{ "best_metric": null, "best_model_checkpoint": null, "epoch": 0.0034190956492007865, "eval_steps": 2, "global_step": 5, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.0006838191298401573, "grad_norm": NaN, "learning_rate": 2e-05, "loss": 14.3757, "step": 1 }, { "epoch": 0.0006838191298401573, "eval_loss": NaN, "eval_runtime": 89.326, "eval_samples_per_second": 6.896, "eval_steps_per_second": 6.896, "step": 1 }, { "epoch": 0.0013676382596803146, "grad_norm": NaN, "learning_rate": 4e-05, "loss": 7.9118, "step": 2 }, { "epoch": 0.0013676382596803146, "eval_loss": NaN, "eval_runtime": 88.7701, "eval_samples_per_second": 6.939, "eval_steps_per_second": 6.939, "step": 2 }, { "epoch": 0.002051457389520472, "grad_norm": NaN, "learning_rate": 6e-05, "loss": 5.5757, "step": 3 }, { "epoch": 0.002735276519360629, "grad_norm": NaN, "learning_rate": 8e-05, "loss": 6.9588, "step": 4 }, { "epoch": 0.002735276519360629, "eval_loss": NaN, "eval_runtime": 88.9657, "eval_samples_per_second": 6.924, "eval_steps_per_second": 6.924, "step": 4 }, { "epoch": 0.0034190956492007865, "grad_norm": NaN, "learning_rate": 0.0001, "loss": 0.2667, "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": 538704573628416.0, "train_batch_size": 1, "trial_name": null, "trial_params": null }