645784996134, "precision": 0.8959576281504931, "recall": 0.9256557016164538, "support": 15899.0 }, "eval_accuracy": 0.909978009895547, "eval_loss": 0.431348592042923, "eval_macro avg": { "f1-score": 0.8689560966998203, "precision": 0.8731943379044659, "recall": 0.8659169011744825, "support": 36380.0 }, "eval_runtime": 1.4449, "eval_samples_per_second": 56.059, "eval_steps_per_second": 7.613, "eval_weighted avg": { "f1-score": 0.9082532456773593, "precision": 0.9073125006515488, "recall": 0.909978009895547, "support": 36380.0 }, "step": 533 } ], "logging_steps": 500, "max_steps": 656, "num_input_tokens_seen": 0, "num_train_epochs": 16, "save_steps": 500, "total_flos": 1863448390276200.0, "train_batch_size": 8, "trial_name": null, "trial_params": null }