| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.4444444444444444, | |
| "eval_steps": 500, | |
| "global_step": 20, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "completion_length": 434.625, | |
| "epoch": 0.14814814814814814, | |
| "grad_norm": 0.9912062287330627, | |
| "kl": 0.0, | |
| "learning_rate": 4.965903258506806e-07, | |
| "loss": 0.0, | |
| "reward": 0.6409656591713428, | |
| "reward_std": 1.1219737641513348, | |
| "rewards/concensus_correctness_reward_func": 0.10987500101327896, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.0625, | |
| "rewards/question_recreation_reward_func": 0.36115315929055214, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.10743749095126987, | |
| "step": 2 | |
| }, | |
| { | |
| "completion_length": 355.0625, | |
| "epoch": 0.2962962962962963, | |
| "grad_norm": 1.8658636808395386, | |
| "kl": 0.0017933516865014099, | |
| "learning_rate": 4.698684378016222e-07, | |
| "loss": 0.0, | |
| "reward": 0.8250661976635456, | |
| "reward_std": 0.9006103053689003, | |
| "rewards/concensus_correctness_reward_func": 0.03831249848008156, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.1875, | |
| "rewards/question_recreation_reward_func": 0.35884744953364134, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.240406250115484, | |
| "step": 4 | |
| }, | |
| { | |
| "completion_length": 409.375, | |
| "epoch": 0.4444444444444444, | |
| "grad_norm": 2.6980597972869873, | |
| "kl": 0.0019965194660471752, | |
| "learning_rate": 4.193203929064353e-07, | |
| "loss": 0.0, | |
| "reward": 0.8316713124513626, | |
| "reward_std": 1.30080346763134, | |
| "rewards/concensus_correctness_reward_func": 0.1365625001490116, | |
| "rewards/consensus_reward_func": 0.125, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.0, | |
| "rewards/question_recreation_reward_func": 0.32526505773421377, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.2448437511920929, | |
| "step": 6 | |
| }, | |
| { | |
| "completion_length": 407.8125, | |
| "epoch": 0.5925925925925926, | |
| "grad_norm": 2.0010826587677, | |
| "kl": 0.0024274730385513976, | |
| "learning_rate": 3.5042385616324236e-07, | |
| "loss": 0.0, | |
| "reward": 2.955460924655199, | |
| "reward_std": 3.671143224462867, | |
| "rewards/concensus_correctness_reward_func": 1.5459375008940697, | |
| "rewards/consensus_reward_func": 0.3125, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.6875, | |
| "rewards/question_recreation_reward_func": 0.4681795723736286, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.03125, | |
| "rewards/xmlcount_reward_func": -0.08990624221041799, | |
| "step": 8 | |
| }, | |
| { | |
| "completion_length": 411.46875, | |
| "epoch": 0.7407407407407407, | |
| "grad_norm": 2.5793025493621826, | |
| "kl": 0.0022733503428753465, | |
| "learning_rate": 2.706448363680831e-07, | |
| "loss": 0.0, | |
| "reward": 2.74161918181926, | |
| "reward_std": 3.8083979981020093, | |
| "rewards/concensus_correctness_reward_func": 1.4272499978542328, | |
| "rewards/consensus_reward_func": 0.3125, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.5, | |
| "rewards/question_recreation_reward_func": 0.37393174413591623, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.12793749920092523, | |
| "step": 10 | |
| }, | |
| { | |
| "completion_length": 297.3125, | |
| "epoch": 0.8888888888888888, | |
| "grad_norm": 2.6540017127990723, | |
| "kl": 0.002707521736738272, | |
| "learning_rate": 1.886286282148002e-07, | |
| "loss": 0.0, | |
| "reward": 2.4840898173861206, | |
| "reward_std": 2.439718505134806, | |
| "rewards/concensus_correctness_reward_func": 1.4270000010728836, | |
| "rewards/consensus_reward_func": 0.0625, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.25, | |
| "rewards/question_recreation_reward_func": 0.34905852330848575, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.3955312464386225, | |
| "step": 12 | |
| }, | |
| { | |
| "completion_length": 455.4166666666667, | |
| "epoch": 1.0, | |
| "grad_norm": 2.082179307937622, | |
| "kl": 0.0017514083980737876, | |
| "learning_rate": 1.1326296046939333e-07, | |
| "loss": 0.0, | |
| "reward": 3.3851648767789206, | |
| "reward_std": 4.538816452026367, | |
| "rewards/concensus_correctness_reward_func": 1.6911666666467984, | |
| "rewards/consensus_reward_func": 0.4166666666666667, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.6666666666666666, | |
| "rewards/question_recreation_reward_func": 0.3862898349761963, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.2243749921520551, | |
| "step": 14 | |
| }, | |
| { | |
| "completion_length": 363.40625, | |
| "epoch": 1.1481481481481481, | |
| "grad_norm": 2.338515043258667, | |
| "kl": 0.003658254208858125, | |
| "learning_rate": 5.271487265090163e-08, | |
| "loss": 0.0, | |
| "reward": 2.821942111477256, | |
| "reward_std": 4.056427404284477, | |
| "rewards/concensus_correctness_reward_func": 1.3848125003278255, | |
| "rewards/consensus_reward_func": 0.1875, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.625, | |
| "rewards/question_recreation_reward_func": 0.4205670952796936, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.20406249328516424, | |
| "step": 16 | |
| }, | |
| { | |
| "completion_length": 428.15625, | |
| "epoch": 1.2962962962962963, | |
| "grad_norm": 2.7204856872558594, | |
| "kl": 0.002255684412375558, | |
| "learning_rate": 1.3545689574841341e-08, | |
| "loss": 0.0, | |
| "reward": 2.408185739070177, | |
| "reward_std": 4.392482792958617, | |
| "rewards/concensus_correctness_reward_func": 1.875, | |
| "rewards/consensus_reward_func": 0.125, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.375, | |
| "rewards/question_recreation_reward_func": 0.2698732349090278, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": -0.23668750002980232, | |
| "step": 18 | |
| }, | |
| { | |
| "completion_length": 432.03125, | |
| "epoch": 1.4444444444444444, | |
| "grad_norm": 2.1787333488464355, | |
| "kl": 0.0016690333577571437, | |
| "learning_rate": 0.0, | |
| "loss": 0.0, | |
| "reward": 2.4449838884174824, | |
| "reward_std": 3.9367115050554276, | |
| "rewards/concensus_correctness_reward_func": 1.25, | |
| "rewards/consensus_reward_func": 0.25, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.5, | |
| "rewards/question_recreation_reward_func": 0.3488589059561491, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.09612499922513962, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 1.4444444444444444, | |
| "step": 20, | |
| "total_flos": 0.0, | |
| "train_loss": 2.0203543867669537e-06, | |
| "train_runtime": 619.8316, | |
| "train_samples_per_second": 0.516, | |
| "train_steps_per_second": 0.032 | |
| } | |
| ], | |
| "logging_steps": 2, | |
| "max_steps": 20, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "save_steps": 25, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |