| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 8.02139037433155, | |
| "eval_steps": 500, | |
| "global_step": 1500, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.748240053653717, | |
| "eval_loss": 0.4897768199443817, | |
| "eval_runtime": 9.3572, | |
| "eval_samples_per_second": 318.792, | |
| "eval_steps_per_second": 19.985, | |
| "step": 187 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7867918014526367, | |
| "eval_loss": 0.4351676404476166, | |
| "eval_runtime": 9.48, | |
| "eval_samples_per_second": 314.662, | |
| "eval_steps_per_second": 19.726, | |
| "step": 374 | |
| }, | |
| { | |
| "epoch": 2.6737967914438503, | |
| "grad_norm": 15.95652961730957, | |
| "learning_rate": 4.723376230943584e-06, | |
| "loss": 0.4992, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8079115152359009, | |
| "eval_loss": 0.413047730922699, | |
| "eval_runtime": 9.447, | |
| "eval_samples_per_second": 315.76, | |
| "eval_steps_per_second": 19.795, | |
| "step": 561 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.805229663848877, | |
| "eval_loss": 0.4279593527317047, | |
| "eval_runtime": 9.494, | |
| "eval_samples_per_second": 314.198, | |
| "eval_steps_per_second": 19.697, | |
| "step": 748 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.8109285831451416, | |
| "eval_loss": 0.46117427945137024, | |
| "eval_runtime": 9.4914, | |
| "eval_samples_per_second": 314.286, | |
| "eval_steps_per_second": 19.702, | |
| "step": 935 | |
| }, | |
| { | |
| "epoch": 5.347593582887701, | |
| "grad_norm": 18.722320556640625, | |
| "learning_rate": 2.9995162926430056e-06, | |
| "loss": 0.258, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.8129400014877319, | |
| "eval_loss": 0.5127686262130737, | |
| "eval_runtime": 9.3459, | |
| "eval_samples_per_second": 319.179, | |
| "eval_steps_per_second": 20.009, | |
| "step": 1122 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.8069058060646057, | |
| "eval_loss": 0.5462248921394348, | |
| "eval_runtime": 9.4811, | |
| "eval_samples_per_second": 314.625, | |
| "eval_steps_per_second": 19.723, | |
| "step": 1309 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.8112637996673584, | |
| "eval_loss": 0.5747013688087463, | |
| "eval_runtime": 9.4581, | |
| "eval_samples_per_second": 315.392, | |
| "eval_steps_per_second": 19.771, | |
| "step": 1496 | |
| }, | |
| { | |
| "epoch": 8.02139037433155, | |
| "grad_norm": 17.311113357543945, | |
| "learning_rate": 1.2756563543424279e-06, | |
| "loss": 0.1447, | |
| "step": 1500 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 1870, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 2.010272908557654e+16, | |
| "train_batch_size": 64, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 6.447236169244162e-06, | |
| "per_device_train_batch_size": 64 | |
| } | |
| } | |