| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 1.0, | |
| "eval_steps": 100, | |
| "global_step": 61, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.02, | |
| "grad_norm": 5.823829876222625, | |
| "learning_rate": 7.142857142857142e-08, | |
| "logits/chosen": -1.0549558401107788, | |
| "logits/rejected": -0.9334831833839417, | |
| "logps/chosen": -681.2974853515625, | |
| "logps/rejected": -846.9592895507812, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "grad_norm": 5.674915170284527, | |
| "learning_rate": 4.96201938253052e-07, | |
| "logits/chosen": -1.1824066638946533, | |
| "logits/rejected": -0.937175989151001, | |
| "logps/chosen": -563.275634765625, | |
| "logps/rejected": -894.625732421875, | |
| "loss": 0.6921, | |
| "rewards/accuracies": 0.5138888955116272, | |
| "rewards/chosen": 0.00011738391185645014, | |
| "rewards/margins": 0.0022102664224803448, | |
| "rewards/rejected": -0.0020928820595145226, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "grad_norm": 6.471345406032975, | |
| "learning_rate": 4.3184341039326217e-07, | |
| "logits/chosen": -1.1980640888214111, | |
| "logits/rejected": -0.9605765342712402, | |
| "logps/chosen": -577.985595703125, | |
| "logps/rejected": -858.2009887695312, | |
| "loss": 0.675, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -0.0019469285616651177, | |
| "rewards/margins": 0.03744823858141899, | |
| "rewards/rejected": -0.03939516842365265, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "grad_norm": 7.131737915304761, | |
| "learning_rate": 3.0765396768561004e-07, | |
| "logits/chosen": -1.18826162815094, | |
| "logits/rejected": -0.9675145149230957, | |
| "logps/chosen": -519.7886962890625, | |
| "logps/rejected": -910.9063720703125, | |
| "loss": 0.6241, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": -0.014793062582612038, | |
| "rewards/margins": 0.13844963908195496, | |
| "rewards/rejected": -0.15324273705482483, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "grad_norm": 7.706616708059976, | |
| "learning_rate": 1.6449496416858282e-07, | |
| "logits/chosen": -1.187648892402649, | |
| "logits/rejected": -0.9329156875610352, | |
| "logps/chosen": -602.1197509765625, | |
| "logps/rejected": -954.8004150390625, | |
| "loss": 0.543, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -0.06608068197965622, | |
| "rewards/margins": 0.4169215261936188, | |
| "rewards/rejected": -0.483002245426178, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "grad_norm": 7.642531846302412, | |
| "learning_rate": 4.9469201811239035e-08, | |
| "logits/chosen": -1.1323978900909424, | |
| "logits/rejected": -0.9658657312393188, | |
| "logps/chosen": -556.5906982421875, | |
| "logps/rejected": -980.1487426757812, | |
| "loss": 0.5093, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -0.13172821700572968, | |
| "rewards/margins": 0.5675451755523682, | |
| "rewards/rejected": -0.6992733478546143, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "grad_norm": 7.37383781369435, | |
| "learning_rate": 4.2296043218295606e-10, | |
| "logits/chosen": -1.1500413417816162, | |
| "logits/rejected": -0.9949632883071899, | |
| "logps/chosen": -577.3805541992188, | |
| "logps/rejected": -966.5217895507812, | |
| "loss": 0.4811, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.1981235146522522, | |
| "rewards/margins": 0.5871651768684387, | |
| "rewards/rejected": -0.7852886915206909, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 61, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5870200258786561, | |
| "train_runtime": 812.4392, | |
| "train_samples_per_second": 4.783, | |
| "train_steps_per_second": 0.075 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 61, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |