| { |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 1.0, |
| "eval_steps": 100, |
| "global_step": 169, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.029585798816568046, |
| "grad_norm": 2.296627017481654, |
| "learning_rate": 5.882352941176471e-06, |
| "loss": 1.1002, |
| "mean_token_accuracy": 0.7101602736353531, |
| "step": 5 |
| }, |
| { |
| "epoch": 0.05917159763313609, |
| "grad_norm": 1.596826543128912, |
| "learning_rate": 1.1764705882352942e-05, |
| "loss": 1.0327, |
| "mean_token_accuracy": 0.7202409034109546, |
| "step": 10 |
| }, |
| { |
| "epoch": 0.08875739644970414, |
| "grad_norm": 0.8465871173559054, |
| "learning_rate": 1.7647058823529414e-05, |
| "loss": 0.9516, |
| "mean_token_accuracy": 0.7326533431949145, |
| "step": 15 |
| }, |
| { |
| "epoch": 0.11834319526627218, |
| "grad_norm": 0.6442903923509904, |
| "learning_rate": 1.9980782984658682e-05, |
| "loss": 0.8804, |
| "mean_token_accuracy": 0.7473519507067801, |
| "step": 20 |
| }, |
| { |
| "epoch": 0.14792899408284024, |
| "grad_norm": 0.5525115864092071, |
| "learning_rate": 1.9863613034027224e-05, |
| "loss": 0.8536, |
| "mean_token_accuracy": 0.7518378164073345, |
| "step": 25 |
| }, |
| { |
| "epoch": 0.17751479289940827, |
| "grad_norm": 0.440737539098077, |
| "learning_rate": 1.9641197940012136e-05, |
| "loss": 0.8404, |
| "mean_token_accuracy": 0.7545693378006253, |
| "step": 30 |
| }, |
| { |
| "epoch": 0.20710059171597633, |
| "grad_norm": 0.39407081780499364, |
| "learning_rate": 1.9315910880512792e-05, |
| "loss": 0.8212, |
| "mean_token_accuracy": 0.7580647532253553, |
| "step": 35 |
| }, |
| { |
| "epoch": 0.23668639053254437, |
| "grad_norm": 0.38687180734685467, |
| "learning_rate": 1.8891222681391853e-05, |
| "loss": 0.805, |
| "mean_token_accuracy": 0.7620870121182095, |
| "step": 40 |
| }, |
| { |
| "epoch": 0.26627218934911245, |
| "grad_norm": 0.42101277394154013, |
| "learning_rate": 1.8371664782625287e-05, |
| "loss": 0.7984, |
| "mean_token_accuracy": 0.7634557897772813, |
| "step": 45 |
| }, |
| { |
| "epoch": 0.2958579881656805, |
| "grad_norm": 0.3927099311549737, |
| "learning_rate": 1.7762780887657576e-05, |
| "loss": 0.7937, |
| "mean_token_accuracy": 0.7643062766147167, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.3254437869822485, |
| "grad_norm": 0.37357593512699977, |
| "learning_rate": 1.7071067811865477e-05, |
| "loss": 0.7894, |
| "mean_token_accuracy": 0.7648615731810274, |
| "step": 55 |
| }, |
| { |
| "epoch": 0.35502958579881655, |
| "grad_norm": 0.3710271041793924, |
| "learning_rate": 1.6303906161279554e-05, |
| "loss": 0.7998, |
| "mean_token_accuracy": 0.7620069255667766, |
| "step": 60 |
| }, |
| { |
| "epoch": 0.38461538461538464, |
| "grad_norm": 0.3955399102575344, |
| "learning_rate": 1.5469481581224274e-05, |
| "loss": 0.7868, |
| "mean_token_accuracy": 0.7647448994635244, |
| "step": 65 |
| }, |
| { |
| "epoch": 0.41420118343195267, |
| "grad_norm": 0.3741661269413078, |
| "learning_rate": 1.4576697415156818e-05, |
| "loss": 0.7782, |
| "mean_token_accuracy": 0.767462910291248, |
| "step": 70 |
| }, |
| { |
| "epoch": 0.4437869822485207, |
| "grad_norm": 0.33831000469886485, |
| "learning_rate": 1.3635079705638298e-05, |
| "loss": 0.777, |
| "mean_token_accuracy": 0.767488969039178, |
| "step": 75 |
| }, |
| { |
| "epoch": 0.47337278106508873, |
| "grad_norm": 0.3449126575086108, |
| "learning_rate": 1.2654675551080724e-05, |
| "loss": 0.7782, |
| "mean_token_accuracy": 0.7665897979720638, |
| "step": 80 |
| }, |
| { |
| "epoch": 0.5029585798816568, |
| "grad_norm": 0.33391154101530296, |
| "learning_rate": 1.164594590280734e-05, |
| "loss": 0.7785, |
| "mean_token_accuracy": 0.7669594110076028, |
| "step": 85 |
| }, |
| { |
| "epoch": 0.5325443786982249, |
| "grad_norm": 0.3483118104625951, |
| "learning_rate": 1.0619653946285948e-05, |
| "loss": 0.7723, |
| "mean_token_accuracy": 0.7682521535948184, |
| "step": 90 |
| }, |
| { |
| "epoch": 0.5621301775147929, |
| "grad_norm": 0.3441520918506831, |
| "learning_rate": 9.586750257511868e-06, |
| "loss": 0.7668, |
| "mean_token_accuracy": 0.7698453409127722, |
| "step": 95 |
| }, |
| { |
| "epoch": 0.591715976331361, |
| "grad_norm": 0.34155263390953944, |
| "learning_rate": 8.558255959926533e-06, |
| "loss": 0.7582, |
| "mean_token_accuracy": 0.7714380710780233, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.591715976331361, |
| "eval_loss": 0.788523256778717, |
| "eval_mean_token_accuracy": 0.7482845227929594, |
| "eval_runtime": 3.1336, |
| "eval_samples_per_second": 41.166, |
| "eval_steps_per_second": 1.596, |
| "step": 100 |
| }, |
| { |
| "epoch": 0.621301775147929, |
| "grad_norm": 0.3526165065371352, |
| "learning_rate": 7.545145128592009e-06, |
| "loss": 0.759, |
| "mean_token_accuracy": 0.7714748062522142, |
| "step": 105 |
| }, |
| { |
| "epoch": 0.650887573964497, |
| "grad_norm": 0.3164819449613246, |
| "learning_rate": 6.558227696373617e-06, |
| "loss": 0.7775, |
| "mean_token_accuracy": 0.7665118577749565, |
| "step": 110 |
| }, |
| { |
| "epoch": 0.6804733727810651, |
| "grad_norm": 0.30863923303250107, |
| "learning_rate": 5.608034111526298e-06, |
| "loss": 0.7636, |
| "mean_token_accuracy": 0.7702796270096404, |
| "step": 115 |
| }, |
| { |
| "epoch": 0.7100591715976331, |
| "grad_norm": 0.2988751601328229, |
| "learning_rate": 4.704702977392914e-06, |
| "loss": 0.7511, |
| "mean_token_accuracy": 0.7742468493685697, |
| "step": 120 |
| }, |
| { |
| "epoch": 0.7396449704142012, |
| "grad_norm": 0.3150637140177592, |
| "learning_rate": 3.857872873103322e-06, |
| "loss": 0.7594, |
| "mean_token_accuracy": 0.7718443409575408, |
| "step": 125 |
| }, |
| { |
| "epoch": 0.7692307692307693, |
| "grad_norm": 0.32049928588878124, |
| "learning_rate": 3.0765795095517026e-06, |
| "loss": 0.7624, |
| "mean_token_accuracy": 0.7706326807866132, |
| "step": 130 |
| }, |
| { |
| "epoch": 0.7988165680473372, |
| "grad_norm": 0.30263238367216605, |
| "learning_rate": 2.369159318001937e-06, |
| "loss": 0.7684, |
| "mean_token_accuracy": 0.7688668175201462, |
| "step": 135 |
| }, |
| { |
| "epoch": 0.8284023668639053, |
| "grad_norm": 0.29351086398490944, |
| "learning_rate": 1.743160500034443e-06, |
| "loss": 0.7596, |
| "mean_token_accuracy": 0.771702211626139, |
| "step": 140 |
| }, |
| { |
| "epoch": 0.8579881656804734, |
| "grad_norm": 0.31214293218264483, |
| "learning_rate": 1.2052624879351105e-06, |
| "loss": 0.7594, |
| "mean_token_accuracy": 0.7714629459028487, |
| "step": 145 |
| }, |
| { |
| "epoch": 0.8875739644970414, |
| "grad_norm": 0.2907168596676241, |
| "learning_rate": 7.612046748871327e-07, |
| "loss": 0.7505, |
| "mean_token_accuracy": 0.7739860872819682, |
| "step": 150 |
| }, |
| { |
| "epoch": 0.9171597633136095, |
| "grad_norm": 0.2786281346223686, |
| "learning_rate": 4.1572517541747294e-07, |
| "loss": 0.741, |
| "mean_token_accuracy": 0.7767968702085823, |
| "step": 155 |
| }, |
| { |
| "epoch": 0.9467455621301775, |
| "grad_norm": 0.29548388474741755, |
| "learning_rate": 1.7251026952640583e-07, |
| "loss": 0.754, |
| "mean_token_accuracy": 0.7731894129375776, |
| "step": 160 |
| }, |
| { |
| "epoch": 0.9763313609467456, |
| "grad_norm": 0.2773197515134327, |
| "learning_rate": 3.4155069933301535e-08, |
| "loss": 0.7467, |
| "mean_token_accuracy": 0.7749469974552831, |
| "step": 165 |
| }, |
| { |
| "epoch": 1.0, |
| "mean_token_accuracy": 0.7704768048458505, |
| "step": 169, |
| "total_flos": 76973799899136.0, |
| "train_loss": 0.8024970934941218, |
| "train_runtime": 942.7889, |
| "train_samples_per_second": 22.935, |
| "train_steps_per_second": 0.179 |
| } |
| ], |
| "logging_steps": 5, |
| "max_steps": 169, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 1, |
| "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": 76973799899136.0, |
| "train_batch_size": 2, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
