| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 20.0, | |
| "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": 745.0, | |
| "epoch": 2.0, | |
| "grad_norm": 0.18274663388729095, | |
| "kl": 0.0, | |
| "learning_rate": 5e-07, | |
| "loss": -0.0, | |
| "reward": 1.084837019443512, | |
| "reward_std": 1.1805158257484436, | |
| "rewards/concensus_correctness_reward_func": 0.0, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.0, | |
| "rewards/question_recreation_reward_func": 0.3814619928598404, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0625, | |
| "rewards/xmlcount_reward_func": 0.6408749967813492, | |
| "step": 2 | |
| }, | |
| { | |
| "completion_length": 778.75, | |
| "epoch": 4.0, | |
| "grad_norm": 0.12910106778144836, | |
| "kl": 0.0005327967810444534, | |
| "learning_rate": 4.864543104251586e-07, | |
| "loss": 0.0, | |
| "reward": 0.07697300612926483, | |
| "reward_std": 1.4115747213363647, | |
| "rewards/concensus_correctness_reward_func": 0.0, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.0, | |
| "rewards/question_recreation_reward_func": 0.4600980132818222, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0625, | |
| "rewards/xmlcount_reward_func": -0.4456250071525574, | |
| "step": 4 | |
| }, | |
| { | |
| "completion_length": 703.125, | |
| "epoch": 6.0, | |
| "grad_norm": 0.1341792196035385, | |
| "kl": 0.0004545120464172214, | |
| "learning_rate": 4.472851273490984e-07, | |
| "loss": 0.0, | |
| "reward": 0.26113903522491455, | |
| "reward_std": 0.8645140528678894, | |
| "rewards/concensus_correctness_reward_func": 0.0, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.0, | |
| "rewards/question_recreation_reward_func": 0.45626404881477356, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": -0.19512498378753662, | |
| "step": 6 | |
| }, | |
| { | |
| "completion_length": 705.75, | |
| "epoch": 8.0, | |
| "grad_norm": 0.13170865178108215, | |
| "kl": 0.0005088421748951077, | |
| "learning_rate": 3.867370395306068e-07, | |
| "loss": 0.0, | |
| "reward": 0.266592875123024, | |
| "reward_std": 1.457129716873169, | |
| "rewards/concensus_correctness_reward_func": 0.0, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.0, | |
| "rewards/question_recreation_reward_func": 0.30996787548065186, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0625, | |
| "rewards/xmlcount_reward_func": -0.10587500035762787, | |
| "step": 8 | |
| }, | |
| { | |
| "completion_length": 679.375, | |
| "epoch": 10.0, | |
| "grad_norm": 0.1503751128911972, | |
| "kl": 0.0005099998670630157, | |
| "learning_rate": 3.1137137178519977e-07, | |
| "loss": 0.0, | |
| "reward": 0.0006678104400634766, | |
| "reward_std": 0.6963486075401306, | |
| "rewards/concensus_correctness_reward_func": 0.0, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.0, | |
| "rewards/question_recreation_reward_func": 0.22504279017448425, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": -0.22437500953674316, | |
| "step": 10 | |
| }, | |
| { | |
| "completion_length": 757.5, | |
| "epoch": 12.0, | |
| "grad_norm": 0.1588054895401001, | |
| "kl": 0.00046218409261200577, | |
| "learning_rate": 2.2935516363191693e-07, | |
| "loss": 0.0, | |
| "reward": -0.19009771943092346, | |
| "reward_std": 1.6218498945236206, | |
| "rewards/concensus_correctness_reward_func": 0.0, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.0, | |
| "rewards/question_recreation_reward_func": 0.17290228977799416, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": -0.3630000352859497, | |
| "step": 12 | |
| }, | |
| { | |
| "completion_length": 654.125, | |
| "epoch": 14.0, | |
| "grad_norm": 0.1310453712940216, | |
| "kl": 0.00050046292017214, | |
| "learning_rate": 1.4957614383675767e-07, | |
| "loss": 0.0, | |
| "reward": 0.7278059124946594, | |
| "reward_std": 1.1783943176269531, | |
| "rewards/concensus_correctness_reward_func": 0.0, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.0, | |
| "rewards/question_recreation_reward_func": 0.2610559053719044, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0625, | |
| "rewards/xmlcount_reward_func": 0.40424999594688416, | |
| "step": 14 | |
| }, | |
| { | |
| "completion_length": 703.5, | |
| "epoch": 16.0, | |
| "grad_norm": 0.1375514566898346, | |
| "kl": 0.00047489034477621317, | |
| "learning_rate": 8.067960709356478e-08, | |
| "loss": 0.0, | |
| "reward": -0.11608892679214478, | |
| "reward_std": 0.8753925263881683, | |
| "rewards/concensus_correctness_reward_func": 0.0, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.0, | |
| "rewards/question_recreation_reward_func": 0.5065360516309738, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0625, | |
| "rewards/xmlcount_reward_func": -0.6851249933242798, | |
| "step": 16 | |
| }, | |
| { | |
| "completion_length": 646.875, | |
| "epoch": 18.0, | |
| "grad_norm": 0.11112207174301147, | |
| "kl": 0.0005517661338672042, | |
| "learning_rate": 3.013156219837776e-08, | |
| "loss": 0.0, | |
| "reward": 0.1678205281496048, | |
| "reward_std": 1.0244586765766144, | |
| "rewards/concensus_correctness_reward_func": 0.0, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.0, | |
| "rewards/question_recreation_reward_func": 0.29194553941488266, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": -0.12412500381469727, | |
| "step": 18 | |
| }, | |
| { | |
| "completion_length": 705.0, | |
| "epoch": 20.0, | |
| "grad_norm": 0.06440959125757217, | |
| "kl": 0.0005306277598720044, | |
| "learning_rate": 3.4096741493194193e-09, | |
| "loss": 0.0, | |
| "reward": 0.6555706560611725, | |
| "reward_std": 0.8972451090812683, | |
| "rewards/concensus_correctness_reward_func": 0.0, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.0, | |
| "rewards/question_recreation_reward_func": 0.47207067906856537, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.1834999993443489, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 20.0, | |
| "step": 20, | |
| "total_flos": 0.0, | |
| "train_loss": 1.1366962233694267e-07, | |
| "train_runtime": 500.2968, | |
| "train_samples_per_second": 0.64, | |
| "train_steps_per_second": 0.04 | |
| } | |
| ], | |
| "logging_steps": 2, | |
| "max_steps": 20, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 20, | |
| "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 | |
| } | |