| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 6.0, | |
| "eval_steps": 20, | |
| "global_step": 108, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 0.0002, | |
| "loss": 1.2637, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "learning_rate": 0.0002, | |
| "loss": 1.1369, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "eval_loss": 1.1003490686416626, | |
| "eval_runtime": 4.4867, | |
| "eval_samples_per_second": 2.675, | |
| "eval_steps_per_second": 0.446, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "learning_rate": 0.0002, | |
| "loss": 1.0287, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "learning_rate": 0.0002, | |
| "loss": 0.944, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 2.22, | |
| "eval_loss": 1.0134328603744507, | |
| "eval_runtime": 4.507, | |
| "eval_samples_per_second": 2.663, | |
| "eval_steps_per_second": 0.444, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 2.78, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8812, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 3.33, | |
| "learning_rate": 0.0002, | |
| "loss": 0.8117, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 3.33, | |
| "eval_loss": 0.9895493388175964, | |
| "eval_runtime": 4.5202, | |
| "eval_samples_per_second": 2.655, | |
| "eval_steps_per_second": 0.442, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 3.89, | |
| "learning_rate": 0.0002, | |
| "loss": 0.7667, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 4.44, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6833, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 4.44, | |
| "eval_loss": 0.9626936316490173, | |
| "eval_runtime": 4.5252, | |
| "eval_samples_per_second": 2.652, | |
| "eval_steps_per_second": 0.442, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "learning_rate": 0.0002, | |
| "loss": 0.6519, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 5.56, | |
| "learning_rate": 0.0002, | |
| "loss": 0.5459, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 5.56, | |
| "eval_loss": 1.000801682472229, | |
| "eval_runtime": 4.5183, | |
| "eval_samples_per_second": 2.656, | |
| "eval_steps_per_second": 0.443, | |
| "step": 100 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 144, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 8, | |
| "save_steps": 500, | |
| "total_flos": 3.736480024992154e+16, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |