{ "best_metric": null, "best_model_checkpoint": null, "epoch": 0.001300326707085155, "eval_steps": 2, "global_step": 5, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.000260065341417031, "grad_norm": NaN, "learning_rate": 1e-05, "loss": 2415.3406, "step": 1 }, { "epoch": 0.000260065341417031, "eval_loss": NaN, "eval_runtime": 515.7593, "eval_samples_per_second": 12.558, "eval_steps_per_second": 6.28, "step": 1 }, { "epoch": 0.000520130682834062, "grad_norm": NaN, "learning_rate": 2e-05, "loss": 80.846, "step": 2 }, { "epoch": 0.000520130682834062, "eval_loss": NaN, "eval_runtime": 327.7405, "eval_samples_per_second": 19.763, "eval_steps_per_second": 9.883, "step": 2 }, { "epoch": 0.0007801960242510931, "grad_norm": NaN, "learning_rate": 3e-05, "loss": 595.1917, "step": 3 }, { "epoch": 0.001040261365668124, "grad_norm": NaN, "learning_rate": 4e-05, "loss": 1650.6531, "step": 4 }, { "epoch": 0.001040261365668124, "eval_loss": NaN, "eval_runtime": 331.5634, "eval_samples_per_second": 19.535, "eval_steps_per_second": 9.769, "step": 4 }, { "epoch": 0.001300326707085155, "grad_norm": NaN, "learning_rate": 5e-05, "loss": 75.1896, "step": 5 } ], "logging_steps": 1, "max_steps": 5, "num_input_tokens_seen": 0, "num_train_epochs": 1, "save_steps": 2, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 8250556258713600.0, "train_batch_size": 2, "trial_name": null, "trial_params": null }