| { | |
| "best_metric": 0.4495352804660797, | |
| "best_model_checkpoint": "data/Gemma-2-2B_task-3_120-samples_config-2/checkpoint-44", | |
| "epoch": 14.909090909090908, | |
| "eval_steps": 500, | |
| "global_step": 82, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.18181818181818182, | |
| "grad_norm": 8.376895904541016, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 3.9078, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.36363636363636365, | |
| "grad_norm": 6.967246055603027, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 3.4781, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.7272727272727273, | |
| "grad_norm": 6.186585426330566, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "loss": 3.2814, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.9090909090909091, | |
| "eval_loss": 3.220660924911499, | |
| "eval_runtime": 7.6385, | |
| "eval_samples_per_second": 3.142, | |
| "eval_steps_per_second": 3.142, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 1.0909090909090908, | |
| "grad_norm": 6.225273132324219, | |
| "learning_rate": 2.4e-05, | |
| "loss": 3.1509, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 1.4545454545454546, | |
| "grad_norm": 7.809123516082764, | |
| "learning_rate": 3.2000000000000005e-05, | |
| "loss": 3.129, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 1.8181818181818183, | |
| "grad_norm": 5.321211338043213, | |
| "learning_rate": 4e-05, | |
| "loss": 2.3853, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 1.7734723091125488, | |
| "eval_runtime": 7.4521, | |
| "eval_samples_per_second": 3.221, | |
| "eval_steps_per_second": 3.221, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 2.1818181818181817, | |
| "grad_norm": 4.0745649337768555, | |
| "learning_rate": 4.8e-05, | |
| "loss": 1.959, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 2.5454545454545454, | |
| "grad_norm": 3.111417293548584, | |
| "learning_rate": 5.6000000000000006e-05, | |
| "loss": 1.4555, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "grad_norm": 2.917243242263794, | |
| "learning_rate": 6.400000000000001e-05, | |
| "loss": 1.0517, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "eval_loss": 0.9646134376525879, | |
| "eval_runtime": 7.4578, | |
| "eval_samples_per_second": 3.218, | |
| "eval_steps_per_second": 3.218, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 3.2727272727272725, | |
| "grad_norm": 2.1608688831329346, | |
| "learning_rate": 7.2e-05, | |
| "loss": 0.7279, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 3.6363636363636362, | |
| "grad_norm": 0.9096205234527588, | |
| "learning_rate": 8e-05, | |
| "loss": 0.6856, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.700407862663269, | |
| "learning_rate": 8.800000000000001e-05, | |
| "loss": 0.6575, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 0.6095558404922485, | |
| "eval_runtime": 7.4569, | |
| "eval_samples_per_second": 3.219, | |
| "eval_steps_per_second": 3.219, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 4.363636363636363, | |
| "grad_norm": 0.7757139801979065, | |
| "learning_rate": 9.6e-05, | |
| "loss": 0.5118, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 4.7272727272727275, | |
| "grad_norm": 0.5130665898323059, | |
| "learning_rate": 9.999512620046522e-05, | |
| "loss": 0.4492, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 4.909090909090909, | |
| "eval_loss": 0.5386365056037903, | |
| "eval_runtime": 7.4523, | |
| "eval_samples_per_second": 3.22, | |
| "eval_steps_per_second": 3.22, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 5.090909090909091, | |
| "grad_norm": 0.6565605401992798, | |
| "learning_rate": 9.995614150494293e-05, | |
| "loss": 0.4765, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 5.454545454545454, | |
| "grad_norm": 0.7080961465835571, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 0.3665, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 5.818181818181818, | |
| "grad_norm": 0.367472767829895, | |
| "learning_rate": 9.976136999909156e-05, | |
| "loss": 0.3535, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.48484835028648376, | |
| "eval_runtime": 7.4527, | |
| "eval_samples_per_second": 3.22, | |
| "eval_steps_per_second": 3.22, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 6.181818181818182, | |
| "grad_norm": 0.3141935169696808, | |
| "learning_rate": 9.96057350657239e-05, | |
| "loss": 0.3497, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 6.545454545454545, | |
| "grad_norm": 0.5521201491355896, | |
| "learning_rate": 9.941141907232765e-05, | |
| "loss": 0.4009, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "grad_norm": 0.5136641263961792, | |
| "learning_rate": 9.917857354066931e-05, | |
| "loss": 0.3521, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "eval_loss": 0.4612942636013031, | |
| "eval_runtime": 7.4513, | |
| "eval_samples_per_second": 3.221, | |
| "eval_steps_per_second": 3.221, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 7.2727272727272725, | |
| "grad_norm": 0.4848323464393616, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 0.3128, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 7.636363636363637, | |
| "grad_norm": 0.34514158964157104, | |
| "learning_rate": 9.859805002892732e-05, | |
| "loss": 0.2022, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.42058125138282776, | |
| "learning_rate": 9.825082472361557e-05, | |
| "loss": 0.2821, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.4495352804660797, | |
| "eval_runtime": 7.4485, | |
| "eval_samples_per_second": 3.222, | |
| "eval_steps_per_second": 3.222, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 8.363636363636363, | |
| "grad_norm": 0.4747871160507202, | |
| "learning_rate": 9.786597487660337e-05, | |
| "loss": 0.2727, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 8.727272727272727, | |
| "grad_norm": 0.41476812958717346, | |
| "learning_rate": 9.744380058222483e-05, | |
| "loss": 0.1878, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 8.909090909090908, | |
| "eval_loss": 0.44962453842163086, | |
| "eval_runtime": 7.4531, | |
| "eval_samples_per_second": 3.22, | |
| "eval_steps_per_second": 3.22, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 9.090909090909092, | |
| "grad_norm": 0.48273763060569763, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.2823, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 9.454545454545455, | |
| "grad_norm": 0.5613682866096497, | |
| "learning_rate": 9.648882429441257e-05, | |
| "loss": 0.215, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 9.818181818181818, | |
| "grad_norm": 0.843788206577301, | |
| "learning_rate": 9.595676696276172e-05, | |
| "loss": 0.1547, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 0.49282336235046387, | |
| "eval_runtime": 7.449, | |
| "eval_samples_per_second": 3.222, | |
| "eval_steps_per_second": 3.222, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 10.181818181818182, | |
| "grad_norm": 0.46964725852012634, | |
| "learning_rate": 9.538887392664544e-05, | |
| "loss": 0.1365, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 10.545454545454545, | |
| "grad_norm": 0.5018220543861389, | |
| "learning_rate": 9.478558801197065e-05, | |
| "loss": 0.1599, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "grad_norm": 0.4618806838989258, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.1539, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "eval_loss": 0.49774304032325745, | |
| "eval_runtime": 7.4564, | |
| "eval_samples_per_second": 3.219, | |
| "eval_steps_per_second": 3.219, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 11.272727272727273, | |
| "grad_norm": 0.5974417924880981, | |
| "learning_rate": 9.347474647526095e-05, | |
| "loss": 0.095, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 11.636363636363637, | |
| "grad_norm": 0.5028112530708313, | |
| "learning_rate": 9.276821300802534e-05, | |
| "loss": 0.1075, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 1.2034255266189575, | |
| "learning_rate": 9.202833017478422e-05, | |
| "loss": 0.1194, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 0.6055799722671509, | |
| "eval_runtime": 7.4506, | |
| "eval_samples_per_second": 3.221, | |
| "eval_steps_per_second": 3.221, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 12.363636363636363, | |
| "grad_norm": 0.3660883903503418, | |
| "learning_rate": 9.125567491391476e-05, | |
| "loss": 0.0486, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 12.727272727272727, | |
| "grad_norm": 0.7464652061462402, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.0872, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 12.909090909090908, | |
| "eval_loss": 0.693767249584198, | |
| "eval_runtime": 7.4557, | |
| "eval_samples_per_second": 3.219, | |
| "eval_steps_per_second": 3.219, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 13.090909090909092, | |
| "grad_norm": 0.45991790294647217, | |
| "learning_rate": 8.961448216775954e-05, | |
| "loss": 0.0513, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 13.454545454545455, | |
| "grad_norm": 0.5081632137298584, | |
| "learning_rate": 8.874722443520899e-05, | |
| "loss": 0.0305, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 13.818181818181818, | |
| "grad_norm": 1.0700777769088745, | |
| "learning_rate": 8.784975278258783e-05, | |
| "loss": 0.0274, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_loss": 0.8378891944885254, | |
| "eval_runtime": 7.4657, | |
| "eval_samples_per_second": 3.215, | |
| "eval_steps_per_second": 3.215, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 14.181818181818182, | |
| "grad_norm": 0.23455728590488434, | |
| "learning_rate": 8.692276703129421e-05, | |
| "loss": 0.0294, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 14.545454545454545, | |
| "grad_norm": 1.548729658126831, | |
| "learning_rate": 8.596699001693255e-05, | |
| "loss": 0.0247, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 14.909090909090908, | |
| "grad_norm": 0.47893524169921875, | |
| "learning_rate": 8.498316702566828e-05, | |
| "loss": 0.0195, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 14.909090909090908, | |
| "eval_loss": 0.8741180896759033, | |
| "eval_runtime": 7.4521, | |
| "eval_samples_per_second": 3.221, | |
| "eval_steps_per_second": 3.221, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 14.909090909090908, | |
| "step": 82, | |
| "total_flos": 3.5471186917851136e+16, | |
| "train_loss": 0.693681104949153, | |
| "train_runtime": 1505.781, | |
| "train_samples_per_second": 2.922, | |
| "train_steps_per_second": 0.166 | |
| } | |
| ], | |
| "logging_steps": 2, | |
| "max_steps": 250, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 50, | |
| "save_steps": 25, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 7, | |
| "early_stopping_threshold": 0.0 | |
| }, | |
| "attributes": { | |
| "early_stopping_patience_counter": 0 | |
| } | |
| }, | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 3.5471186917851136e+16, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |