| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 6.289308176100629, | |
| "eval_steps": 500, | |
| "global_step": 2000, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.6135483870967742, | |
| "eval_loss": 0.23216405510902405, | |
| "eval_runtime": 14.5556, | |
| "eval_samples_per_second": 212.976, | |
| "eval_steps_per_second": 26.656, | |
| "step": 318 | |
| }, | |
| { | |
| "epoch": 1.5723270440251573, | |
| "grad_norm": 0.5930498838424683, | |
| "learning_rate": 1.606918238993711e-05, | |
| "loss": 0.3696, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8325806451612904, | |
| "eval_loss": 0.1069631278514862, | |
| "eval_runtime": 15.7006, | |
| "eval_samples_per_second": 197.444, | |
| "eval_steps_per_second": 24.712, | |
| "step": 636 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8890322580645161, | |
| "eval_loss": 0.0675421878695488, | |
| "eval_runtime": 15.266, | |
| "eval_samples_per_second": 203.065, | |
| "eval_steps_per_second": 25.416, | |
| "step": 954 | |
| }, | |
| { | |
| "epoch": 3.1446540880503147, | |
| "grad_norm": 0.593765914440155, | |
| "learning_rate": 1.2138364779874214e-05, | |
| "loss": 0.1241, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.9048387096774193, | |
| "eval_loss": 0.05093123018741608, | |
| "eval_runtime": 14.5589, | |
| "eval_samples_per_second": 212.928, | |
| "eval_steps_per_second": 26.65, | |
| "step": 1272 | |
| }, | |
| { | |
| "epoch": 4.716981132075472, | |
| "grad_norm": 0.48284056782722473, | |
| "learning_rate": 8.207547169811321e-06, | |
| "loss": 0.0767, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9083870967741936, | |
| "eval_loss": 0.0422496534883976, | |
| "eval_runtime": 15.2188, | |
| "eval_samples_per_second": 203.695, | |
| "eval_steps_per_second": 25.495, | |
| "step": 1590 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.915483870967742, | |
| "eval_loss": 0.037340711802244186, | |
| "eval_runtime": 15.6107, | |
| "eval_samples_per_second": 198.581, | |
| "eval_steps_per_second": 24.855, | |
| "step": 1908 | |
| }, | |
| { | |
| "epoch": 6.289308176100629, | |
| "grad_norm": 0.24097558856010437, | |
| "learning_rate": 4.276729559748428e-06, | |
| "loss": 0.0607, | |
| "step": 2000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 2544, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 8, | |
| "save_steps": 500, | |
| "total_flos": 520991326672152.0, | |
| "train_batch_size": 48, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.3077189989743373, | |
| "num_train_epochs": 8, | |
| "temperature": 5 | |
| } | |
| } | |