| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 4.0, | |
| "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": 537.6875, | |
| "epoch": 0.4, | |
| "grad_norm": 1.4502513408660889, | |
| "kl": 0.0045772524463245645, | |
| "learning_rate": 4.965903258506806e-07, | |
| "loss": 0.0, | |
| "reward": 1.4177430057898164, | |
| "reward_std": 2.1470614364370704, | |
| "rewards/concensus_correctness_reward_func": 0.7554374933242798, | |
| "rewards/consensus_reward_func": 0.125, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.25, | |
| "rewards/question_recreation_reward_func": 0.25036793970502913, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.03693749010562897, | |
| "step": 2 | |
| }, | |
| { | |
| "completion_length": 468.53125, | |
| "epoch": 0.8, | |
| "grad_norm": 1.9520736932754517, | |
| "kl": 0.001855591282946989, | |
| "learning_rate": 4.698684378016222e-07, | |
| "loss": 0.0, | |
| "reward": 0.3959203027188778, | |
| "reward_std": 1.1887737782672048, | |
| "rewards/concensus_correctness_reward_func": 0.03125, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.25, | |
| "rewards/question_recreation_reward_func": 0.2753265555948019, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": -0.16065625473856926, | |
| "step": 4 | |
| }, | |
| { | |
| "completion_length": 441.78125, | |
| "epoch": 1.2, | |
| "grad_norm": 1.735103726387024, | |
| "kl": 0.0016164953267434612, | |
| "learning_rate": 4.193203929064353e-07, | |
| "loss": 0.0, | |
| "reward": 0.623163734562695, | |
| "reward_std": 0.7843756377696991, | |
| "rewards/concensus_correctness_reward_func": 0.06006250157952309, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.25, | |
| "rewards/question_recreation_reward_func": 0.24881997099146247, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.06428124755620956, | |
| "step": 6 | |
| }, | |
| { | |
| "completion_length": 435.59375, | |
| "epoch": 1.6, | |
| "grad_norm": 2.2628138065338135, | |
| "kl": 0.0013706091995118186, | |
| "learning_rate": 3.5042385616324236e-07, | |
| "loss": 0.0, | |
| "reward": 0.3420835845172405, | |
| "reward_std": 0.797295784112066, | |
| "rewards/concensus_correctness_reward_func": 0.0, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.0625, | |
| "rewards/question_recreation_reward_func": 0.3639585729688406, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": -0.08437500428408384, | |
| "step": 8 | |
| }, | |
| { | |
| "completion_length": 319.5625, | |
| "epoch": 2.0, | |
| "grad_norm": 2.1416268348693848, | |
| "kl": 0.0025234216154785827, | |
| "learning_rate": 2.706448363680831e-07, | |
| "loss": 0.0, | |
| "reward": 1.0948227606713772, | |
| "reward_std": 1.085303759202361, | |
| "rewards/concensus_correctness_reward_func": 0.08193749934434891, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.25, | |
| "rewards/question_recreation_reward_func": 0.37897900864481926, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.38390624336898327, | |
| "step": 10 | |
| }, | |
| { | |
| "completion_length": 525.09375, | |
| "epoch": 2.4, | |
| "grad_norm": 2.21820068359375, | |
| "kl": 0.001763832857250236, | |
| "learning_rate": 1.886286282148002e-07, | |
| "loss": 0.0, | |
| "reward": 0.4163860874250531, | |
| "reward_std": 1.0169010870158672, | |
| "rewards/concensus_correctness_reward_func": 0.0, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.0625, | |
| "rewards/question_recreation_reward_func": 0.413198578171432, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": -0.0593124907463789, | |
| "step": 12 | |
| }, | |
| { | |
| "completion_length": 311.84375, | |
| "epoch": 2.8, | |
| "grad_norm": 2.4503183364868164, | |
| "kl": 0.0024037769326241687, | |
| "learning_rate": 1.1326296046939333e-07, | |
| "loss": 0.0, | |
| "reward": 1.1352357491850853, | |
| "reward_std": 1.1741696512326598, | |
| "rewards/concensus_correctness_reward_func": 0.15131249651312828, | |
| "rewards/consensus_reward_func": 0.0625, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.25, | |
| "rewards/question_recreation_reward_func": 0.37954823300242424, | |
| "rewards/soft_format_reward_func": 0.015625, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.2762500010430813, | |
| "step": 14 | |
| }, | |
| { | |
| "completion_length": 371.1875, | |
| "epoch": 3.2, | |
| "grad_norm": 2.3564932346343994, | |
| "kl": 0.003548893568222411, | |
| "learning_rate": 5.271487265090163e-08, | |
| "loss": 0.0, | |
| "reward": 0.6700164806097746, | |
| "reward_std": 0.8069747863337398, | |
| "rewards/concensus_correctness_reward_func": 0.0, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.125, | |
| "rewards/question_recreation_reward_func": 0.3270789897069335, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.2179374946281314, | |
| "step": 16 | |
| }, | |
| { | |
| "completion_length": 364.21875, | |
| "epoch": 3.6, | |
| "grad_norm": 3.2016563415527344, | |
| "kl": 0.0021984638296999037, | |
| "learning_rate": 1.3545689574841341e-08, | |
| "loss": 0.0, | |
| "reward": 0.9208500199019909, | |
| "reward_std": 1.1289626825600863, | |
| "rewards/concensus_correctness_reward_func": 0.21381250023841858, | |
| "rewards/consensus_reward_func": 0.0625, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.1875, | |
| "rewards/question_recreation_reward_func": 0.184193748049438, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.2728437576442957, | |
| "step": 18 | |
| }, | |
| { | |
| "completion_length": 364.34375, | |
| "epoch": 4.0, | |
| "grad_norm": 2.2764692306518555, | |
| "kl": 0.0035686935880221426, | |
| "learning_rate": 0.0, | |
| "loss": 0.0, | |
| "reward": 1.4730796124786139, | |
| "reward_std": 1.3868739902973175, | |
| "rewards/concensus_correctness_reward_func": 0.32718750089406967, | |
| "rewards/consensus_reward_func": 0.125, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.1875, | |
| "rewards/question_recreation_reward_func": 0.46451712027192116, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.3688750071451068, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "step": 20, | |
| "total_flos": 0.0, | |
| "train_loss": 2.546286020788102e-06, | |
| "train_runtime": 864.6962, | |
| "train_samples_per_second": 0.37, | |
| "train_steps_per_second": 0.023 | |
| } | |
| ], | |
| "logging_steps": 2, | |
| "max_steps": 20, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "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 | |
| } | |