| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 30.0, | |
| "eval_steps": 500, | |
| "global_step": 24210, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 6.501698017120361, | |
| "learning_rate": 9.66914498141264e-06, | |
| "loss": 1.2014, | |
| "step": 807 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.3661971688270569, | |
| "eval_loss": 1.183009147644043, | |
| "eval_runtime": 5.7015, | |
| "eval_samples_per_second": 12.453, | |
| "eval_steps_per_second": 12.453, | |
| "step": 807 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 11.675583839416504, | |
| "learning_rate": 9.335811648079307e-06, | |
| "loss": 1.0915, | |
| "step": 1614 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.3239436745643616, | |
| "eval_loss": 1.5119783878326416, | |
| "eval_runtime": 5.7636, | |
| "eval_samples_per_second": 12.319, | |
| "eval_steps_per_second": 12.319, | |
| "step": 1614 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 16.97098731994629, | |
| "learning_rate": 9.002478314745973e-06, | |
| "loss": 1.1433, | |
| "step": 2421 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.4084506928920746, | |
| "eval_loss": 1.5698989629745483, | |
| "eval_runtime": 5.8045, | |
| "eval_samples_per_second": 12.232, | |
| "eval_steps_per_second": 12.232, | |
| "step": 2421 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 127.35404205322266, | |
| "learning_rate": 8.669558033870302e-06, | |
| "loss": 1.2819, | |
| "step": 3228 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.47887325286865234, | |
| "eval_loss": 1.737151861190796, | |
| "eval_runtime": 5.8167, | |
| "eval_samples_per_second": 12.206, | |
| "eval_steps_per_second": 12.206, | |
| "step": 3228 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "grad_norm": 0.40140867233276367, | |
| "learning_rate": 8.336224700536968e-06, | |
| "loss": 1.2718, | |
| "step": 4035 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.4647887349128723, | |
| "eval_loss": 2.216925859451294, | |
| "eval_runtime": 5.8112, | |
| "eval_samples_per_second": 12.218, | |
| "eval_steps_per_second": 12.218, | |
| "step": 4035 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "grad_norm": 30.236343383789062, | |
| "learning_rate": 8.003304419661297e-06, | |
| "loss": 1.4535, | |
| "step": 4842 | |
| }, | |
| { | |
| "epoch": 6.0, | |
| "eval_accuracy": 0.577464759349823, | |
| "eval_loss": 1.7295811176300049, | |
| "eval_runtime": 5.8523, | |
| "eval_samples_per_second": 12.132, | |
| "eval_steps_per_second": 12.132, | |
| "step": 4842 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "grad_norm": 0.2374914586544037, | |
| "learning_rate": 7.670384138785627e-06, | |
| "loss": 1.3433, | |
| "step": 5649 | |
| }, | |
| { | |
| "epoch": 7.0, | |
| "eval_accuracy": 0.5492957830429077, | |
| "eval_loss": 2.268446683883667, | |
| "eval_runtime": 5.8095, | |
| "eval_samples_per_second": 12.221, | |
| "eval_steps_per_second": 12.221, | |
| "step": 5649 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "grad_norm": 32.27313232421875, | |
| "learning_rate": 7.3370508054522925e-06, | |
| "loss": 1.4086, | |
| "step": 6456 | |
| }, | |
| { | |
| "epoch": 8.0, | |
| "eval_accuracy": 0.6478873491287231, | |
| "eval_loss": 1.8598684072494507, | |
| "eval_runtime": 5.7903, | |
| "eval_samples_per_second": 12.262, | |
| "eval_steps_per_second": 12.262, | |
| "step": 6456 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "grad_norm": 42.521121978759766, | |
| "learning_rate": 7.003717472118959e-06, | |
| "loss": 1.3923, | |
| "step": 7263 | |
| }, | |
| { | |
| "epoch": 9.0, | |
| "eval_accuracy": 0.6197183132171631, | |
| "eval_loss": 1.9419935941696167, | |
| "eval_runtime": 5.8464, | |
| "eval_samples_per_second": 12.144, | |
| "eval_steps_per_second": 12.144, | |
| "step": 7263 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "grad_norm": 0.37923476099967957, | |
| "learning_rate": 6.670384138785626e-06, | |
| "loss": 1.3353, | |
| "step": 8070 | |
| }, | |
| { | |
| "epoch": 10.0, | |
| "eval_accuracy": 0.577464759349823, | |
| "eval_loss": 2.2150189876556396, | |
| "eval_runtime": 5.8548, | |
| "eval_samples_per_second": 12.127, | |
| "eval_steps_per_second": 12.127, | |
| "step": 8070 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "grad_norm": 19.937904357910156, | |
| "learning_rate": 6.337463857909954e-06, | |
| "loss": 1.367, | |
| "step": 8877 | |
| }, | |
| { | |
| "epoch": 11.0, | |
| "eval_accuracy": 0.6338028311729431, | |
| "eval_loss": 1.9825525283813477, | |
| "eval_runtime": 5.9069, | |
| "eval_samples_per_second": 12.02, | |
| "eval_steps_per_second": 12.02, | |
| "step": 8877 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "grad_norm": 454.2120666503906, | |
| "learning_rate": 6.004543577034284e-06, | |
| "loss": 1.1848, | |
| "step": 9684 | |
| }, | |
| { | |
| "epoch": 12.0, | |
| "eval_accuracy": 0.6478873491287231, | |
| "eval_loss": 1.9545217752456665, | |
| "eval_runtime": 5.8618, | |
| "eval_samples_per_second": 12.112, | |
| "eval_steps_per_second": 12.112, | |
| "step": 9684 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "grad_norm": 0.09417306631803513, | |
| "learning_rate": 5.671210243700951e-06, | |
| "loss": 1.1355, | |
| "step": 10491 | |
| }, | |
| { | |
| "epoch": 13.0, | |
| "eval_accuracy": 0.6619718074798584, | |
| "eval_loss": 1.9863765239715576, | |
| "eval_runtime": 5.8578, | |
| "eval_samples_per_second": 12.12, | |
| "eval_steps_per_second": 12.12, | |
| "step": 10491 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "grad_norm": 33.132957458496094, | |
| "learning_rate": 5.337876910367618e-06, | |
| "loss": 1.1549, | |
| "step": 11298 | |
| }, | |
| { | |
| "epoch": 14.0, | |
| "eval_accuracy": 0.6338028311729431, | |
| "eval_loss": 1.9428231716156006, | |
| "eval_runtime": 5.8665, | |
| "eval_samples_per_second": 12.103, | |
| "eval_steps_per_second": 12.103, | |
| "step": 11298 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "grad_norm": 0.12686021625995636, | |
| "learning_rate": 5.004543577034285e-06, | |
| "loss": 1.0505, | |
| "step": 12105 | |
| }, | |
| { | |
| "epoch": 15.0, | |
| "eval_accuracy": 0.6901408433914185, | |
| "eval_loss": 1.9100552797317505, | |
| "eval_runtime": 5.8853, | |
| "eval_samples_per_second": 12.064, | |
| "eval_steps_per_second": 12.064, | |
| "step": 12105 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "grad_norm": 1043.1195068359375, | |
| "learning_rate": 4.6712102437009505e-06, | |
| "loss": 1.0442, | |
| "step": 12912 | |
| }, | |
| { | |
| "epoch": 16.0, | |
| "eval_accuracy": 0.6478873491287231, | |
| "eval_loss": 2.1706087589263916, | |
| "eval_runtime": 5.8878, | |
| "eval_samples_per_second": 12.059, | |
| "eval_steps_per_second": 12.059, | |
| "step": 12912 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "grad_norm": 26.500118255615234, | |
| "learning_rate": 4.338289962825279e-06, | |
| "loss": 0.9922, | |
| "step": 13719 | |
| }, | |
| { | |
| "epoch": 17.0, | |
| "eval_accuracy": 0.6197183132171631, | |
| "eval_loss": 2.462040901184082, | |
| "eval_runtime": 5.8428, | |
| "eval_samples_per_second": 12.152, | |
| "eval_steps_per_second": 12.152, | |
| "step": 13719 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "grad_norm": 162.029052734375, | |
| "learning_rate": 4.004956629491946e-06, | |
| "loss": 0.8698, | |
| "step": 14526 | |
| }, | |
| { | |
| "epoch": 18.0, | |
| "eval_accuracy": 0.6619718074798584, | |
| "eval_loss": 2.142850637435913, | |
| "eval_runtime": 5.9312, | |
| "eval_samples_per_second": 11.971, | |
| "eval_steps_per_second": 11.971, | |
| "step": 14526 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "grad_norm": 0.0350579209625721, | |
| "learning_rate": 3.6716232961586124e-06, | |
| "loss": 0.8202, | |
| "step": 15333 | |
| }, | |
| { | |
| "epoch": 19.0, | |
| "eval_accuracy": 0.6197183132171631, | |
| "eval_loss": 2.372546911239624, | |
| "eval_runtime": 5.8453, | |
| "eval_samples_per_second": 12.147, | |
| "eval_steps_per_second": 12.147, | |
| "step": 15333 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "grad_norm": 0.042548421770334244, | |
| "learning_rate": 3.3387030152829415e-06, | |
| "loss": 0.8612, | |
| "step": 16140 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "eval_accuracy": 0.6619718074798584, | |
| "eval_loss": 2.1631431579589844, | |
| "eval_runtime": 5.8804, | |
| "eval_samples_per_second": 12.074, | |
| "eval_steps_per_second": 12.074, | |
| "step": 16140 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "grad_norm": 0.09584546834230423, | |
| "learning_rate": 3.0053696819496083e-06, | |
| "loss": 0.8197, | |
| "step": 16947 | |
| }, | |
| { | |
| "epoch": 21.0, | |
| "eval_accuracy": 0.6338028311729431, | |
| "eval_loss": 2.393216848373413, | |
| "eval_runtime": 5.8299, | |
| "eval_samples_per_second": 12.179, | |
| "eval_steps_per_second": 12.179, | |
| "step": 16947 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "grad_norm": 0.0448361337184906, | |
| "learning_rate": 2.672036348616274e-06, | |
| "loss": 0.7858, | |
| "step": 17754 | |
| }, | |
| { | |
| "epoch": 22.0, | |
| "eval_accuracy": 0.6478873491287231, | |
| "eval_loss": 2.25321364402771, | |
| "eval_runtime": 5.8111, | |
| "eval_samples_per_second": 12.218, | |
| "eval_steps_per_second": 12.218, | |
| "step": 17754 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "grad_norm": 0.057099100202322006, | |
| "learning_rate": 2.3391160677406034e-06, | |
| "loss": 0.7717, | |
| "step": 18561 | |
| }, | |
| { | |
| "epoch": 23.0, | |
| "eval_accuracy": 0.5633803009986877, | |
| "eval_loss": 2.813150644302368, | |
| "eval_runtime": 5.8476, | |
| "eval_samples_per_second": 12.142, | |
| "eval_steps_per_second": 12.142, | |
| "step": 18561 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "grad_norm": 202.8040008544922, | |
| "learning_rate": 2.00578273440727e-06, | |
| "loss": 0.6282, | |
| "step": 19368 | |
| }, | |
| { | |
| "epoch": 24.0, | |
| "eval_accuracy": 0.6197183132171631, | |
| "eval_loss": 2.549309492111206, | |
| "eval_runtime": 5.839, | |
| "eval_samples_per_second": 12.16, | |
| "eval_steps_per_second": 12.16, | |
| "step": 19368 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "grad_norm": 0.06747495383024216, | |
| "learning_rate": 1.6728624535315987e-06, | |
| "loss": 0.7394, | |
| "step": 20175 | |
| }, | |
| { | |
| "epoch": 25.0, | |
| "eval_accuracy": 0.6619718074798584, | |
| "eval_loss": 2.3195266723632812, | |
| "eval_runtime": 5.8894, | |
| "eval_samples_per_second": 12.056, | |
| "eval_steps_per_second": 12.056, | |
| "step": 20175 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "grad_norm": 4.30875301361084, | |
| "learning_rate": 1.3395291201982652e-06, | |
| "loss": 0.5895, | |
| "step": 20982 | |
| }, | |
| { | |
| "epoch": 26.0, | |
| "eval_accuracy": 0.6619718074798584, | |
| "eval_loss": 2.43306565284729, | |
| "eval_runtime": 5.8283, | |
| "eval_samples_per_second": 12.182, | |
| "eval_steps_per_second": 12.182, | |
| "step": 20982 | |
| }, | |
| { | |
| "epoch": 27.0, | |
| "grad_norm": 0.05214543268084526, | |
| "learning_rate": 1.006195786864932e-06, | |
| "loss": 0.5854, | |
| "step": 21789 | |
| }, | |
| { | |
| "epoch": 27.0, | |
| "eval_accuracy": 0.6760563254356384, | |
| "eval_loss": 2.428130626678467, | |
| "eval_runtime": 5.8154, | |
| "eval_samples_per_second": 12.209, | |
| "eval_steps_per_second": 12.209, | |
| "step": 21789 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "grad_norm": 0.1047598198056221, | |
| "learning_rate": 6.732755059892606e-07, | |
| "loss": 0.6911, | |
| "step": 22596 | |
| }, | |
| { | |
| "epoch": 28.0, | |
| "eval_accuracy": 0.6619718074798584, | |
| "eval_loss": 2.499279737472534, | |
| "eval_runtime": 5.8567, | |
| "eval_samples_per_second": 12.123, | |
| "eval_steps_per_second": 12.123, | |
| "step": 22596 | |
| }, | |
| { | |
| "epoch": 29.0, | |
| "grad_norm": 92.8244857788086, | |
| "learning_rate": 3.3994217265592734e-07, | |
| "loss": 0.5502, | |
| "step": 23403 | |
| }, | |
| { | |
| "epoch": 29.0, | |
| "eval_accuracy": 0.6338028311729431, | |
| "eval_loss": 2.64582896232605, | |
| "eval_runtime": 5.8325, | |
| "eval_samples_per_second": 12.173, | |
| "eval_steps_per_second": 12.173, | |
| "step": 23403 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "grad_norm": 0.047546908259391785, | |
| "learning_rate": 6.60883932259397e-09, | |
| "loss": 0.584, | |
| "step": 24210 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "eval_accuracy": 0.6338028311729431, | |
| "eval_loss": 2.586568593978882, | |
| "eval_runtime": 5.8563, | |
| "eval_samples_per_second": 12.124, | |
| "eval_steps_per_second": 12.124, | |
| "step": 24210 | |
| }, | |
| { | |
| "epoch": 30.0, | |
| "step": 24210, | |
| "total_flos": 2.2475408052256202e+18, | |
| "train_loss": 1.0049422518175921, | |
| "train_runtime": 3971.9322, | |
| "train_samples_per_second": 6.095, | |
| "train_steps_per_second": 6.095 | |
| } | |
| ], | |
| "logging_steps": 12, | |
| "max_steps": 24210, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 30, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": false, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 2.2475408052256202e+18, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |