| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.4922698253980463, | |
| "eval_steps": 500, | |
| "global_step": 200, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 2.9999999999999997e-05, | |
| "loss": 3.623, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 5.9999999999999995e-05, | |
| "loss": 3.605, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 8.999999999999999e-05, | |
| "loss": 3.5168, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 0.00011999999999999999, | |
| "loss": 3.3511, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 0.00015, | |
| "loss": 3.1952, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 0.00017999999999999998, | |
| "loss": 3.0575, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 0.00020999999999999998, | |
| "loss": 2.8575, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 0.00023999999999999998, | |
| "loss": 2.7029, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 0.00027, | |
| "loss": 2.6392, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 0.0003, | |
| "loss": 2.5679, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 0.00027, | |
| "loss": 2.5392, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 0.00023999999999999998, | |
| "loss": 2.5153, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 0.00020999999999999998, | |
| "loss": 2.4822, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 0.00017999999999999998, | |
| "loss": 2.4678, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 0.00015, | |
| "loss": 2.4438, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 0.00011999999999999999, | |
| "loss": 2.4351, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 8.999999999999999e-05, | |
| "loss": 2.4147, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 5.9999999999999995e-05, | |
| "loss": 2.3935, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 2.9999999999999997e-05, | |
| "loss": 2.4104, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 0.0, | |
| "loss": 2.3838, | |
| "step": 200 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 200, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 2.723845771535155e+16, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |