| { | |
| "best_metric": 0.5567514677103719, | |
| "best_model_checkpoint": "tiny-bert-sst2-distilled/run-4/checkpoint-704", | |
| "epoch": 8.0, | |
| "eval_steps": 500, | |
| "global_step": 704, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 0.4811249077320099, | |
| "learning_rate": 6.084357840038042e-05, | |
| "loss": 0.24, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.5430528375733855, | |
| "eval_f1": 0.2098138747884941, | |
| "eval_loss": 0.21115726232528687, | |
| "eval_precision": 0.775, | |
| "eval_recall": 0.12133072407045009, | |
| "eval_runtime": 31.2118, | |
| "eval_samples_per_second": 32.744, | |
| "eval_steps_per_second": 1.025, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 0.9260004758834839, | |
| "learning_rate": 5.408318080033815e-05, | |
| "loss": 0.2091, | |
| "step": 176 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.541095890410959, | |
| "eval_f1": 0.20642978003384094, | |
| "eval_loss": 0.1988278329372406, | |
| "eval_precision": 0.7625, | |
| "eval_recall": 0.11937377690802348, | |
| "eval_runtime": 30.8172, | |
| "eval_samples_per_second": 33.163, | |
| "eval_steps_per_second": 1.038, | |
| "step": 176 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 0.5467262864112854, | |
| "learning_rate": 4.732278320029588e-05, | |
| "loss": 0.2019, | |
| "step": 264 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.5420743639921722, | |
| "eval_f1": 0.20945945945945946, | |
| "eval_loss": 0.1968851089477539, | |
| "eval_precision": 0.7654320987654321, | |
| "eval_recall": 0.12133072407045009, | |
| "eval_runtime": 31.5996, | |
| "eval_samples_per_second": 32.342, | |
| "eval_steps_per_second": 1.013, | |
| "step": 264 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.3741915822029114, | |
| "learning_rate": 4.0562385600253614e-05, | |
| "loss": 0.2013, | |
| "step": 352 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.5391389432485323, | |
| "eval_f1": 0.19761499148211242, | |
| "eval_loss": 0.19684568047523499, | |
| "eval_precision": 0.7631578947368421, | |
| "eval_recall": 0.11350293542074363, | |
| "eval_runtime": 30.3074, | |
| "eval_samples_per_second": 33.721, | |
| "eval_steps_per_second": 1.056, | |
| "step": 352 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.39911171793937683, | |
| "learning_rate": 3.3801988000211345e-05, | |
| "loss": 0.2002, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.547945205479452, | |
| "eval_f1": 0.23255813953488372, | |
| "eval_loss": 0.1965479999780655, | |
| "eval_precision": 0.7692307692307693, | |
| "eval_recall": 0.136986301369863, | |
| "eval_runtime": 29.369, | |
| "eval_samples_per_second": 34.799, | |
| "eval_steps_per_second": 1.09, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 1.0663460493087769, | |
| "learning_rate": 2.7041590400169076e-05, | |
| "loss": 0.1998, | |
| "step": 528 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.5557729941291585, | |
| "eval_f1": 0.26058631921824105, | |
| "eval_loss": 0.19687490165233612, | |
| "eval_precision": 0.7766990291262136, | |
| "eval_recall": 0.15655577299412915, | |
| "eval_runtime": 29.649, | |
| "eval_samples_per_second": 34.47, | |
| "eval_steps_per_second": 1.079, | |
| "step": 528 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 0.4529750347137451, | |
| "learning_rate": 2.0281192800126807e-05, | |
| "loss": 0.1992, | |
| "step": 616 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.5499021526418787, | |
| "eval_f1": 0.24092409240924093, | |
| "eval_loss": 0.19784095883369446, | |
| "eval_precision": 0.7684210526315789, | |
| "eval_recall": 0.14285714285714285, | |
| "eval_runtime": 29.7832, | |
| "eval_samples_per_second": 34.315, | |
| "eval_steps_per_second": 1.074, | |
| "step": 616 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.8743975162506104, | |
| "learning_rate": 1.3520795200084538e-05, | |
| "loss": 0.1989, | |
| "step": 704 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.5567514677103719, | |
| "eval_f1": 0.2634146341463414, | |
| "eval_loss": 0.19613005220890045, | |
| "eval_precision": 0.7788461538461539, | |
| "eval_recall": 0.15851272015655576, | |
| "eval_runtime": 28.6336, | |
| "eval_samples_per_second": 35.692, | |
| "eval_steps_per_second": 1.118, | |
| "step": 704 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 880, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "save_steps": 500, | |
| "total_flos": 1885561578240.0, | |
| "train_batch_size": 35, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.2634542499188758, | |
| "learning_rate": 6.760397600042269e-05, | |
| "num_train_epochs": 10, | |
| "per_device_train_batch_size": 35, | |
| "temperature": 2 | |
| } | |
| } | |