| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 9.8, | |
| "eval_steps": 500, | |
| "global_step": 10, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 0.3125, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 812.5, | |
| "completions/mean_length": 647.75, | |
| "completions/mean_terminated_length": 475.3166809082031, | |
| "completions/min_length": 138.5, | |
| "completions/min_terminated_length": 138.5, | |
| "epoch": 1.8, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 3.236729621887207, | |
| "kl": 0.00027297700398776215, | |
| "learning_rate": 5e-07, | |
| "loss": -0.001, | |
| "num_tokens": 14460.0, | |
| "reward": 0.14074324816465378, | |
| "reward_std": 0.12783230934292078, | |
| "rewards/concensus_correctness_reward_func/mean": 0.07774999737739563, | |
| "rewards/concensus_correctness_reward_func/std": 0.21991021931171417, | |
| "rewards/consensus_reward_func/mean": 0.0, | |
| "rewards/consensus_reward_func/std": 0.0, | |
| "rewards/cumulative_reward_2/mean": 0.0, | |
| "rewards/cumulative_reward_2/std": 0.0, | |
| "rewards/final_correctness_reward_func/mean": 0.0, | |
| "rewards/final_correctness_reward_func/std": 0.0, | |
| "rewards/question_recreation_reward_func/mean": 0.06299325078725815, | |
| "rewards/question_recreation_reward_func/std": 0.018077346496284008, | |
| "rewards/soft_format_reward_func/mean": 0.0, | |
| "rewards/soft_format_reward_func/std": 0.0, | |
| "rewards/strict_format_reward_func/mean": 0.0, | |
| "rewards/strict_format_reward_func/std": 0.0, | |
| "rewards/xmlcount_reward_func/mean": 0.0, | |
| "rewards/xmlcount_reward_func/std": 0.0, | |
| "step": 2 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 0.1875, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 754.0, | |
| "completions/mean_length": 397.25, | |
| "completions/mean_terminated_length": 251.08333587646484, | |
| "completions/min_length": 4.0, | |
| "completions/min_terminated_length": 4.0, | |
| "epoch": 3.8, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 4.27091646194458, | |
| "kl": 0.0006203544398886152, | |
| "learning_rate": 4.864543104251586e-07, | |
| "loss": 0.0127, | |
| "num_tokens": 24912.0, | |
| "reward": 0.1337380139157176, | |
| "reward_std": 0.1742639576550573, | |
| "rewards/concensus_correctness_reward_func/mean": 0.11999999731779099, | |
| "rewards/concensus_correctness_reward_func/std": 0.33941125869750977, | |
| "rewards/consensus_reward_func/mean": 0.0, | |
| "rewards/consensus_reward_func/std": 0.0, | |
| "rewards/cumulative_reward_2/mean": 0.0, | |
| "rewards/cumulative_reward_2/std": 0.0, | |
| "rewards/final_correctness_reward_func/mean": 0.0, | |
| "rewards/final_correctness_reward_func/std": 0.0, | |
| "rewards/question_recreation_reward_func/mean": 0.013738006353378296, | |
| "rewards/question_recreation_reward_func/std": 0.010505724931135774, | |
| "rewards/soft_format_reward_func/mean": 0.0, | |
| "rewards/soft_format_reward_func/std": 0.0, | |
| "rewards/strict_format_reward_func/mean": 0.0, | |
| "rewards/strict_format_reward_func/std": 0.0, | |
| "rewards/xmlcount_reward_func/mean": 0.0, | |
| "rewards/xmlcount_reward_func/std": 0.0, | |
| "step": 4 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 0.1875, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 846.5, | |
| "completions/mean_length": 528.4375, | |
| "completions/mean_terminated_length": 422.8214416503906, | |
| "completions/min_length": 4.0, | |
| "completions/min_terminated_length": 4.0, | |
| "epoch": 5.8, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 3.0706632137298584, | |
| "kl": 0.0008973750846053008, | |
| "learning_rate": 4.472851273490984e-07, | |
| "loss": 0.0403, | |
| "num_tokens": 37463.0, | |
| "reward": 0.12599835265427828, | |
| "reward_std": 0.16678897803649306, | |
| "rewards/concensus_correctness_reward_func/mean": 0.1107499971985817, | |
| "rewards/concensus_correctness_reward_func/std": 0.31324827671051025, | |
| "rewards/consensus_reward_func/mean": 0.0, | |
| "rewards/consensus_reward_func/std": 0.0, | |
| "rewards/cumulative_reward_2/mean": 0.0, | |
| "rewards/cumulative_reward_2/std": 0.0, | |
| "rewards/final_correctness_reward_func/mean": 0.0, | |
| "rewards/final_correctness_reward_func/std": 0.0, | |
| "rewards/question_recreation_reward_func/mean": 0.015248352661728859, | |
| "rewards/question_recreation_reward_func/std": 0.015421947930008173, | |
| "rewards/soft_format_reward_func/mean": 0.0, | |
| "rewards/soft_format_reward_func/std": 0.0, | |
| "rewards/strict_format_reward_func/mean": 0.0, | |
| "rewards/strict_format_reward_func/std": 0.0, | |
| "rewards/xmlcount_reward_func/mean": 0.0, | |
| "rewards/xmlcount_reward_func/std": 0.0, | |
| "step": 6 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 0.3125, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 744.5, | |
| "completions/mean_length": 588.75, | |
| "completions/mean_terminated_length": 384.3666687011719, | |
| "completions/min_length": 119.0, | |
| "completions/min_terminated_length": 119.0, | |
| "epoch": 7.8, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 3.6464202404022217, | |
| "kl": 0.0010676352321752347, | |
| "learning_rate": 3.867370395306068e-07, | |
| "loss": 0.1472, | |
| "num_tokens": 50979.0, | |
| "reward": 0.05322604067623615, | |
| "reward_std": 0.02834410360082984, | |
| "rewards/concensus_correctness_reward_func/mean": 0.0, | |
| "rewards/concensus_correctness_reward_func/std": 0.0, | |
| "rewards/consensus_reward_func/mean": 0.0, | |
| "rewards/consensus_reward_func/std": 0.0, | |
| "rewards/cumulative_reward_2/mean": 0.0, | |
| "rewards/cumulative_reward_2/std": 0.0, | |
| "rewards/final_correctness_reward_func/mean": 0.0, | |
| "rewards/final_correctness_reward_func/std": 0.0, | |
| "rewards/question_recreation_reward_func/mean": 0.05322604067623615, | |
| "rewards/question_recreation_reward_func/std": 0.03214400727301836, | |
| "rewards/soft_format_reward_func/mean": 0.0, | |
| "rewards/soft_format_reward_func/std": 0.0, | |
| "rewards/strict_format_reward_func/mean": 0.0, | |
| "rewards/strict_format_reward_func/std": 0.0, | |
| "rewards/xmlcount_reward_func/mean": 0.0, | |
| "rewards/xmlcount_reward_func/std": 0.0, | |
| "step": 8 | |
| }, | |
| { | |
| "clip_ratio/high_max": 0.0, | |
| "clip_ratio/high_mean": 0.0, | |
| "clip_ratio/low_mean": 0.0, | |
| "clip_ratio/low_min": 0.0, | |
| "clip_ratio/region_mean": 0.0, | |
| "completions/clipped_ratio": 0.375, | |
| "completions/max_length": 1024.0, | |
| "completions/max_terminated_length": 817.5, | |
| "completions/mean_length": 669.4375, | |
| "completions/mean_terminated_length": 470.0, | |
| "completions/min_length": 148.5, | |
| "completions/min_terminated_length": 148.5, | |
| "epoch": 9.8, | |
| "frac_reward_zero_std": 0.0, | |
| "grad_norm": 3.48903226852417, | |
| "kl": 0.0014760679550818168, | |
| "learning_rate": 3.1137137178519977e-07, | |
| "loss": 0.2054, | |
| "num_tokens": 65786.0, | |
| "reward": 0.011943170800805092, | |
| "reward_std": 0.0064218416810035706, | |
| "rewards/concensus_correctness_reward_func/mean": 0.0, | |
| "rewards/concensus_correctness_reward_func/std": 0.0, | |
| "rewards/consensus_reward_func/mean": 0.0, | |
| "rewards/consensus_reward_func/std": 0.0, | |
| "rewards/cumulative_reward_2/mean": 0.0, | |
| "rewards/cumulative_reward_2/std": 0.0, | |
| "rewards/final_correctness_reward_func/mean": 0.0, | |
| "rewards/final_correctness_reward_func/std": 0.0, | |
| "rewards/question_recreation_reward_func/mean": 0.011943170800805092, | |
| "rewards/question_recreation_reward_func/std": 0.0063162968726828694, | |
| "rewards/soft_format_reward_func/mean": 0.0, | |
| "rewards/soft_format_reward_func/std": 0.0, | |
| "rewards/strict_format_reward_func/mean": 0.0, | |
| "rewards/strict_format_reward_func/std": 0.0, | |
| "rewards/xmlcount_reward_func/mean": 0.0, | |
| "rewards/xmlcount_reward_func/std": 0.0, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 9.8, | |
| "step": 10, | |
| "total_flos": 0.0, | |
| "train_loss": 0.08094322010874748, | |
| "train_runtime": 956.1563, | |
| "train_samples_per_second": 0.167, | |
| "train_steps_per_second": 0.021 | |
| } | |
| ], | |
| "logging_steps": 2, | |
| "max_steps": 20, | |
| "num_input_tokens_seen": 65786, | |
| "num_train_epochs": 10, | |
| "save_steps": 25, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": false, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |