{ "best_metric": 0.5472163817533595, "best_model_checkpoint": "/tmp/mecha-tagger/checkpoint-1573", "epoch": 13.0, "global_step": 1573, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.83, "learning_rate": 2e-05, "loss": 0.437, "step": 100 }, { "epoch": 1.0, "eval_accuracy": 0.06247574699262708, "eval_f1": 0.431281999799217, "eval_loss": 0.32573798298835754, "eval_roc_auc": 0.6429553592437186, "eval_runtime": 50.2381, "eval_samples_per_second": 51.296, "eval_steps_per_second": 0.537, "step": 121 }, { "epoch": 1.65, "learning_rate": 2e-05, "loss": 0.3211, "step": 200 }, { "epoch": 2.0, "eval_accuracy": 0.055490880869227786, "eval_f1": 0.45432441820076414, "eval_loss": 0.3099991977214813, "eval_roc_auc": 0.6526660416613327, "eval_runtime": 50.3523, "eval_samples_per_second": 51.179, "eval_steps_per_second": 0.536, "step": 242 }, { "epoch": 2.48, "learning_rate": 2e-05, "loss": 0.309, "step": 300 }, { "epoch": 3.0, "eval_accuracy": 0.05820721769499418, "eval_f1": 0.4483176720218248, "eval_loss": 0.30198442935943604, "eval_roc_auc": 0.6485457459740307, "eval_runtime": 50.3568, "eval_samples_per_second": 51.175, "eval_steps_per_second": 0.536, "step": 363 }, { "epoch": 3.31, "learning_rate": 2e-05, "loss": 0.303, "step": 400 }, { "epoch": 4.0, "eval_accuracy": 0.05743112145906092, "eval_f1": 0.4746113989637306, "eval_loss": 0.295987069606781, "eval_roc_auc": 0.6608536486091969, "eval_runtime": 50.3091, "eval_samples_per_second": 51.223, "eval_steps_per_second": 0.537, "step": 484 }, { "epoch": 4.13, "learning_rate": 2e-05, "loss": 0.2965, "step": 500 }, { "epoch": 4.96, "learning_rate": 2e-05, "loss": 0.294, "step": 600 }, { "epoch": 5.0, "eval_accuracy": 0.0682964687621265, "eval_f1": 0.4646505308710311, "eval_loss": 0.28789642453193665, "eval_roc_auc": 0.6549506720523524, "eval_runtime": 50.2808, "eval_samples_per_second": 51.252, "eval_steps_per_second": 0.537, "step": 605 }, { "epoch": 5.79, "learning_rate": 2e-05, "loss": 0.2879, "step": 700 }, { "epoch": 6.0, "eval_accuracy": 0.06674427629025999, "eval_f1": 0.4727673207849419, "eval_loss": 0.2825814187526703, "eval_roc_auc": 0.658515066442945, "eval_runtime": 50.324, "eval_samples_per_second": 51.208, "eval_steps_per_second": 0.537, "step": 726 }, { "epoch": 6.61, "learning_rate": 2e-05, "loss": 0.2846, "step": 800 }, { "epoch": 7.0, "eval_accuracy": 0.06984866123399301, "eval_f1": 0.48976385684127843, "eval_loss": 0.27861288189888, "eval_roc_auc": 0.6671462002192768, "eval_runtime": 50.1175, "eval_samples_per_second": 51.419, "eval_steps_per_second": 0.539, "step": 847 }, { "epoch": 7.44, "learning_rate": 2e-05, "loss": 0.2781, "step": 900 }, { "epoch": 8.0, "eval_accuracy": 0.06713232440822663, "eval_f1": 0.4918627450980392, "eval_loss": 0.27516159415245056, "eval_roc_auc": 0.6681229472140713, "eval_runtime": 50.4646, "eval_samples_per_second": 51.066, "eval_steps_per_second": 0.535, "step": 968 }, { "epoch": 8.26, "learning_rate": 2e-05, "loss": 0.2626, "step": 1000 }, { "epoch": 9.0, "eval_accuracy": 0.06635622817229336, "eval_f1": 0.5017455391776571, "eval_loss": 0.26821279525756836, "eval_roc_auc": 0.6732412875217908, "eval_runtime": 51.4118, "eval_samples_per_second": 50.125, "eval_steps_per_second": 0.525, "step": 1089 }, { "epoch": 9.09, "learning_rate": 2e-05, "loss": 0.2625, "step": 1100 }, { "epoch": 9.92, "learning_rate": 2e-05, "loss": 0.2528, "step": 1200 }, { "epoch": 10.0, "eval_accuracy": 0.06674427629025999, "eval_f1": 0.5120230381569474, "eval_loss": 0.26574626564979553, "eval_roc_auc": 0.6785095938529904, "eval_runtime": 51.4008, "eval_samples_per_second": 50.135, "eval_steps_per_second": 0.525, "step": 1210 }, { "epoch": 10.74, "learning_rate": 2e-05, "loss": 0.2451, "step": 1300 }, { "epoch": 11.0, "eval_accuracy": 0.0682964687621265, "eval_f1": 0.5220403321470937, "eval_loss": 0.2608773112297058, "eval_roc_auc": 0.6839487133042365, "eval_runtime": 51.3505, "eval_samples_per_second": 50.184, "eval_steps_per_second": 0.526, "step": 1331 }, { "epoch": 11.57, "learning_rate": 2e-05, "loss": 0.2362, "step": 1400 }, { "epoch": 12.0, "eval_accuracy": 0.07140085370585952, "eval_f1": 0.5350377904264253, "eval_loss": 0.2600255012512207, "eval_roc_auc": 0.6914645249590691, "eval_runtime": 51.3565, "eval_samples_per_second": 50.179, "eval_steps_per_second": 0.526, "step": 1452 }, { "epoch": 12.4, "learning_rate": 2e-05, "loss": 0.2281, "step": 1500 }, { "epoch": 13.0, "eval_accuracy": 0.07023670935195964, "eval_f1": 0.5472163817533595, "eval_loss": 0.2586032450199127, "eval_roc_auc": 0.6994249977444085, "eval_runtime": 51.2515, "eval_samples_per_second": 50.281, "eval_steps_per_second": 0.527, "step": 1573 } ], "max_steps": 6050, "num_train_epochs": 50, "total_flos": 6.882242130850185e+19, "trial_name": null, "trial_params": null }