| { | |
| "best_metric": 0.6712512373924255, | |
| "best_model_checkpoint": "data/Gemma-2-2B_task-2_120-samples_config-2/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": 0.94005286693573, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 1.1048, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.36363636363636365, | |
| "grad_norm": 0.8046804070472717, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 0.9651, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.7272727272727273, | |
| "grad_norm": 0.8616108894348145, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "loss": 1.0825, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.9090909090909091, | |
| "eval_loss": 1.0681871175765991, | |
| "eval_runtime": 9.3935, | |
| "eval_samples_per_second": 2.555, | |
| "eval_steps_per_second": 2.555, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 1.0909090909090908, | |
| "grad_norm": 0.8416345715522766, | |
| "learning_rate": 2.4e-05, | |
| "loss": 0.9768, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 1.4545454545454546, | |
| "grad_norm": 0.8554661870002747, | |
| "learning_rate": 3.2000000000000005e-05, | |
| "loss": 1.0333, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 1.8181818181818183, | |
| "grad_norm": 0.6826609969139099, | |
| "learning_rate": 4e-05, | |
| "loss": 0.9795, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.9157736897468567, | |
| "eval_runtime": 9.1295, | |
| "eval_samples_per_second": 2.629, | |
| "eval_steps_per_second": 2.629, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 2.1818181818181817, | |
| "grad_norm": 0.5249937176704407, | |
| "learning_rate": 4.8e-05, | |
| "loss": 0.8301, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 2.5454545454545454, | |
| "grad_norm": 0.5049210786819458, | |
| "learning_rate": 5.6000000000000006e-05, | |
| "loss": 0.8504, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "grad_norm": 0.5010665059089661, | |
| "learning_rate": 6.400000000000001e-05, | |
| "loss": 0.7555, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "eval_loss": 0.8180524706840515, | |
| "eval_runtime": 9.1466, | |
| "eval_samples_per_second": 2.624, | |
| "eval_steps_per_second": 2.624, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 3.2727272727272725, | |
| "grad_norm": 0.3780786097049713, | |
| "learning_rate": 7.2e-05, | |
| "loss": 0.7806, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 3.6363636363636362, | |
| "grad_norm": 0.44708988070487976, | |
| "learning_rate": 8e-05, | |
| "loss": 0.7026, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.42761316895484924, | |
| "learning_rate": 8.800000000000001e-05, | |
| "loss": 0.7129, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 0.7546100616455078, | |
| "eval_runtime": 9.1285, | |
| "eval_samples_per_second": 2.629, | |
| "eval_steps_per_second": 2.629, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 4.363636363636363, | |
| "grad_norm": 0.33841633796691895, | |
| "learning_rate": 9.6e-05, | |
| "loss": 0.6834, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 4.7272727272727275, | |
| "grad_norm": 0.3635419011116028, | |
| "learning_rate": 9.999512620046522e-05, | |
| "loss": 0.6824, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 4.909090909090909, | |
| "eval_loss": 0.7212942242622375, | |
| "eval_runtime": 9.1335, | |
| "eval_samples_per_second": 2.628, | |
| "eval_steps_per_second": 2.628, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 5.090909090909091, | |
| "grad_norm": 0.26798659563064575, | |
| "learning_rate": 9.995614150494293e-05, | |
| "loss": 0.5945, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 5.454545454545454, | |
| "grad_norm": 0.32639092206954956, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 0.6373, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 5.818181818181818, | |
| "grad_norm": 0.3496810793876648, | |
| "learning_rate": 9.976136999909156e-05, | |
| "loss": 0.6146, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.6979790329933167, | |
| "eval_runtime": 9.1292, | |
| "eval_samples_per_second": 2.629, | |
| "eval_steps_per_second": 2.629, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 6.181818181818182, | |
| "grad_norm": 0.2730304002761841, | |
| "learning_rate": 9.96057350657239e-05, | |
| "loss": 0.5836, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 6.545454545454545, | |
| "grad_norm": 0.3141060471534729, | |
| "learning_rate": 9.941141907232765e-05, | |
| "loss": 0.5661, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "grad_norm": 0.27712708711624146, | |
| "learning_rate": 9.917857354066931e-05, | |
| "loss": 0.5689, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "eval_loss": 0.6853704452514648, | |
| "eval_runtime": 9.1303, | |
| "eval_samples_per_second": 2.629, | |
| "eval_steps_per_second": 2.629, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 7.2727272727272725, | |
| "grad_norm": 0.28335508704185486, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 0.5345, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 7.636363636363637, | |
| "grad_norm": 0.2927487790584564, | |
| "learning_rate": 9.859805002892732e-05, | |
| "loss": 0.5174, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.276620090007782, | |
| "learning_rate": 9.825082472361557e-05, | |
| "loss": 0.5226, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.6730685234069824, | |
| "eval_runtime": 9.1488, | |
| "eval_samples_per_second": 2.623, | |
| "eval_steps_per_second": 2.623, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 8.363636363636363, | |
| "grad_norm": 0.3022850751876831, | |
| "learning_rate": 9.786597487660337e-05, | |
| "loss": 0.475, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 8.727272727272727, | |
| "grad_norm": 0.32451099157333374, | |
| "learning_rate": 9.744380058222483e-05, | |
| "loss": 0.4847, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 8.909090909090908, | |
| "eval_loss": 0.6712512373924255, | |
| "eval_runtime": 9.1393, | |
| "eval_samples_per_second": 2.626, | |
| "eval_steps_per_second": 2.626, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 9.090909090909092, | |
| "grad_norm": 0.34935930371284485, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.467, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 9.454545454545455, | |
| "grad_norm": 0.3449249863624573, | |
| "learning_rate": 9.648882429441257e-05, | |
| "loss": 0.4225, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 9.818181818181818, | |
| "grad_norm": 0.3710680902004242, | |
| "learning_rate": 9.595676696276172e-05, | |
| "loss": 0.4181, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 0.6911107897758484, | |
| "eval_runtime": 9.1312, | |
| "eval_samples_per_second": 2.628, | |
| "eval_steps_per_second": 2.628, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 10.181818181818182, | |
| "grad_norm": 0.37772881984710693, | |
| "learning_rate": 9.538887392664544e-05, | |
| "loss": 0.4087, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 10.545454545454545, | |
| "grad_norm": 0.4116469621658325, | |
| "learning_rate": 9.478558801197065e-05, | |
| "loss": 0.3675, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "grad_norm": 0.6264581680297852, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.3633, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "eval_loss": 0.7228239178657532, | |
| "eval_runtime": 9.1352, | |
| "eval_samples_per_second": 2.627, | |
| "eval_steps_per_second": 2.627, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 11.272727272727273, | |
| "grad_norm": 0.45233646035194397, | |
| "learning_rate": 9.347474647526095e-05, | |
| "loss": 0.3166, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 11.636363636363637, | |
| "grad_norm": 0.5267932415008545, | |
| "learning_rate": 9.276821300802534e-05, | |
| "loss": 0.3261, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 0.6076023578643799, | |
| "learning_rate": 9.202833017478422e-05, | |
| "loss": 0.2845, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 0.7944682240486145, | |
| "eval_runtime": 9.1296, | |
| "eval_samples_per_second": 2.629, | |
| "eval_steps_per_second": 2.629, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 12.363636363636363, | |
| "grad_norm": 0.5722086429595947, | |
| "learning_rate": 9.125567491391476e-05, | |
| "loss": 0.2619, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 12.727272727272727, | |
| "grad_norm": 0.728240430355072, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.2149, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 12.909090909090908, | |
| "eval_loss": 0.9006109237670898, | |
| "eval_runtime": 9.1357, | |
| "eval_samples_per_second": 2.627, | |
| "eval_steps_per_second": 2.627, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 13.090909090909092, | |
| "grad_norm": 0.6510495543479919, | |
| "learning_rate": 8.961448216775954e-05, | |
| "loss": 0.2073, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 13.454545454545455, | |
| "grad_norm": 0.6451771855354309, | |
| "learning_rate": 8.874722443520899e-05, | |
| "loss": 0.1907, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 13.818181818181818, | |
| "grad_norm": 0.8164005279541016, | |
| "learning_rate": 8.784975278258783e-05, | |
| "loss": 0.1446, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_loss": 1.0793968439102173, | |
| "eval_runtime": 9.13, | |
| "eval_samples_per_second": 2.629, | |
| "eval_steps_per_second": 2.629, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 14.181818181818182, | |
| "grad_norm": 0.665549099445343, | |
| "learning_rate": 8.692276703129421e-05, | |
| "loss": 0.112, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 14.545454545454545, | |
| "grad_norm": 0.7149677872657776, | |
| "learning_rate": 8.596699001693255e-05, | |
| "loss": 0.121, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 14.909090909090908, | |
| "grad_norm": 0.7177186012268066, | |
| "learning_rate": 8.498316702566828e-05, | |
| "loss": 0.0753, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 14.909090909090908, | |
| "eval_loss": 1.2625266313552856, | |
| "eval_runtime": 9.1337, | |
| "eval_samples_per_second": 2.628, | |
| "eval_steps_per_second": 2.628, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 15.272727272727273, | |
| "grad_norm": 0.8046224117279053, | |
| "learning_rate": 8.397206521307584e-05, | |
| "loss": 0.0674, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 15.636363636363637, | |
| "grad_norm": 1.23111891746521, | |
| "learning_rate": 8.293447300593402e-05, | |
| "loss": 0.0563, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 0.6707900166511536, | |
| "learning_rate": 8.18711994874345e-05, | |
| "loss": 0.0483, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_loss": 1.4372771978378296, | |
| "eval_runtime": 9.1347, | |
| "eval_samples_per_second": 2.627, | |
| "eval_steps_per_second": 2.627, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "step": 88, | |
| "total_flos": 4.588577999762227e+16, | |
| "train_loss": 0.5017756496302106, | |
| "train_runtime": 1960.1661, | |
| "train_samples_per_second": 2.245, | |
| "train_steps_per_second": 0.128 | |
| } | |
| ], | |
| "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": 4.588577999762227e+16, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |