| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "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": 122.6875, | |
| "epoch": 0.1, | |
| "grad_norm": 0.0, | |
| "kl": 0.0, | |
| "learning_rate": 4.965903258506806e-07, | |
| "loss": 0.0, | |
| "reward": 6.851499974727631, | |
| "reward_std": 0.0, | |
| "rewards/concensus_correctness_reward_func": 1.9765000082552433, | |
| "rewards/consensus_reward_func": 2.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.125, | |
| "rewards/question_recreation_reward_func": 1.0, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.5, | |
| "rewards/xmlcount_reward_func": 1.25, | |
| "step": 2 | |
| }, | |
| { | |
| "completion_length": 117.71875, | |
| "epoch": 0.2, | |
| "grad_norm": 0.0025590641889721155, | |
| "kl": 0.00010481959861863288, | |
| "learning_rate": 4.698684378016222e-07, | |
| "loss": 0.0, | |
| "reward": 6.8094218000769615, | |
| "reward_std": 0.02787156030535698, | |
| "rewards/concensus_correctness_reward_func": 2.0442499965429306, | |
| "rewards/consensus_reward_func": 1.875, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.25, | |
| "rewards/question_recreation_reward_func": 0.9839218780398369, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.46875, | |
| "rewards/xmlcount_reward_func": 1.1875, | |
| "step": 4 | |
| }, | |
| { | |
| "completion_length": 124.75, | |
| "epoch": 0.3, | |
| "grad_norm": 0.004200837574899197, | |
| "kl": 0.0009384811905874813, | |
| "learning_rate": 4.193203929064353e-07, | |
| "loss": 0.0, | |
| "reward": 6.331388913094997, | |
| "reward_std": 0.0, | |
| "rewards/concensus_correctness_reward_func": 1.7947500124573708, | |
| "rewards/consensus_reward_func": 1.875, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.0, | |
| "rewards/question_recreation_reward_func": 0.989763967692852, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.46875, | |
| "rewards/xmlcount_reward_func": 1.203125, | |
| "step": 6 | |
| }, | |
| { | |
| "completion_length": 137.53125, | |
| "epoch": 0.4, | |
| "grad_norm": 1.0187238454818726, | |
| "kl": 0.0013889123110857327, | |
| "learning_rate": 3.5042385616324236e-07, | |
| "loss": 0.0, | |
| "reward": 6.525790655985475, | |
| "reward_std": 0.023460550233721733, | |
| "rewards/concensus_correctness_reward_func": 1.9200000166893005, | |
| "rewards/consensus_reward_func": 1.875, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.125, | |
| "rewards/question_recreation_reward_func": 0.9807907268404961, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.46875, | |
| "rewards/xmlcount_reward_func": 1.15625, | |
| "step": 8 | |
| }, | |
| { | |
| "completion_length": 116.21875, | |
| "epoch": 0.5, | |
| "grad_norm": 0.0037905466742813587, | |
| "kl": 0.0010511610335015575, | |
| "learning_rate": 2.706448363680831e-07, | |
| "loss": 0.0, | |
| "reward": 7.413999915122986, | |
| "reward_std": 0.0, | |
| "rewards/concensus_correctness_reward_func": 2.289000004529953, | |
| "rewards/consensus_reward_func": 2.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.375, | |
| "rewards/question_recreation_reward_func": 1.0, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.5, | |
| "rewards/xmlcount_reward_func": 1.25, | |
| "step": 10 | |
| }, | |
| { | |
| "completion_length": 145.96875, | |
| "epoch": 0.6, | |
| "grad_norm": 0.6090595126152039, | |
| "kl": 0.0015178388712229207, | |
| "learning_rate": 1.886286282148002e-07, | |
| "loss": 0.0, | |
| "reward": 6.296343315392733, | |
| "reward_std": 0.047967955470085144, | |
| "rewards/concensus_correctness_reward_func": 1.7946250140666962, | |
| "rewards/consensus_reward_func": 1.875, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.0, | |
| "rewards/question_recreation_reward_func": 0.9704683646559715, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.46875, | |
| "rewards/xmlcount_reward_func": 1.1875, | |
| "step": 12 | |
| }, | |
| { | |
| "completion_length": 118.6875, | |
| "epoch": 0.7, | |
| "grad_norm": 1.6188795566558838, | |
| "kl": 0.002664562603968079, | |
| "learning_rate": 1.1326296046939333e-07, | |
| "loss": -0.0, | |
| "reward": 6.830424338579178, | |
| "reward_std": 0.00013725121971219778, | |
| "rewards/concensus_correctness_reward_func": 2.044375002384186, | |
| "rewards/consensus_reward_func": 1.875, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.25, | |
| "rewards/question_recreation_reward_func": 0.9891744069755077, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.46875, | |
| "rewards/xmlcount_reward_func": 1.203125, | |
| "step": 14 | |
| }, | |
| { | |
| "completion_length": 120.5625, | |
| "epoch": 0.8, | |
| "grad_norm": 0.008192956447601318, | |
| "kl": 0.007514026697208465, | |
| "learning_rate": 5.271487265090163e-08, | |
| "loss": 0.0, | |
| "reward": 6.4971517994999886, | |
| "reward_std": 0.11160904914140701, | |
| "rewards/concensus_correctness_reward_func": 1.924625001847744, | |
| "rewards/consensus_reward_func": 1.75, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.3125, | |
| "rewards/question_recreation_reward_func": 0.931901843752712, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.4375, | |
| "rewards/xmlcount_reward_func": 1.140625, | |
| "step": 16 | |
| }, | |
| { | |
| "completion_length": 115.53125, | |
| "epoch": 0.9, | |
| "grad_norm": 0.009131981059908867, | |
| "kl": 0.004858428852458019, | |
| "learning_rate": 1.3545689574841341e-08, | |
| "loss": 0.0, | |
| "reward": 6.914499908685684, | |
| "reward_std": 0.0, | |
| "rewards/concensus_correctness_reward_func": 2.039500005543232, | |
| "rewards/consensus_reward_func": 2.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.125, | |
| "rewards/question_recreation_reward_func": 1.0, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.5, | |
| "rewards/xmlcount_reward_func": 1.25, | |
| "step": 18 | |
| }, | |
| { | |
| "completion_length": 118.0625, | |
| "epoch": 1.0, | |
| "grad_norm": 0.005396973341703415, | |
| "kl": 0.004226890665677274, | |
| "learning_rate": 0.0, | |
| "loss": 0.0, | |
| "reward": 6.9989491030573845, | |
| "reward_std": 0.00043361730422475375, | |
| "rewards/concensus_correctness_reward_func": 2.174249991774559, | |
| "rewards/consensus_reward_func": 1.75, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.5, | |
| "rewards/question_recreation_reward_func": 0.980949155986309, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.4375, | |
| "rewards/xmlcount_reward_func": 1.15625, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 20, | |
| "total_flos": 0.0, | |
| "train_loss": 1.7776284941817266e-06, | |
| "train_runtime": 5637.9637, | |
| "train_samples_per_second": 0.057, | |
| "train_steps_per_second": 0.004 | |
| } | |
| ], | |
| "logging_steps": 2, | |
| "max_steps": 20, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "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": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |