| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 15.0, | |
| "eval_steps": 500, | |
| "global_step": 4770, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.9968553459119497, | |
| "grad_norm": 0.8092947006225586, | |
| "learning_rate": 0.0004835195985179114, | |
| "loss": 0.2576, | |
| "step": 317 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7422580645161291, | |
| "eval_loss": 0.11835014075040817, | |
| "eval_runtime": 5.4376, | |
| "eval_samples_per_second": 570.107, | |
| "eval_steps_per_second": 11.954, | |
| "step": 318 | |
| }, | |
| { | |
| "epoch": 1.9937106918238994, | |
| "grad_norm": 0.4659373164176941, | |
| "learning_rate": 0.0004678272940120885, | |
| "loss": 0.0912, | |
| "step": 634 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8132258064516129, | |
| "eval_loss": 0.08229727298021317, | |
| "eval_runtime": 5.3212, | |
| "eval_samples_per_second": 582.581, | |
| "eval_steps_per_second": 12.215, | |
| "step": 636 | |
| }, | |
| { | |
| "epoch": 2.990566037735849, | |
| "grad_norm": 0.39092108607292175, | |
| "learning_rate": 0.0004424318300788979, | |
| "loss": 0.058, | |
| "step": 951 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.853225806451613, | |
| "eval_loss": 0.07101583480834961, | |
| "eval_runtime": 5.4193, | |
| "eval_samples_per_second": 572.031, | |
| "eval_steps_per_second": 11.994, | |
| "step": 954 | |
| }, | |
| { | |
| "epoch": 3.9874213836477987, | |
| "grad_norm": 0.18997839093208313, | |
| "learning_rate": 0.00040843616613818045, | |
| "loss": 0.0468, | |
| "step": 1268 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8619354838709677, | |
| "eval_loss": 0.061981625854969025, | |
| "eval_runtime": 5.4233, | |
| "eval_samples_per_second": 571.613, | |
| "eval_steps_per_second": 11.985, | |
| "step": 1272 | |
| }, | |
| { | |
| "epoch": 4.984276729559748, | |
| "grad_norm": 0.57562655210495, | |
| "learning_rate": 0.00036731677995288685, | |
| "loss": 0.0404, | |
| "step": 1585 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.8870967741935484, | |
| "eval_loss": 0.05836557596921921, | |
| "eval_runtime": 5.4115, | |
| "eval_samples_per_second": 572.852, | |
| "eval_steps_per_second": 12.011, | |
| "step": 1590 | |
| }, | |
| { | |
| "epoch": 5.981132075471698, | |
| "grad_norm": 0.656391978263855, | |
| "learning_rate": 0.0003208595421986017, | |
| "loss": 0.0363, | |
| "step": 1902 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.8767741935483871, | |
| "eval_loss": 0.057680290192365646, | |
| "eval_runtime": 5.4159, | |
| "eval_samples_per_second": 572.384, | |
| "eval_steps_per_second": 12.002, | |
| "step": 1908 | |
| }, | |
| { | |
| "epoch": 6.977987421383648, | |
| "grad_norm": 0.21630945801734924, | |
| "learning_rate": 0.00027108215367951916, | |
| "loss": 0.0284, | |
| "step": 2219 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.8964516129032258, | |
| "eval_loss": 0.05861014127731323, | |
| "eval_runtime": 5.4131, | |
| "eval_samples_per_second": 572.683, | |
| "eval_steps_per_second": 12.008, | |
| "step": 2226 | |
| }, | |
| { | |
| "epoch": 7.9748427672955975, | |
| "grad_norm": 0.07545796036720276, | |
| "learning_rate": 0.00022014651384770874, | |
| "loss": 0.0245, | |
| "step": 2536 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.9048387096774193, | |
| "eval_loss": 0.04624096304178238, | |
| "eval_runtime": 5.409, | |
| "eval_samples_per_second": 573.114, | |
| "eval_steps_per_second": 12.017, | |
| "step": 2544 | |
| }, | |
| { | |
| "epoch": 8.971698113207546, | |
| "grad_norm": 0.08013833314180374, | |
| "learning_rate": 0.000170264826579655, | |
| "loss": 0.0209, | |
| "step": 2853 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.9106451612903226, | |
| "eval_loss": 0.042855095118284225, | |
| "eval_runtime": 5.3995, | |
| "eval_samples_per_second": 574.13, | |
| "eval_steps_per_second": 12.038, | |
| "step": 2862 | |
| }, | |
| { | |
| "epoch": 9.968553459119496, | |
| "grad_norm": 0.14883077144622803, | |
| "learning_rate": 0.00012360352116346234, | |
| "loss": 0.0192, | |
| "step": 3170 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.9132258064516129, | |
| "eval_loss": 0.04050453379750252, | |
| "eval_runtime": 5.3574, | |
| "eval_samples_per_second": 578.637, | |
| "eval_steps_per_second": 12.133, | |
| "step": 3180 | |
| }, | |
| { | |
| "epoch": 10.965408805031446, | |
| "grad_norm": 0.19123013317584991, | |
| "learning_rate": 8.218916133847685e-05, | |
| "loss": 0.0179, | |
| "step": 3487 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_accuracy": 0.9125806451612903, | |
| "eval_loss": 0.04078555479645729, | |
| "eval_runtime": 5.4382, | |
| "eval_samples_per_second": 570.045, | |
| "eval_steps_per_second": 11.953, | |
| "step": 3498 | |
| }, | |
| { | |
| "epoch": 11.962264150943396, | |
| "grad_norm": 0.04665813222527504, | |
| "learning_rate": 4.782042888526468e-05, | |
| "loss": 0.0168, | |
| "step": 3804 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_accuracy": 0.9141935483870968, | |
| "eval_loss": 0.039393454790115356, | |
| "eval_runtime": 5.4122, | |
| "eval_samples_per_second": 572.776, | |
| "eval_steps_per_second": 12.01, | |
| "step": 3816 | |
| }, | |
| { | |
| "epoch": 12.959119496855346, | |
| "grad_norm": 0.05185122787952423, | |
| "learning_rate": 2.1990004437934068e-05, | |
| "loss": 0.016, | |
| "step": 4121 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_accuracy": 0.9141935483870968, | |
| "eval_loss": 0.038763027638196945, | |
| "eval_runtime": 5.3902, | |
| "eval_samples_per_second": 575.122, | |
| "eval_steps_per_second": 12.059, | |
| "step": 4134 | |
| }, | |
| { | |
| "epoch": 13.955974842767295, | |
| "grad_norm": 0.04909258708357811, | |
| "learning_rate": 5.819738341004267e-06, | |
| "loss": 0.0156, | |
| "step": 4438 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_accuracy": 0.912258064516129, | |
| "eval_loss": 0.03870353475213051, | |
| "eval_runtime": 5.403, | |
| "eval_samples_per_second": 573.76, | |
| "eval_steps_per_second": 12.03, | |
| "step": 4452 | |
| }, | |
| { | |
| "epoch": 14.952830188679245, | |
| "grad_norm": 0.04776826128363609, | |
| "learning_rate": 1.1927168377902104e-08, | |
| "loss": 0.0153, | |
| "step": 4755 | |
| } | |
| ], | |
| "logging_steps": 317, | |
| "max_steps": 4770, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 15, | |
| "save_steps": 1000000000.0, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1236646073993904.0, | |
| "train_batch_size": 48, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.6530130862589958, | |
| "learning_rate": 0.0004888272055421989, | |
| "lr_scheduler_type": "cosine", | |
| "num_train_epochs": 15, | |
| "temperature": 12.988061249662817, | |
| "weight_decay": 0.12001185240531148 | |
| } | |
| } | |