| { | |
| "best_metric": 0.2674808204174042, | |
| "best_model_checkpoint": "./results/checkpoint-2416", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 7248, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.20695364238410596, | |
| "grad_norm": 0.06496760994195938, | |
| "learning_rate": 1.862030905077263e-05, | |
| "loss": 0.1879, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4139072847682119, | |
| "grad_norm": 0.2046840637922287, | |
| "learning_rate": 1.7240618101545256e-05, | |
| "loss": 0.1544, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.6208609271523179, | |
| "grad_norm": 78.30413818359375, | |
| "learning_rate": 1.5860927152317882e-05, | |
| "loss": 0.1443, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.8278145695364238, | |
| "grad_norm": 13.444316864013672, | |
| "learning_rate": 1.448123620309051e-05, | |
| "loss": 0.1214, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 0.2674808204174042, | |
| "eval_runtime": 68.0862, | |
| "eval_samples_per_second": 31.548, | |
| "eval_steps_per_second": 3.951, | |
| "step": 2416 | |
| }, | |
| { | |
| "epoch": 1.0347682119205297, | |
| "grad_norm": 11.859747886657715, | |
| "learning_rate": 1.3101545253863135e-05, | |
| "loss": 0.1299, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.2417218543046358, | |
| "grad_norm": 7.801605701446533, | |
| "learning_rate": 1.1721854304635763e-05, | |
| "loss": 0.1358, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.4486754966887418, | |
| "grad_norm": 9.417176246643066, | |
| "learning_rate": 1.034216335540839e-05, | |
| "loss": 0.1331, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.6556291390728477, | |
| "grad_norm": 0.017003627493977547, | |
| "learning_rate": 8.962472406181017e-06, | |
| "loss": 0.1325, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.8625827814569536, | |
| "grad_norm": 7.609455585479736, | |
| "learning_rate": 7.582781456953643e-06, | |
| "loss": 0.1273, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.2778768837451935, | |
| "eval_runtime": 67.8572, | |
| "eval_samples_per_second": 31.655, | |
| "eval_steps_per_second": 3.964, | |
| "step": 4832 | |
| }, | |
| { | |
| "epoch": 2.0695364238410594, | |
| "grad_norm": 0.0077050491236150265, | |
| "learning_rate": 6.203090507726269e-06, | |
| "loss": 0.125, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.2764900662251657, | |
| "grad_norm": 9.524530410766602, | |
| "learning_rate": 4.823399558498897e-06, | |
| "loss": 0.0672, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.4834437086092715, | |
| "grad_norm": 3.2392263412475586, | |
| "learning_rate": 3.443708609271523e-06, | |
| "loss": 0.0731, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.6903973509933774, | |
| "grad_norm": 0.4321236312389374, | |
| "learning_rate": 2.06401766004415e-06, | |
| "loss": 0.0818, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 2.8973509933774837, | |
| "grad_norm": 0.006676756776869297, | |
| "learning_rate": 6.843267108167771e-07, | |
| "loss": 0.0691, | |
| "step": 7000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 0.30005592107772827, | |
| "eval_runtime": 67.9967, | |
| "eval_samples_per_second": 31.59, | |
| "eval_steps_per_second": 3.956, | |
| "step": 7248 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 7248, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 1.52529694887168e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |