| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 4.0, | |
| "eval_steps": 500, | |
| "global_step": 124, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.16129032258064516, | |
| "grad_norm": 1.3462869408886267, | |
| "learning_rate": 1.923076923076923e-05, | |
| "loss": 0.1869, | |
| "mean_token_accuracy": 0.9672520697116852, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.3225806451612903, | |
| "grad_norm": 0.16032580673129507, | |
| "learning_rate": 3.846153846153846e-05, | |
| "loss": 0.0853, | |
| "mean_token_accuracy": 0.9812823057174682, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.4838709677419355, | |
| "grad_norm": 0.10052593361659107, | |
| "learning_rate": 4.996396285840363e-05, | |
| "loss": 0.0619, | |
| "mean_token_accuracy": 0.986013001203537, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.6451612903225806, | |
| "grad_norm": 0.07255406696478843, | |
| "learning_rate": 4.955986957912985e-05, | |
| "loss": 0.0584, | |
| "mean_token_accuracy": 0.986441147327423, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.8064516129032258, | |
| "grad_norm": 0.08920993926011508, | |
| "learning_rate": 4.87147425228744e-05, | |
| "loss": 0.052, | |
| "mean_token_accuracy": 0.9883647918701172, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.967741935483871, | |
| "grad_norm": 0.06573996870091518, | |
| "learning_rate": 4.7445477962587224e-05, | |
| "loss": 0.0524, | |
| "mean_token_accuracy": 0.9878623425960541, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 1.129032258064516, | |
| "grad_norm": 0.056610436447897504, | |
| "learning_rate": 4.5777451775537437e-05, | |
| "loss": 0.0441, | |
| "mean_token_accuracy": 0.9883042335510254, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 1.2903225806451613, | |
| "grad_norm": 0.05793088867879483, | |
| "learning_rate": 4.37440121139768e-05, | |
| "loss": 0.0449, | |
| "mean_token_accuracy": 0.9885231554508209, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.4516129032258065, | |
| "grad_norm": 0.051892819891969894, | |
| "learning_rate": 4.138581268943275e-05, | |
| "loss": 0.0392, | |
| "mean_token_accuracy": 0.9899458229541779, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 1.6129032258064515, | |
| "grad_norm": 0.0646130897394382, | |
| "learning_rate": 3.875e-05, | |
| "loss": 0.0421, | |
| "mean_token_accuracy": 0.9895950436592102, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.7741935483870968, | |
| "grad_norm": 0.04950171648116134, | |
| "learning_rate": 3.5889270750056945e-05, | |
| "loss": 0.031, | |
| "mean_token_accuracy": 0.9919082045555114, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 1.935483870967742, | |
| "grad_norm": 0.0606413837545455, | |
| "learning_rate": 3.286081830702121e-05, | |
| "loss": 0.0347, | |
| "mean_token_accuracy": 0.9910365581512451, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 2.096774193548387, | |
| "grad_norm": 0.07623335467494033, | |
| "learning_rate": 2.9725189258195623e-05, | |
| "loss": 0.03, | |
| "mean_token_accuracy": 0.9921543657779693, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 2.258064516129032, | |
| "grad_norm": 0.04875134102847451, | |
| "learning_rate": 2.6545072928086673e-05, | |
| "loss": 0.0251, | |
| "mean_token_accuracy": 0.9926283597946167, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 2.4193548387096775, | |
| "grad_norm": 0.059635109510899856, | |
| "learning_rate": 2.3384048056871306e-05, | |
| "loss": 0.0233, | |
| "mean_token_accuracy": 0.9933796107769013, | |
| "step": 75 | |
| }, | |
| { | |
| "epoch": 2.5806451612903225, | |
| "grad_norm": 0.06668699113879667, | |
| "learning_rate": 2.0305311697146985e-05, | |
| "loss": 0.0235, | |
| "mean_token_accuracy": 0.993266487121582, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 2.741935483870968, | |
| "grad_norm": 0.06406854664489343, | |
| "learning_rate": 1.7370415741602346e-05, | |
| "loss": 0.0255, | |
| "mean_token_accuracy": 0.9926106989383697, | |
| "step": 85 | |
| }, | |
| { | |
| "epoch": 2.903225806451613, | |
| "grad_norm": 0.06289941290219148, | |
| "learning_rate": 1.463803634168423e-05, | |
| "loss": 0.0229, | |
| "mean_token_accuracy": 0.9931235194206238, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 3.064516129032258, | |
| "grad_norm": 0.05401269074344149, | |
| "learning_rate": 1.2162800819760516e-05, | |
| "loss": 0.0218, | |
| "mean_token_accuracy": 0.9935337245464325, | |
| "step": 95 | |
| }, | |
| { | |
| "epoch": 3.225806451612903, | |
| "grad_norm": 0.05694790107472112, | |
| "learning_rate": 9.994195527835116e-06, | |
| "loss": 0.0125, | |
| "mean_token_accuracy": 0.9963877737522125, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 3.3870967741935485, | |
| "grad_norm": 0.07090477863149891, | |
| "learning_rate": 8.175576487540416e-06, | |
| "loss": 0.0123, | |
| "mean_token_accuracy": 0.99611736536026, | |
| "step": 105 | |
| }, | |
| { | |
| "epoch": 3.5483870967741935, | |
| "grad_norm": 0.05390988555062966, | |
| "learning_rate": 6.743302591269457e-06, | |
| "loss": 0.013, | |
| "mean_token_accuracy": 0.9960679054260254, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 3.709677419354839, | |
| "grad_norm": 0.07736825400243297, | |
| "learning_rate": 5.726008693996276e-06, | |
| "loss": 0.0138, | |
| "mean_token_accuracy": 0.9958937764167786, | |
| "step": 115 | |
| }, | |
| { | |
| "epoch": 3.870967741935484, | |
| "grad_norm": 0.05845945058343013, | |
| "learning_rate": 5.1440331285566846e-06, | |
| "loss": 0.0135, | |
| "mean_token_accuracy": 0.9959296882152557, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "mean_token_accuracy": 0.9968161135911942, | |
| "step": 124, | |
| "total_flos": 57119042306048.0, | |
| "train_loss": 0.03955811998175998, | |
| "train_runtime": 751.6795, | |
| "train_samples_per_second": 2.629, | |
| "train_steps_per_second": 0.165 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 124, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 1000, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 57119042306048.0, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |