| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 126, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.24, | |
| "grad_norm": 0.58479905128479, | |
| "learning_rate": 8.000000000000001e-07, | |
| "loss": 0.3585, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "grad_norm": 0.508639395236969, | |
| "learning_rate": 1.6000000000000001e-06, | |
| "loss": 0.3844, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "grad_norm": 0.4961947202682495, | |
| "learning_rate": 2.4000000000000003e-06, | |
| "loss": 0.3441, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "grad_norm": 0.5687706470489502, | |
| "learning_rate": 3.2000000000000003e-06, | |
| "loss": 0.3524, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "grad_norm": 0.41114819049835205, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 0.3529, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.43, | |
| "grad_norm": 0.49048012495040894, | |
| "learning_rate": 4.800000000000001e-06, | |
| "loss": 0.374, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.67, | |
| "grad_norm": 0.39151516556739807, | |
| "learning_rate": 5.600000000000001e-06, | |
| "loss": 0.3365, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 1.9, | |
| "grad_norm": 0.443233460187912, | |
| "learning_rate": 6.24e-06, | |
| "loss": 0.3333, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 2.14, | |
| "grad_norm": 0.3584277331829071, | |
| "learning_rate": 7.04e-06, | |
| "loss": 0.3153, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 2.38, | |
| "grad_norm": 0.3396163582801819, | |
| "learning_rate": 7.840000000000001e-06, | |
| "loss": 0.3337, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 2.62, | |
| "grad_norm": 0.4359325170516968, | |
| "learning_rate": 8.64e-06, | |
| "loss": 0.3309, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 2.86, | |
| "grad_norm": 0.35861968994140625, | |
| "learning_rate": 9.440000000000001e-06, | |
| "loss": 0.34, | |
| "step": 120 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 126, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 608348932669440.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |