| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 9.433962264150944, | |
| "eval_steps": 500, | |
| "global_step": 3000, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.6016129032258064, | |
| "eval_loss": 0.2116461843252182, | |
| "eval_runtime": 16.545, | |
| "eval_samples_per_second": 187.367, | |
| "eval_steps_per_second": 23.451, | |
| "step": 318 | |
| }, | |
| { | |
| "epoch": 1.5723270440251573, | |
| "grad_norm": 0.5623254179954529, | |
| "learning_rate": 1.685534591194969e-05, | |
| "loss": 0.3397, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8306451612903226, | |
| "eval_loss": 0.0993104875087738, | |
| "eval_runtime": 15.946, | |
| "eval_samples_per_second": 194.406, | |
| "eval_steps_per_second": 24.332, | |
| "step": 636 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8909677419354839, | |
| "eval_loss": 0.06301400810480118, | |
| "eval_runtime": 20.5764, | |
| "eval_samples_per_second": 150.658, | |
| "eval_steps_per_second": 18.857, | |
| "step": 954 | |
| }, | |
| { | |
| "epoch": 3.1446540880503147, | |
| "grad_norm": 0.5391496419906616, | |
| "learning_rate": 1.371069182389937e-05, | |
| "loss": 0.1148, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.9045161290322581, | |
| "eval_loss": 0.04739077761769295, | |
| "eval_runtime": 15.9991, | |
| "eval_samples_per_second": 193.761, | |
| "eval_steps_per_second": 24.251, | |
| "step": 1272 | |
| }, | |
| { | |
| "epoch": 4.716981132075472, | |
| "grad_norm": 0.44755664467811584, | |
| "learning_rate": 1.0566037735849058e-05, | |
| "loss": 0.0708, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.9129032258064517, | |
| "eval_loss": 0.038904447108507156, | |
| "eval_runtime": 17.1562, | |
| "eval_samples_per_second": 180.692, | |
| "eval_steps_per_second": 22.616, | |
| "step": 1590 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.9219354838709677, | |
| "eval_loss": 0.03372984379529953, | |
| "eval_runtime": 18.1417, | |
| "eval_samples_per_second": 170.877, | |
| "eval_steps_per_second": 21.387, | |
| "step": 1908 | |
| }, | |
| { | |
| "epoch": 6.289308176100629, | |
| "grad_norm": 0.19874028861522675, | |
| "learning_rate": 7.421383647798742e-06, | |
| "loss": 0.0549, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.9261290322580645, | |
| "eval_loss": 0.03085845522582531, | |
| "eval_runtime": 17.1033, | |
| "eval_samples_per_second": 181.251, | |
| "eval_steps_per_second": 22.686, | |
| "step": 2226 | |
| }, | |
| { | |
| "epoch": 7.861635220125786, | |
| "grad_norm": 0.22561663389205933, | |
| "learning_rate": 4.276729559748428e-06, | |
| "loss": 0.0473, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.9287096774193548, | |
| "eval_loss": 0.028838176280260086, | |
| "eval_runtime": 15.9646, | |
| "eval_samples_per_second": 194.18, | |
| "eval_steps_per_second": 24.304, | |
| "step": 2544 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.93, | |
| "eval_loss": 0.02777356281876564, | |
| "eval_runtime": 15.8351, | |
| "eval_samples_per_second": 195.768, | |
| "eval_steps_per_second": 24.503, | |
| "step": 2862 | |
| }, | |
| { | |
| "epoch": 9.433962264150944, | |
| "grad_norm": 0.3520292639732361, | |
| "learning_rate": 1.1320754716981133e-06, | |
| "loss": 0.0436, | |
| "step": 3000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 3180, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 780738843279612.0, | |
| "train_batch_size": 48, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.7461709946754571, | |
| "num_train_epochs": 10, | |
| "temperature": 7 | |
| } | |
| } | |