{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 2.0, "eval_steps": 100, "global_step": 42, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "entropy": 2.08540016412735, "epoch": 0.963855421686747, "grad_norm": 1.4502723217010498, "learning_rate": 1.9000000000000002e-06, "logits/chosen": 1.9373448673414857, "logits/rejected": 1.4560297831585156, "logps/chosen": -43.992462682723996, "logps/rejected": -51.50791831016541, "loss": 0.6897519111633301, "mean_token_accuracy": 0.2624862600117922, "num_tokens": 37803.0, "rewards/accuracies": 0.7875, "rewards/chosen": 0.0014421886393392925, "rewards/margins": 0.0068238718846259875, "rewards/rejected": -0.005381683184623398, "step": 20 }, { "entropy": 2.1433570188812063, "epoch": 1.9156626506024095, "grad_norm": 1.7149038314819336, "learning_rate": 3.900000000000001e-06, "logits/chosen": 2.0310179220420284, "logits/rejected": 1.5091439161978566, "logps/chosen": -43.38176954245265, "logps/rejected": -52.56174864950059, "loss": 0.628478717803955, "mean_token_accuracy": 0.2560643789511693, "num_tokens": 74911.0, "rewards/accuracies": 0.9968354430379747, "rewards/chosen": 0.02222161173903018, "rewards/margins": 0.13752426909683627, "rewards/rejected": -0.11530265806219246, "step": 40 } ], "logging_steps": 20, "max_steps": 63, "num_input_tokens_seen": 0, "num_train_epochs": 3, "save_steps": 500, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 2006470394830848.0, "train_batch_size": 4, "trial_name": null, "trial_params": null }