{ "best_metric": null, "best_model_checkpoint": null, "epoch": 20.0, "global_step": 580, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 1.0, "learning_rate": 3.6559821951932435e-06, "loss": 2.0662, "step": 29 }, { "epoch": 2.0, "learning_rate": 5.463687482021196e-06, "loss": 1.413, "step": 58 }, { "epoch": 3.0, "learning_rate": 6.2507068274056945e-06, "loss": 1.1678, "step": 87 }, { "epoch": 4.0, "learning_rate": 6.760687110038668e-06, "loss": 1.0382, "step": 116 }, { "epoch": 5.0, "learning_rate": 7.13867309942701e-06, "loss": 0.954, "step": 145 }, { "epoch": 6.0, "learning_rate": 7.439130415804967e-06, "loss": 0.8951, "step": 174 }, { "epoch": 7.0, "learning_rate": 7.688512885309291e-06, "loss": 0.8487, "step": 203 }, { "epoch": 8.0, "learning_rate": 7.901687980648141e-06, "loss": 0.8119, "step": 232 }, { "epoch": 9.0, "learning_rate": 8.087848755754509e-06, "loss": 0.7811, "step": 261 }, { "epoch": 10.0, "learning_rate": 8.253078243154342e-06, "loss": 0.7546, "step": 290 }, { "epoch": 11.0, "learning_rate": 8.401611184973133e-06, "loss": 0.7308, "step": 319 }, { "epoch": 12.0, "learning_rate": 8.53651443066959e-06, "loss": 0.7102, "step": 348 }, { "epoch": 13.0, "learning_rate": 8.66008046615066e-06, "loss": 0.6909, "step": 377 }, { "epoch": 14.0, "learning_rate": 8.774067992353493e-06, "loss": 0.6728, "step": 406 }, { "epoch": 15.0, "learning_rate": 8.879855771890845e-06, "loss": 0.6564, "step": 435 }, { "epoch": 16.0, "learning_rate": 8.9785447549646e-06, "loss": 0.64, "step": 464 }, { "epoch": 17.0, "learning_rate": 9.071028041604564e-06, "loss": 0.6247, "step": 493 }, { "epoch": 18.0, "learning_rate": 9.15804011885607e-06, "loss": 0.611, "step": 522 }, { "epoch": 19.0, "learning_rate": 9.240192329946528e-06, "loss": 0.5964, "step": 551 }, { "epoch": 20.0, "learning_rate": 9.317998951644404e-06, "loss": 0.5837, "step": 580 }, { "epoch": 20.0, "step": 580, "total_flos": 1134007009935360.0, "train_loss": 0.8623725891113281, "train_runtime": 17895.3422, "train_samples_per_second": 8.201, "train_steps_per_second": 0.032 } ], "max_steps": 580, "num_train_epochs": 20, "total_flos": 1134007009935360.0, "trial_name": null, "trial_params": null }