| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.005442547112048439, | |
| "eval_steps": 2, | |
| "global_step": 5, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0010885094224096878, | |
| "grad_norm": 4.611927032470703, | |
| "learning_rate": 1e-05, | |
| "loss": 9.2995, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0010885094224096878, | |
| "eval_loss": 8.37597942352295, | |
| "eval_runtime": 79.8413, | |
| "eval_samples_per_second": 19.388, | |
| "eval_steps_per_second": 9.694, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0021770188448193757, | |
| "grad_norm": 5.37861967086792, | |
| "learning_rate": 2e-05, | |
| "loss": 7.6651, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.0021770188448193757, | |
| "eval_loss": 8.245399475097656, | |
| "eval_runtime": 17.5091, | |
| "eval_samples_per_second": 88.411, | |
| "eval_steps_per_second": 44.206, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.0032655282672290633, | |
| "grad_norm": 4.811445713043213, | |
| "learning_rate": 3e-05, | |
| "loss": 8.629, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.004354037689638751, | |
| "grad_norm": 4.4482831954956055, | |
| "learning_rate": 4e-05, | |
| "loss": 8.2228, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.004354037689638751, | |
| "eval_loss": 8.056740760803223, | |
| "eval_runtime": 17.3084, | |
| "eval_samples_per_second": 89.436, | |
| "eval_steps_per_second": 44.718, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.005442547112048439, | |
| "grad_norm": 5.2335686683654785, | |
| "learning_rate": 5e-05, | |
| "loss": 7.7954, | |
| "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": 933006126612480.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |