| { | |
| "best_metric": 10.954463005065918, | |
| "best_model_checkpoint": "miner_id_24/checkpoint-75", | |
| "epoch": 0.645682001614205, | |
| "eval_steps": 25, | |
| "global_step": 75, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.008609093354856066, | |
| "grad_norm": 10.340320587158203, | |
| "learning_rate": 3.3333333333333335e-05, | |
| "loss": 355.0044, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.008609093354856066, | |
| "eval_loss": 11.096755027770996, | |
| "eval_runtime": 0.6051, | |
| "eval_samples_per_second": 82.636, | |
| "eval_steps_per_second": 21.485, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.017218186709712133, | |
| "grad_norm": 10.717597961425781, | |
| "learning_rate": 6.666666666666667e-05, | |
| "loss": 354.975, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.0258272800645682, | |
| "grad_norm": 13.101455688476562, | |
| "learning_rate": 0.0001, | |
| "loss": 355.0461, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.034436373419424265, | |
| "grad_norm": 11.34632396697998, | |
| "learning_rate": 9.99571699711836e-05, | |
| "loss": 354.9361, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.04304546677428033, | |
| "grad_norm": 12.115002632141113, | |
| "learning_rate": 9.982876141412856e-05, | |
| "loss": 354.7938, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.0516545601291364, | |
| "grad_norm": 11.92343807220459, | |
| "learning_rate": 9.961501876182148e-05, | |
| "loss": 354.764, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.060263653483992465, | |
| "grad_norm": 12.599183082580566, | |
| "learning_rate": 9.931634888554937e-05, | |
| "loss": 354.6257, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.06887274683884853, | |
| "grad_norm": 12.820220947265625, | |
| "learning_rate": 9.893332032039701e-05, | |
| "loss": 354.4836, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.0774818401937046, | |
| "grad_norm": 14.551854133605957, | |
| "learning_rate": 9.846666218300807e-05, | |
| "loss": 354.3333, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.08609093354856066, | |
| "grad_norm": 16.782869338989258, | |
| "learning_rate": 9.791726278367022e-05, | |
| "loss": 354.0781, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.09470002690341674, | |
| "grad_norm": 21.504772186279297, | |
| "learning_rate": 9.728616793536588e-05, | |
| "loss": 353.7289, | |
| "step": 11 | |
| }, | |
| { | |
| "epoch": 0.1033091202582728, | |
| "grad_norm": 24.859169006347656, | |
| "learning_rate": 9.657457896300791e-05, | |
| "loss": 353.4381, | |
| "step": 12 | |
| }, | |
| { | |
| "epoch": 0.11191821361312887, | |
| "grad_norm": 18.659643173217773, | |
| "learning_rate": 9.578385041664925e-05, | |
| "loss": 353.6805, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.12052730696798493, | |
| "grad_norm": 10.551276206970215, | |
| "learning_rate": 9.491548749301997e-05, | |
| "loss": 353.9667, | |
| "step": 14 | |
| }, | |
| { | |
| "epoch": 0.129136400322841, | |
| "grad_norm": 11.915667533874512, | |
| "learning_rate": 9.397114317029975e-05, | |
| "loss": 353.8238, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.13774549367769706, | |
| "grad_norm": 11.559496879577637, | |
| "learning_rate": 9.295261506157986e-05, | |
| "loss": 353.7797, | |
| "step": 16 | |
| }, | |
| { | |
| "epoch": 0.14635458703255314, | |
| "grad_norm": 10.912751197814941, | |
| "learning_rate": 9.186184199300464e-05, | |
| "loss": 353.7299, | |
| "step": 17 | |
| }, | |
| { | |
| "epoch": 0.1549636803874092, | |
| "grad_norm": 11.342243194580078, | |
| "learning_rate": 9.070090031310558e-05, | |
| "loss": 353.5555, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 0.16357277374226525, | |
| "grad_norm": 11.279723167419434, | |
| "learning_rate": 8.947199994035401e-05, | |
| "loss": 353.4576, | |
| "step": 19 | |
| }, | |
| { | |
| "epoch": 0.17218186709712133, | |
| "grad_norm": 11.607893943786621, | |
| "learning_rate": 8.817748015645558e-05, | |
| "loss": 353.2998, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.1807909604519774, | |
| "grad_norm": 12.31462287902832, | |
| "learning_rate": 8.681980515339464e-05, | |
| "loss": 353.1127, | |
| "step": 21 | |
| }, | |
| { | |
| "epoch": 0.18940005380683347, | |
| "grad_norm": 13.772215843200684, | |
| "learning_rate": 8.540155934270471e-05, | |
| "loss": 352.8197, | |
| "step": 22 | |
| }, | |
| { | |
| "epoch": 0.19800914716168955, | |
| "grad_norm": 16.35397720336914, | |
| "learning_rate": 8.392544243589427e-05, | |
| "loss": 352.1693, | |
| "step": 23 | |
| }, | |
| { | |
| "epoch": 0.2066182405165456, | |
| "grad_norm": 19.223989486694336, | |
| "learning_rate": 8.239426430539243e-05, | |
| "loss": 351.5058, | |
| "step": 24 | |
| }, | |
| { | |
| "epoch": 0.21522733387140167, | |
| "grad_norm": 22.75308609008789, | |
| "learning_rate": 8.081093963579707e-05, | |
| "loss": 350.7887, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.21522733387140167, | |
| "eval_loss": 11.010751724243164, | |
| "eval_runtime": 0.0775, | |
| "eval_samples_per_second": 645.508, | |
| "eval_steps_per_second": 167.832, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.22383642722625774, | |
| "grad_norm": 8.562657356262207, | |
| "learning_rate": 7.917848237560709e-05, | |
| "loss": 353.2953, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.2324455205811138, | |
| "grad_norm": 9.411365509033203, | |
| "learning_rate": 7.75e-05, | |
| "loss": 352.9915, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 0.24105461393596986, | |
| "grad_norm": 10.57524299621582, | |
| "learning_rate": 7.577868759557654e-05, | |
| "loss": 352.6193, | |
| "step": 28 | |
| }, | |
| { | |
| "epoch": 0.24966370729082593, | |
| "grad_norm": 9.406920433044434, | |
| "learning_rate": 7.401782177833148e-05, | |
| "loss": 352.7668, | |
| "step": 29 | |
| }, | |
| { | |
| "epoch": 0.258272800645682, | |
| "grad_norm": 9.367744445800781, | |
| "learning_rate": 7.222075445642904e-05, | |
| "loss": 352.7045, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.2668818940005381, | |
| "grad_norm": 9.058300971984863, | |
| "learning_rate": 7.03909064496551e-05, | |
| "loss": 352.7024, | |
| "step": 31 | |
| }, | |
| { | |
| "epoch": 0.2754909873553941, | |
| "grad_norm": 9.112210273742676, | |
| "learning_rate": 6.853176097769229e-05, | |
| "loss": 352.6205, | |
| "step": 32 | |
| }, | |
| { | |
| "epoch": 0.2841000807102502, | |
| "grad_norm": 9.437152862548828, | |
| "learning_rate": 6.664685702961344e-05, | |
| "loss": 352.4774, | |
| "step": 33 | |
| }, | |
| { | |
| "epoch": 0.29270917406510627, | |
| "grad_norm": 10.079466819763184, | |
| "learning_rate": 6.473978262721463e-05, | |
| "loss": 352.2588, | |
| "step": 34 | |
| }, | |
| { | |
| "epoch": 0.3013182674199623, | |
| "grad_norm": 11.751006126403809, | |
| "learning_rate": 6.281416799501188e-05, | |
| "loss": 351.5457, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.3099273607748184, | |
| "grad_norm": 14.352370262145996, | |
| "learning_rate": 6.087367864990233e-05, | |
| "loss": 350.6714, | |
| "step": 36 | |
| }, | |
| { | |
| "epoch": 0.31853645412967446, | |
| "grad_norm": 17.41858673095703, | |
| "learning_rate": 5.8922008423644624e-05, | |
| "loss": 349.8581, | |
| "step": 37 | |
| }, | |
| { | |
| "epoch": 0.3271455474845305, | |
| "grad_norm": 13.782036781311035, | |
| "learning_rate": 5.696287243144013e-05, | |
| "loss": 350.8174, | |
| "step": 38 | |
| }, | |
| { | |
| "epoch": 0.3357546408393866, | |
| "grad_norm": 8.684106826782227, | |
| "learning_rate": 5.500000000000001e-05, | |
| "loss": 352.3159, | |
| "step": 39 | |
| }, | |
| { | |
| "epoch": 0.34436373419424265, | |
| "grad_norm": 8.738736152648926, | |
| "learning_rate": 5.303712756855988e-05, | |
| "loss": 352.193, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.35297282754909876, | |
| "grad_norm": 8.415644645690918, | |
| "learning_rate": 5.107799157635538e-05, | |
| "loss": 352.1501, | |
| "step": 41 | |
| }, | |
| { | |
| "epoch": 0.3615819209039548, | |
| "grad_norm": 8.467161178588867, | |
| "learning_rate": 4.912632135009769e-05, | |
| "loss": 352.097, | |
| "step": 42 | |
| }, | |
| { | |
| "epoch": 0.37019101425881085, | |
| "grad_norm": 8.014203071594238, | |
| "learning_rate": 4.718583200498814e-05, | |
| "loss": 352.2586, | |
| "step": 43 | |
| }, | |
| { | |
| "epoch": 0.37880010761366695, | |
| "grad_norm": 8.003595352172852, | |
| "learning_rate": 4.526021737278538e-05, | |
| "loss": 352.1916, | |
| "step": 44 | |
| }, | |
| { | |
| "epoch": 0.387409200968523, | |
| "grad_norm": 8.26496410369873, | |
| "learning_rate": 4.3353142970386564e-05, | |
| "loss": 351.9877, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.3960182943233791, | |
| "grad_norm": 9.2656888961792, | |
| "learning_rate": 4.146823902230772e-05, | |
| "loss": 351.5651, | |
| "step": 46 | |
| }, | |
| { | |
| "epoch": 0.40462738767823514, | |
| "grad_norm": 10.289533615112305, | |
| "learning_rate": 3.960909355034491e-05, | |
| "loss": 351.0219, | |
| "step": 47 | |
| }, | |
| { | |
| "epoch": 0.4132364810330912, | |
| "grad_norm": 12.535999298095703, | |
| "learning_rate": 3.777924554357096e-05, | |
| "loss": 350.1457, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 0.4218455743879473, | |
| "grad_norm": 15.549736976623535, | |
| "learning_rate": 3.598217822166854e-05, | |
| "loss": 349.2452, | |
| "step": 49 | |
| }, | |
| { | |
| "epoch": 0.43045466774280333, | |
| "grad_norm": 19.12620735168457, | |
| "learning_rate": 3.422131240442349e-05, | |
| "loss": 348.492, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.43045466774280333, | |
| "eval_loss": 10.966361045837402, | |
| "eval_runtime": 0.0781, | |
| "eval_samples_per_second": 640.442, | |
| "eval_steps_per_second": 166.515, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.4390637610976594, | |
| "grad_norm": 7.249837875366211, | |
| "learning_rate": 3.250000000000001e-05, | |
| "loss": 352.3247, | |
| "step": 51 | |
| }, | |
| { | |
| "epoch": 0.4476728544525155, | |
| "grad_norm": 7.920052528381348, | |
| "learning_rate": 3.082151762439293e-05, | |
| "loss": 351.9609, | |
| "step": 52 | |
| }, | |
| { | |
| "epoch": 0.4562819478073715, | |
| "grad_norm": 8.327653884887695, | |
| "learning_rate": 2.9189060364202943e-05, | |
| "loss": 351.6255, | |
| "step": 53 | |
| }, | |
| { | |
| "epoch": 0.4648910411622276, | |
| "grad_norm": 8.325037002563477, | |
| "learning_rate": 2.760573569460757e-05, | |
| "loss": 351.5259, | |
| "step": 54 | |
| }, | |
| { | |
| "epoch": 0.47350013451708367, | |
| "grad_norm": 8.068108558654785, | |
| "learning_rate": 2.6074557564105727e-05, | |
| "loss": 351.6653, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.4821092278719397, | |
| "grad_norm": 7.1570658683776855, | |
| "learning_rate": 2.459844065729529e-05, | |
| "loss": 352.1607, | |
| "step": 56 | |
| }, | |
| { | |
| "epoch": 0.4907183212267958, | |
| "grad_norm": 7.453501224517822, | |
| "learning_rate": 2.3180194846605367e-05, | |
| "loss": 351.8083, | |
| "step": 57 | |
| }, | |
| { | |
| "epoch": 0.49932741458165186, | |
| "grad_norm": 7.861183166503906, | |
| "learning_rate": 2.1822519843544424e-05, | |
| "loss": 351.535, | |
| "step": 58 | |
| }, | |
| { | |
| "epoch": 0.5079365079365079, | |
| "grad_norm": 8.656818389892578, | |
| "learning_rate": 2.0528000059645997e-05, | |
| "loss": 351.2567, | |
| "step": 59 | |
| }, | |
| { | |
| "epoch": 0.516545601291364, | |
| "grad_norm": 10.01552677154541, | |
| "learning_rate": 1.9299099686894423e-05, | |
| "loss": 350.7032, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.5251546946462201, | |
| "grad_norm": 12.777168273925781, | |
| "learning_rate": 1.8138158006995364e-05, | |
| "loss": 349.638, | |
| "step": 61 | |
| }, | |
| { | |
| "epoch": 0.5337637880010762, | |
| "grad_norm": 16.425859451293945, | |
| "learning_rate": 1.7047384938420154e-05, | |
| "loss": 348.5255, | |
| "step": 62 | |
| }, | |
| { | |
| "epoch": 0.5423728813559322, | |
| "grad_norm": 12.349985122680664, | |
| "learning_rate": 1.602885682970026e-05, | |
| "loss": 350.1203, | |
| "step": 63 | |
| }, | |
| { | |
| "epoch": 0.5509819747107882, | |
| "grad_norm": 7.621433734893799, | |
| "learning_rate": 1.5084512506980026e-05, | |
| "loss": 351.9437, | |
| "step": 64 | |
| }, | |
| { | |
| "epoch": 0.5595910680656443, | |
| "grad_norm": 8.107879638671875, | |
| "learning_rate": 1.4216149583350754e-05, | |
| "loss": 351.5311, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.5682001614205004, | |
| "grad_norm": 7.896660327911377, | |
| "learning_rate": 1.3425421036992098e-05, | |
| "loss": 351.4348, | |
| "step": 66 | |
| }, | |
| { | |
| "epoch": 0.5768092547753565, | |
| "grad_norm": 8.302569389343262, | |
| "learning_rate": 1.2713832064634126e-05, | |
| "loss": 351.1709, | |
| "step": 67 | |
| }, | |
| { | |
| "epoch": 0.5854183481302125, | |
| "grad_norm": 8.209525108337402, | |
| "learning_rate": 1.2082737216329794e-05, | |
| "loss": 351.1588, | |
| "step": 68 | |
| }, | |
| { | |
| "epoch": 0.5940274414850686, | |
| "grad_norm": 7.311443328857422, | |
| "learning_rate": 1.1533337816991932e-05, | |
| "loss": 351.5625, | |
| "step": 69 | |
| }, | |
| { | |
| "epoch": 0.6026365348399246, | |
| "grad_norm": 7.658151626586914, | |
| "learning_rate": 1.1066679679603e-05, | |
| "loss": 351.509, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.6112456281947808, | |
| "grad_norm": 8.324136734008789, | |
| "learning_rate": 1.0683651114450641e-05, | |
| "loss": 351.2198, | |
| "step": 71 | |
| }, | |
| { | |
| "epoch": 0.6198547215496368, | |
| "grad_norm": 9.308731079101562, | |
| "learning_rate": 1.0384981238178534e-05, | |
| "loss": 350.7391, | |
| "step": 72 | |
| }, | |
| { | |
| "epoch": 0.6284638149044929, | |
| "grad_norm": 11.327068328857422, | |
| "learning_rate": 1.017123858587145e-05, | |
| "loss": 349.8664, | |
| "step": 73 | |
| }, | |
| { | |
| "epoch": 0.6370729082593489, | |
| "grad_norm": 15.463874816894531, | |
| "learning_rate": 1.00428300288164e-05, | |
| "loss": 348.5386, | |
| "step": 74 | |
| }, | |
| { | |
| "epoch": 0.645682001614205, | |
| "grad_norm": 19.38113784790039, | |
| "learning_rate": 1e-05, | |
| "loss": 347.904, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 0.645682001614205, | |
| "eval_loss": 10.954463005065918, | |
| "eval_runtime": 0.0787, | |
| "eval_samples_per_second": 635.445, | |
| "eval_steps_per_second": 165.216, | |
| "step": 75 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 75, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 25, | |
| "stateful_callbacks": { | |
| "EarlyStoppingCallback": { | |
| "args": { | |
| "early_stopping_patience": 1, | |
| "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": 6323739033600.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |