| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 21, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.14, | |
| "grad_norm": 1.3965061424468534, | |
| "learning_rate": 6.666666666666667e-06, | |
| "loss": 1.1352, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 1.68859600319114, | |
| "learning_rate": 1.9396926207859085e-05, | |
| "loss": 1.1023, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 1.3855385780334473, | |
| "eval_runtime": 1.3434, | |
| "eval_samples_per_second": 5.955, | |
| "eval_steps_per_second": 0.744, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "grad_norm": 0.9181756094384841, | |
| "learning_rate": 1.342020143325669e-05, | |
| "loss": 0.983, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 1.40538489818573, | |
| "eval_runtime": 1.3773, | |
| "eval_samples_per_second": 5.809, | |
| "eval_steps_per_second": 0.726, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 2.14, | |
| "grad_norm": 0.7917274913498488, | |
| "learning_rate": 5.000000000000003e-06, | |
| "loss": 0.8697, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 2.86, | |
| "grad_norm": 0.6206861453282275, | |
| "learning_rate": 1.519224698779198e-07, | |
| "loss": 0.7795, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 1.411104440689087, | |
| "eval_runtime": 1.3903, | |
| "eval_samples_per_second": 5.754, | |
| "eval_steps_per_second": 0.719, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 21, | |
| "total_flos": 4344627855360.0, | |
| "train_loss": 0.9259367443266369, | |
| "train_runtime": 345.6861, | |
| "train_samples_per_second": 3.775, | |
| "train_steps_per_second": 0.061 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 21, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 4344627855360.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |