| { | |
| "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": 253.96875, | |
| "epoch": 0.1, | |
| "grad_norm": 54.46091842651367, | |
| "kl": 0.0, | |
| "learning_rate": 4.965903258506806e-07, | |
| "loss": -0.0, | |
| "reward": 1.0164429601281881, | |
| "reward_std": 1.3832301030051894, | |
| "rewards/concensus_correctness_reward_func": 0.625, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.25, | |
| "rewards/question_recreation_reward_func": 0.157911735936068, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": -0.016468747053295374, | |
| "step": 2 | |
| }, | |
| { | |
| "completion_length": 300.0625, | |
| "epoch": 0.2, | |
| "grad_norm": 11.377876281738281, | |
| "kl": 0.0016310415630869102, | |
| "learning_rate": 4.698684378016222e-07, | |
| "loss": 0.0, | |
| "reward": 1.3625886285444722, | |
| "reward_std": 1.3904136032506358, | |
| "rewards/concensus_correctness_reward_func": 0.625, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.5625, | |
| "rewards/question_recreation_reward_func": 0.19358859345084056, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": -0.01849999837577343, | |
| "step": 4 | |
| }, | |
| { | |
| "completion_length": 329.25, | |
| "epoch": 0.3, | |
| "grad_norm": 7.467350482940674, | |
| "kl": 0.0011920945034944452, | |
| "learning_rate": 4.193203929064353e-07, | |
| "loss": 0.0, | |
| "reward": 0.679933191742748, | |
| "reward_std": 0.5611653565429151, | |
| "rewards/concensus_correctness_reward_func": 0.0, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.3125, | |
| "rewards/question_recreation_reward_func": 0.15724571567261592, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.2101875024382025, | |
| "step": 6 | |
| }, | |
| { | |
| "completion_length": 340.28125, | |
| "epoch": 0.4, | |
| "grad_norm": 12.62759780883789, | |
| "kl": 0.0016059228400990833, | |
| "learning_rate": 3.5042385616324236e-07, | |
| "loss": 0.0, | |
| "reward": 0.22430133307352662, | |
| "reward_std": 0.7189028215361759, | |
| "rewards/concensus_correctness_reward_func": 0.0, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.3125, | |
| "rewards/question_recreation_reward_func": 0.138676329923328, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": -0.2268749950453639, | |
| "step": 8 | |
| }, | |
| { | |
| "completion_length": 304.40625, | |
| "epoch": 0.5, | |
| "grad_norm": 12.517777442932129, | |
| "kl": 0.002796674736600835, | |
| "learning_rate": 2.706448363680831e-07, | |
| "loss": 0.0, | |
| "reward": 0.7383955912664533, | |
| "reward_std": 0.8703088565671351, | |
| "rewards/concensus_correctness_reward_func": 0.11275000125169754, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.1875, | |
| "rewards/question_recreation_reward_func": 0.2436143477098085, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.1945312488824129, | |
| "step": 10 | |
| }, | |
| { | |
| "completion_length": 261.65625, | |
| "epoch": 0.6, | |
| "grad_norm": 25.072782516479492, | |
| "kl": 0.008284762378025334, | |
| "learning_rate": 1.886286282148002e-07, | |
| "loss": 0.0, | |
| "reward": 1.612264923285693, | |
| "reward_std": 1.8028453141450882, | |
| "rewards/concensus_correctness_reward_func": 1.25, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.0625, | |
| "rewards/question_recreation_reward_func": 0.13995241769589484, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.15981250256299973, | |
| "step": 12 | |
| }, | |
| { | |
| "completion_length": 277.1875, | |
| "epoch": 0.7, | |
| "grad_norm": 17.136720657348633, | |
| "kl": 0.0029835424866178073, | |
| "learning_rate": 1.1326296046939333e-07, | |
| "loss": 0.0, | |
| "reward": 0.6826112208073027, | |
| "reward_std": 0.627162155113183, | |
| "rewards/concensus_correctness_reward_func": 0.06012500077486038, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.3125, | |
| "rewards/question_recreation_reward_func": 0.24014246050501242, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.06984375091269612, | |
| "step": 14 | |
| }, | |
| { | |
| "completion_length": 375.34375, | |
| "epoch": 0.8, | |
| "grad_norm": 59.931880950927734, | |
| "kl": 0.003028268714842852, | |
| "learning_rate": 5.271487265090163e-08, | |
| "loss": 0.0, | |
| "reward": 0.21182444714941084, | |
| "reward_std": 0.2998874420009088, | |
| "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": 0.09560569762834348, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": -0.008781249634921554, | |
| "step": 16 | |
| }, | |
| { | |
| "completion_length": 356.25, | |
| "epoch": 0.9, | |
| "grad_norm": 43.81071853637695, | |
| "kl": 0.006780517935112584, | |
| "learning_rate": 1.3545689574841341e-08, | |
| "loss": 0.0, | |
| "reward": 0.6197623474290594, | |
| "reward_std": 0.3784800235880539, | |
| "rewards/concensus_correctness_reward_func": 0.0, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.25, | |
| "rewards/question_recreation_reward_func": 0.19623111432883888, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": 0.17353125242516398, | |
| "step": 18 | |
| }, | |
| { | |
| "completion_length": 297.375, | |
| "epoch": 1.0, | |
| "grad_norm": 16.618606567382812, | |
| "kl": 0.0019668052918859757, | |
| "learning_rate": 0.0, | |
| "loss": 0.0, | |
| "reward": 0.4587072639260441, | |
| "reward_std": 0.46473248617257923, | |
| "rewards/concensus_correctness_reward_func": 0.0, | |
| "rewards/consensus_reward_func": 0.0, | |
| "rewards/cumulative_reward_2": 0.0, | |
| "rewards/final_correctness_reward_func": 0.375, | |
| "rewards/question_recreation_reward_func": 0.156519764859695, | |
| "rewards/soft_format_reward_func": 0.0, | |
| "rewards/strict_format_reward_func": 0.0, | |
| "rewards/xmlcount_reward_func": -0.07281250366941094, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 20, | |
| "total_flos": 0.0, | |
| "train_loss": 3.016861137439264e-06, | |
| "train_runtime": 243.3421, | |
| "train_samples_per_second": 1.315, | |
| "train_steps_per_second": 0.082 | |
| } | |
| ], | |
| "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 | |
| } | |