| { | |
| "best_metric": 0.9282341599464417, | |
| "best_model_checkpoint": "data/Gemma-2-2B_task-1_120-samples_config-2_full_auto/checkpoint-49", | |
| "epoch": 16.0, | |
| "eval_steps": 500, | |
| "global_step": 88, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.18181818181818182, | |
| "grad_norm": 1.2380177974700928, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 2.2393, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.36363636363636365, | |
| "grad_norm": 1.3111598491668701, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 2.3571, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.7272727272727273, | |
| "grad_norm": 1.2179521322250366, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "loss": 2.2743, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.9090909090909091, | |
| "eval_loss": 2.277392625808716, | |
| "eval_runtime": 2.1055, | |
| "eval_samples_per_second": 11.398, | |
| "eval_steps_per_second": 11.398, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 1.0909090909090908, | |
| "grad_norm": 1.292388677597046, | |
| "learning_rate": 2.4e-05, | |
| "loss": 2.3235, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 1.4545454545454546, | |
| "grad_norm": 1.1814697980880737, | |
| "learning_rate": 3.2000000000000005e-05, | |
| "loss": 2.2247, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 1.8181818181818183, | |
| "grad_norm": 1.2490177154541016, | |
| "learning_rate": 4e-05, | |
| "loss": 2.2016, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 1.9580740928649902, | |
| "eval_runtime": 2.1, | |
| "eval_samples_per_second": 11.428, | |
| "eval_steps_per_second": 11.428, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 2.1818181818181817, | |
| "grad_norm": 1.0402225255966187, | |
| "learning_rate": 4.8e-05, | |
| "loss": 1.9202, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 2.5454545454545454, | |
| "grad_norm": 0.9636721014976501, | |
| "learning_rate": 5.6000000000000006e-05, | |
| "loss": 1.8448, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "grad_norm": 0.8552557229995728, | |
| "learning_rate": 6.400000000000001e-05, | |
| "loss": 1.6799, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "eval_loss": 1.5710538625717163, | |
| "eval_runtime": 2.1155, | |
| "eval_samples_per_second": 11.345, | |
| "eval_steps_per_second": 11.345, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 3.2727272727272725, | |
| "grad_norm": 0.7890507578849792, | |
| "learning_rate": 7.2e-05, | |
| "loss": 1.5023, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 3.6363636363636362, | |
| "grad_norm": 0.8660544157028198, | |
| "learning_rate": 8e-05, | |
| "loss": 1.4211, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.921505868434906, | |
| "learning_rate": 8.800000000000001e-05, | |
| "loss": 1.3079, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 1.2307935953140259, | |
| "eval_runtime": 2.1412, | |
| "eval_samples_per_second": 11.209, | |
| "eval_steps_per_second": 11.209, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 4.363636363636363, | |
| "grad_norm": 0.8225371241569519, | |
| "learning_rate": 9.6e-05, | |
| "loss": 1.2072, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 4.7272727272727275, | |
| "grad_norm": 0.6062935590744019, | |
| "learning_rate": 9.999512620046522e-05, | |
| "loss": 1.0919, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 4.909090909090909, | |
| "eval_loss": 1.0582603216171265, | |
| "eval_runtime": 2.1482, | |
| "eval_samples_per_second": 11.172, | |
| "eval_steps_per_second": 11.172, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 5.090909090909091, | |
| "grad_norm": 0.5227900147438049, | |
| "learning_rate": 9.995614150494293e-05, | |
| "loss": 1.0025, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 5.454545454545454, | |
| "grad_norm": 0.42430010437965393, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 0.9657, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 5.818181818181818, | |
| "grad_norm": 0.40398141741752625, | |
| "learning_rate": 9.976136999909156e-05, | |
| "loss": 0.9651, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.9889283776283264, | |
| "eval_runtime": 2.1411, | |
| "eval_samples_per_second": 11.209, | |
| "eval_steps_per_second": 11.209, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 6.181818181818182, | |
| "grad_norm": 0.4089219868183136, | |
| "learning_rate": 9.96057350657239e-05, | |
| "loss": 0.9109, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 6.545454545454545, | |
| "grad_norm": 0.41858747601509094, | |
| "learning_rate": 9.941141907232765e-05, | |
| "loss": 0.9576, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "grad_norm": 0.4120303690433502, | |
| "learning_rate": 9.917857354066931e-05, | |
| "loss": 0.8719, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "eval_loss": 0.9587691426277161, | |
| "eval_runtime": 2.1442, | |
| "eval_samples_per_second": 11.193, | |
| "eval_steps_per_second": 11.193, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 7.2727272727272725, | |
| "grad_norm": 0.3998332619667053, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 0.8755, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 7.636363636363637, | |
| "grad_norm": 0.4282527565956116, | |
| "learning_rate": 9.859805002892732e-05, | |
| "loss": 0.7937, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.4126659035682678, | |
| "learning_rate": 9.825082472361557e-05, | |
| "loss": 0.8299, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.933229386806488, | |
| "eval_runtime": 2.1461, | |
| "eval_samples_per_second": 11.183, | |
| "eval_steps_per_second": 11.183, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 8.363636363636363, | |
| "grad_norm": 0.43723729252815247, | |
| "learning_rate": 9.786597487660337e-05, | |
| "loss": 0.7659, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 8.727272727272727, | |
| "grad_norm": 0.4273925721645355, | |
| "learning_rate": 9.744380058222483e-05, | |
| "loss": 0.7912, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 8.909090909090908, | |
| "eval_loss": 0.9282341599464417, | |
| "eval_runtime": 2.1547, | |
| "eval_samples_per_second": 11.138, | |
| "eval_steps_per_second": 11.138, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 9.090909090909092, | |
| "grad_norm": 0.48125049471855164, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.7628, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 9.454545454545455, | |
| "grad_norm": 0.4407140910625458, | |
| "learning_rate": 9.648882429441257e-05, | |
| "loss": 0.7257, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 9.818181818181818, | |
| "grad_norm": 0.46118631958961487, | |
| "learning_rate": 9.595676696276172e-05, | |
| "loss": 0.692, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 0.9331913590431213, | |
| "eval_runtime": 2.1089, | |
| "eval_samples_per_second": 11.38, | |
| "eval_steps_per_second": 11.38, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 10.181818181818182, | |
| "grad_norm": 0.4512542486190796, | |
| "learning_rate": 9.538887392664544e-05, | |
| "loss": 0.6853, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 10.545454545454545, | |
| "grad_norm": 0.4982048273086548, | |
| "learning_rate": 9.478558801197065e-05, | |
| "loss": 0.6613, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "grad_norm": 0.5843939185142517, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.63, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "eval_loss": 0.9538729190826416, | |
| "eval_runtime": 2.1016, | |
| "eval_samples_per_second": 11.42, | |
| "eval_steps_per_second": 11.42, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 11.272727272727273, | |
| "grad_norm": 0.5563876628875732, | |
| "learning_rate": 9.347474647526095e-05, | |
| "loss": 0.5712, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 11.636363636363637, | |
| "grad_norm": 0.5716339349746704, | |
| "learning_rate": 9.276821300802534e-05, | |
| "loss": 0.5642, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 0.6912764310836792, | |
| "learning_rate": 9.202833017478422e-05, | |
| "loss": 0.5784, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 1.0077999830245972, | |
| "eval_runtime": 2.0975, | |
| "eval_samples_per_second": 11.442, | |
| "eval_steps_per_second": 11.442, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 12.363636363636363, | |
| "grad_norm": 0.6151651740074158, | |
| "learning_rate": 9.125567491391476e-05, | |
| "loss": 0.4975, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 12.727272727272727, | |
| "grad_norm": 0.6696138978004456, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.4937, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 12.909090909090908, | |
| "eval_loss": 1.081881046295166, | |
| "eval_runtime": 2.104, | |
| "eval_samples_per_second": 11.407, | |
| "eval_steps_per_second": 11.407, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 13.090909090909092, | |
| "grad_norm": 0.760191798210144, | |
| "learning_rate": 8.961448216775954e-05, | |
| "loss": 0.4553, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 13.454545454545455, | |
| "grad_norm": 0.742333710193634, | |
| "learning_rate": 8.874722443520899e-05, | |
| "loss": 0.3933, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 13.818181818181818, | |
| "grad_norm": 0.8903003931045532, | |
| "learning_rate": 8.784975278258783e-05, | |
| "loss": 0.4029, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_loss": 1.2031666040420532, | |
| "eval_runtime": 2.102, | |
| "eval_samples_per_second": 11.418, | |
| "eval_steps_per_second": 11.418, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 14.181818181818182, | |
| "grad_norm": 0.7762343287467957, | |
| "learning_rate": 8.692276703129421e-05, | |
| "loss": 0.3706, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 14.545454545454545, | |
| "grad_norm": 0.9183315634727478, | |
| "learning_rate": 8.596699001693255e-05, | |
| "loss": 0.3174, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 14.909090909090908, | |
| "grad_norm": 1.0166970491409302, | |
| "learning_rate": 8.498316702566828e-05, | |
| "loss": 0.3117, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 14.909090909090908, | |
| "eval_loss": 1.391228199005127, | |
| "eval_runtime": 2.112, | |
| "eval_samples_per_second": 11.364, | |
| "eval_steps_per_second": 11.364, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 15.272727272727273, | |
| "grad_norm": 0.8107355833053589, | |
| "learning_rate": 8.397206521307584e-05, | |
| "loss": 0.2381, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 15.636363636363637, | |
| "grad_norm": 0.8892229795455933, | |
| "learning_rate": 8.293447300593402e-05, | |
| "loss": 0.2356, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 1.222302794456482, | |
| "learning_rate": 8.18711994874345e-05, | |
| "loss": 0.2521, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_loss": 1.5847429037094116, | |
| "eval_runtime": 2.1083, | |
| "eval_samples_per_second": 11.384, | |
| "eval_steps_per_second": 11.384, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "step": 88, | |
| "total_flos": 6108365199507456.0, | |
| "train_loss": 0.974235578355464, | |
| "train_runtime": 616.5144, | |
| "train_samples_per_second": 7.137, | |
| "train_steps_per_second": 0.406 | |
| } | |
| ], | |
| "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": 6108365199507456.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |