{ "best_metric": null, "best_model_checkpoint": null, "epoch": 3.0, "eval_steps": 1115, "global_step": 2229, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 1.0, "eval_loss": 0.25535184144973755, "eval_runtime": 14.8054, "eval_samples_per_second": 32.421, "eval_steps_per_second": 1.013, "step": 743 }, { "epoch": 1.3458950201884252, "grad_norm": 0.663228788046308, "learning_rate": 3.3664863364891915e-06, "loss": 0.2863, "step": 1000 }, { "epoch": 2.0, "eval_loss": 0.2506025433540344, "eval_runtime": 14.2808, "eval_samples_per_second": 33.611, "eval_steps_per_second": 1.05, "step": 1486 }, { "epoch": 2.6917900403768504, "grad_norm": 0.6701798076996605, "learning_rate": 1.5905925700480885e-07, "loss": 0.1776, "step": 2000 }, { "epoch": 3.0, "eval_loss": 0.26738250255584717, "eval_runtime": 14.1687, "eval_samples_per_second": 33.877, "eval_steps_per_second": 1.059, "step": 2229 }, { "epoch": 3.0, "step": 2229, "tflops": 0.10843344625184329, "token/s": 550.5949428046324, "total_flos": 1253049251135488.0, "train_loss": 0.22336662569726298, "train_runtime": 11593.2658, "train_samples_per_second": 12.297, "train_steps_per_second": 0.192 } ], "log_save_evaluate_time": 1083.2308747768402, "logging_steps": 1000, "max_steps": 2229, "num_input_tokens_seen": 0, "num_train_epochs": 3, "save_steps": 1, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 1253049251135488.0, "total_tokens": 5786784.0, "train_batch_size": 2, "trial_name": null, "trial_params": null }