| { | |
| "best_metric": 0.874152660369873, | |
| "best_model_checkpoint": "data/Gemma-2-2B_task-2_120-samples_config-2_full/checkpoint-66", | |
| "epoch": 18.90909090909091, | |
| "eval_steps": 500, | |
| "global_step": 104, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.18181818181818182, | |
| "grad_norm": 0.5122515559196472, | |
| "learning_rate": 4.000000000000001e-06, | |
| "loss": 1.3574, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.36363636363636365, | |
| "grad_norm": 0.4222266674041748, | |
| "learning_rate": 8.000000000000001e-06, | |
| "loss": 1.2763, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.7272727272727273, | |
| "grad_norm": 0.46601182222366333, | |
| "learning_rate": 1.6000000000000003e-05, | |
| "loss": 1.356, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.9090909090909091, | |
| "eval_loss": 1.3236860036849976, | |
| "eval_runtime": 9.3936, | |
| "eval_samples_per_second": 2.555, | |
| "eval_steps_per_second": 2.555, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 1.0909090909090908, | |
| "grad_norm": 0.4952509105205536, | |
| "learning_rate": 2.4e-05, | |
| "loss": 1.3224, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 1.4545454545454546, | |
| "grad_norm": 0.4738887846469879, | |
| "learning_rate": 3.2000000000000005e-05, | |
| "loss": 1.3152, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 1.8181818181818183, | |
| "grad_norm": 0.41166675090789795, | |
| "learning_rate": 4e-05, | |
| "loss": 1.3153, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 1.2305984497070312, | |
| "eval_runtime": 9.2736, | |
| "eval_samples_per_second": 2.588, | |
| "eval_steps_per_second": 2.588, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 2.1818181818181817, | |
| "grad_norm": 0.32252660393714905, | |
| "learning_rate": 4.8e-05, | |
| "loss": 1.1922, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 2.5454545454545454, | |
| "grad_norm": 0.27899450063705444, | |
| "learning_rate": 5.6000000000000006e-05, | |
| "loss": 1.182, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "grad_norm": 0.24849611520767212, | |
| "learning_rate": 6.400000000000001e-05, | |
| "loss": 1.1344, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 2.909090909090909, | |
| "eval_loss": 1.1312867403030396, | |
| "eval_runtime": 9.1243, | |
| "eval_samples_per_second": 2.63, | |
| "eval_steps_per_second": 2.63, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 3.2727272727272725, | |
| "grad_norm": 0.25325286388397217, | |
| "learning_rate": 7.2e-05, | |
| "loss": 1.1237, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 3.6363636363636362, | |
| "grad_norm": 0.27949851751327515, | |
| "learning_rate": 8e-05, | |
| "loss": 1.0603, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 0.24920938909053802, | |
| "learning_rate": 8.800000000000001e-05, | |
| "loss": 1.0582, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 1.051779866218567, | |
| "eval_runtime": 9.1224, | |
| "eval_samples_per_second": 2.631, | |
| "eval_steps_per_second": 2.631, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 4.363636363636363, | |
| "grad_norm": 0.235067680478096, | |
| "learning_rate": 9.6e-05, | |
| "loss": 1.0156, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 4.7272727272727275, | |
| "grad_norm": 0.2814384698867798, | |
| "learning_rate": 9.999512620046522e-05, | |
| "loss": 0.9757, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 4.909090909090909, | |
| "eval_loss": 0.9752205014228821, | |
| "eval_runtime": 9.1196, | |
| "eval_samples_per_second": 2.632, | |
| "eval_steps_per_second": 2.632, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 5.090909090909091, | |
| "grad_norm": 0.2683999538421631, | |
| "learning_rate": 9.995614150494293e-05, | |
| "loss": 0.9355, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 5.454545454545454, | |
| "grad_norm": 0.1930312216281891, | |
| "learning_rate": 9.987820251299122e-05, | |
| "loss": 0.9116, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 5.818181818181818, | |
| "grad_norm": 0.15850737690925598, | |
| "learning_rate": 9.976136999909156e-05, | |
| "loss": 0.8903, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_loss": 0.9264867901802063, | |
| "eval_runtime": 9.1206, | |
| "eval_samples_per_second": 2.631, | |
| "eval_steps_per_second": 2.631, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 6.181818181818182, | |
| "grad_norm": 0.15357622504234314, | |
| "learning_rate": 9.96057350657239e-05, | |
| "loss": 0.8729, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 6.545454545454545, | |
| "grad_norm": 0.1535169631242752, | |
| "learning_rate": 9.941141907232765e-05, | |
| "loss": 0.8513, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "grad_norm": 0.13679227232933044, | |
| "learning_rate": 9.917857354066931e-05, | |
| "loss": 0.8676, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 6.909090909090909, | |
| "eval_loss": 0.9079222679138184, | |
| "eval_runtime": 9.1249, | |
| "eval_samples_per_second": 2.63, | |
| "eval_steps_per_second": 2.63, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 7.2727272727272725, | |
| "grad_norm": 0.13816682994365692, | |
| "learning_rate": 9.890738003669029e-05, | |
| "loss": 0.8599, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 7.636363636363637, | |
| "grad_norm": 0.11960936337709427, | |
| "learning_rate": 9.859805002892732e-05, | |
| "loss": 0.8157, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 0.1116834282875061, | |
| "learning_rate": 9.825082472361557e-05, | |
| "loss": 0.8368, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_loss": 0.8935267329216003, | |
| "eval_runtime": 9.1211, | |
| "eval_samples_per_second": 2.631, | |
| "eval_steps_per_second": 2.631, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 8.363636363636363, | |
| "grad_norm": 0.1247890517115593, | |
| "learning_rate": 9.786597487660337e-05, | |
| "loss": 0.805, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 8.727272727272727, | |
| "grad_norm": 0.14664214849472046, | |
| "learning_rate": 9.744380058222483e-05, | |
| "loss": 0.8107, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 8.909090909090908, | |
| "eval_loss": 0.8844764232635498, | |
| "eval_runtime": 9.1225, | |
| "eval_samples_per_second": 2.631, | |
| "eval_steps_per_second": 2.631, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 9.090909090909092, | |
| "grad_norm": 0.1276860386133194, | |
| "learning_rate": 9.698463103929542e-05, | |
| "loss": 0.8177, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 9.454545454545455, | |
| "grad_norm": 0.12827570736408234, | |
| "learning_rate": 9.648882429441257e-05, | |
| "loss": 0.7772, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 9.818181818181818, | |
| "grad_norm": 0.12181173264980316, | |
| "learning_rate": 9.595676696276172e-05, | |
| "loss": 0.8061, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_loss": 0.8781268000602722, | |
| "eval_runtime": 9.1186, | |
| "eval_samples_per_second": 2.632, | |
| "eval_steps_per_second": 2.632, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 10.181818181818182, | |
| "grad_norm": 0.12815453112125397, | |
| "learning_rate": 9.538887392664544e-05, | |
| "loss": 0.7701, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 10.545454545454545, | |
| "grad_norm": 0.1292680948972702, | |
| "learning_rate": 9.478558801197065e-05, | |
| "loss": 0.7632, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "grad_norm": 0.13942602276802063, | |
| "learning_rate": 9.414737964294636e-05, | |
| "loss": 0.7712, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 10.909090909090908, | |
| "eval_loss": 0.8747226595878601, | |
| "eval_runtime": 9.1211, | |
| "eval_samples_per_second": 2.631, | |
| "eval_steps_per_second": 2.631, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 11.272727272727273, | |
| "grad_norm": 0.12621885538101196, | |
| "learning_rate": 9.347474647526095e-05, | |
| "loss": 0.7481, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 11.636363636363637, | |
| "grad_norm": 0.13754956424236298, | |
| "learning_rate": 9.276821300802534e-05, | |
| "loss": 0.7603, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 0.17661471664905548, | |
| "learning_rate": 9.202833017478422e-05, | |
| "loss": 0.747, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_loss": 0.874152660369873, | |
| "eval_runtime": 9.1255, | |
| "eval_samples_per_second": 2.63, | |
| "eval_steps_per_second": 2.63, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 12.363636363636363, | |
| "grad_norm": 0.16321919858455658, | |
| "learning_rate": 9.125567491391476e-05, | |
| "loss": 0.7276, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 12.727272727272727, | |
| "grad_norm": 0.17005178332328796, | |
| "learning_rate": 9.045084971874738e-05, | |
| "loss": 0.7182, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 12.909090909090908, | |
| "eval_loss": 0.877945601940155, | |
| "eval_runtime": 9.1275, | |
| "eval_samples_per_second": 2.629, | |
| "eval_steps_per_second": 2.629, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 13.090909090909092, | |
| "grad_norm": 0.18697452545166016, | |
| "learning_rate": 8.961448216775954e-05, | |
| "loss": 0.7105, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 13.454545454545455, | |
| "grad_norm": 0.17018142342567444, | |
| "learning_rate": 8.874722443520899e-05, | |
| "loss": 0.7383, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 13.818181818181818, | |
| "grad_norm": 0.1911424696445465, | |
| "learning_rate": 8.784975278258783e-05, | |
| "loss": 0.7025, | |
| "step": 76 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_loss": 0.8850860595703125, | |
| "eval_runtime": 9.1224, | |
| "eval_samples_per_second": 2.631, | |
| "eval_steps_per_second": 2.631, | |
| "step": 77 | |
| }, | |
| { | |
| "epoch": 14.181818181818182, | |
| "grad_norm": 0.19358296692371368, | |
| "learning_rate": 8.692276703129421e-05, | |
| "loss": 0.6621, | |
| "step": 78 | |
| }, | |
| { | |
| "epoch": 14.545454545454545, | |
| "grad_norm": 0.23677457869052887, | |
| "learning_rate": 8.596699001693255e-05, | |
| "loss": 0.7115, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 14.909090909090908, | |
| "grad_norm": 0.26214173436164856, | |
| "learning_rate": 8.498316702566828e-05, | |
| "loss": 0.626, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 14.909090909090908, | |
| "eval_loss": 0.894670307636261, | |
| "eval_runtime": 9.1272, | |
| "eval_samples_per_second": 2.63, | |
| "eval_steps_per_second": 2.63, | |
| "step": 82 | |
| }, | |
| { | |
| "epoch": 15.272727272727273, | |
| "grad_norm": 0.22342948615550995, | |
| "learning_rate": 8.397206521307584e-05, | |
| "loss": 0.6646, | |
| "step": 84 | |
| }, | |
| { | |
| "epoch": 15.636363636363637, | |
| "grad_norm": 0.26301610469818115, | |
| "learning_rate": 8.293447300593402e-05, | |
| "loss": 0.6354, | |
| "step": 86 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 0.27089810371398926, | |
| "learning_rate": 8.18711994874345e-05, | |
| "loss": 0.633, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_loss": 0.9076154828071594, | |
| "eval_runtime": 9.1276, | |
| "eval_samples_per_second": 2.629, | |
| "eval_steps_per_second": 2.629, | |
| "step": 88 | |
| }, | |
| { | |
| "epoch": 16.363636363636363, | |
| "grad_norm": 0.25885578989982605, | |
| "learning_rate": 8.07830737662829e-05, | |
| "loss": 0.6663, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 16.727272727272727, | |
| "grad_norm": 0.4284859895706177, | |
| "learning_rate": 7.967094433018508e-05, | |
| "loss": 0.583, | |
| "step": 92 | |
| }, | |
| { | |
| "epoch": 16.90909090909091, | |
| "eval_loss": 0.9299036860466003, | |
| "eval_runtime": 9.1252, | |
| "eval_samples_per_second": 2.63, | |
| "eval_steps_per_second": 2.63, | |
| "step": 93 | |
| }, | |
| { | |
| "epoch": 17.09090909090909, | |
| "grad_norm": 0.3989965617656708, | |
| "learning_rate": 7.85356783842216e-05, | |
| "loss": 0.5729, | |
| "step": 94 | |
| }, | |
| { | |
| "epoch": 17.454545454545453, | |
| "grad_norm": 0.3654530644416809, | |
| "learning_rate": 7.737816117462752e-05, | |
| "loss": 0.5678, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 17.818181818181817, | |
| "grad_norm": 0.44815194606781006, | |
| "learning_rate": 7.619929529850397e-05, | |
| "loss": 0.5357, | |
| "step": 98 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_loss": 0.9466567039489746, | |
| "eval_runtime": 9.1232, | |
| "eval_samples_per_second": 2.631, | |
| "eval_steps_per_second": 2.631, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 18.181818181818183, | |
| "grad_norm": 0.5155482888221741, | |
| "learning_rate": 7.500000000000001e-05, | |
| "loss": 0.5782, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 18.545454545454547, | |
| "grad_norm": 0.6486256718635559, | |
| "learning_rate": 7.378121045351378e-05, | |
| "loss": 0.5235, | |
| "step": 102 | |
| }, | |
| { | |
| "epoch": 18.90909090909091, | |
| "grad_norm": 0.49874356389045715, | |
| "learning_rate": 7.254387703447154e-05, | |
| "loss": 0.5232, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 18.90909090909091, | |
| "eval_loss": 0.9673776030540466, | |
| "eval_runtime": 9.1256, | |
| "eval_samples_per_second": 2.63, | |
| "eval_steps_per_second": 2.63, | |
| "step": 104 | |
| }, | |
| { | |
| "epoch": 18.90909090909091, | |
| "step": 104, | |
| "total_flos": 5.448936376080794e+16, | |
| "train_loss": 0.8463961791533691, | |
| "train_runtime": 2327.6689, | |
| "train_samples_per_second": 1.89, | |
| "train_steps_per_second": 0.107 | |
| } | |
| ], | |
| "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": 5.448936376080794e+16, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |