| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 4.976525821596244, | |
| "eval_steps": 500, | |
| "global_step": 265, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.47, | |
| "grad_norm": 6.918335914611816, | |
| "learning_rate": 2.5555555555555557e-05, | |
| "loss": 2.6706, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "grad_norm": 1.054337501525879, | |
| "learning_rate": 2.735294117647059e-05, | |
| "loss": 2.1693, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.41, | |
| "grad_norm": 1.9611724615097046, | |
| "learning_rate": 2.420168067226891e-05, | |
| "loss": 1.8135, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "grad_norm": 0.7985315322875977, | |
| "learning_rate": 2.1050420168067226e-05, | |
| "loss": 1.6577, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.35, | |
| "grad_norm": 0.72879958152771, | |
| "learning_rate": 1.7899159663865546e-05, | |
| "loss": 1.5306, | |
| "step": 125 | |
| }, | |
| { | |
| "epoch": 2.82, | |
| "grad_norm": 0.7586284875869751, | |
| "learning_rate": 1.4747899159663864e-05, | |
| "loss": 1.5247, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 3.29, | |
| "grad_norm": 0.8593073487281799, | |
| "learning_rate": 1.1596638655462186e-05, | |
| "loss": 1.469, | |
| "step": 175 | |
| }, | |
| { | |
| "epoch": 3.76, | |
| "grad_norm": 0.8334183096885681, | |
| "learning_rate": 8.445378151260505e-06, | |
| "loss": 1.454, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 4.23, | |
| "grad_norm": 0.8410723805427551, | |
| "learning_rate": 5.294117647058824e-06, | |
| "loss": 1.4421, | |
| "step": 225 | |
| }, | |
| { | |
| "epoch": 4.69, | |
| "grad_norm": 0.8686198592185974, | |
| "learning_rate": 2.1428571428571427e-06, | |
| "loss": 1.4399, | |
| "step": 250 | |
| } | |
| ], | |
| "logging_steps": 25, | |
| "max_steps": 265, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 1.303502332624896e+16, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |