| { |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 0.9874476987447699, |
| "eval_steps": 500, |
| "global_step": 59, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.02, |
| "learning_rate": 8.333333333333333e-08, |
| "logits/chosen": -2.8012306690216064, |
| "logits/rejected": -2.7200045585632324, |
| "logps/chosen": -308.75421142578125, |
| "logps/pi_response": -233.9468231201172, |
| "logps/ref_response": -233.9468231201172, |
| "logps/rejected": -311.24969482421875, |
| "loss": 0.6931, |
| "rewards/accuracies": 0.0, |
| "rewards/chosen": 0.0, |
| "rewards/margins": 0.0, |
| "rewards/rejected": 0.0, |
| "step": 1 |
| }, |
| { |
| "epoch": 0.17, |
| "learning_rate": 4.930057285201027e-07, |
| "logits/chosen": -2.7711262702941895, |
| "logits/rejected": -2.709352731704712, |
| "logps/chosen": -263.98291015625, |
| "logps/pi_response": -193.27850341796875, |
| "logps/ref_response": -193.61187744140625, |
| "logps/rejected": -278.3876037597656, |
| "loss": 0.69, |
| "rewards/accuracies": 0.59375, |
| "rewards/chosen": 0.011274036020040512, |
| "rewards/margins": 0.006776580587029457, |
| "rewards/rejected": 0.004497454967349768, |
| "step": 10 |
| }, |
| { |
| "epoch": 0.33, |
| "learning_rate": 4.187457503795526e-07, |
| "logits/chosen": -2.811276912689209, |
| "logits/rejected": -2.751648426055908, |
| "logps/chosen": -266.3084411621094, |
| "logps/pi_response": -196.0972137451172, |
| "logps/ref_response": -193.69471740722656, |
| "logps/rejected": -272.2674255371094, |
| "loss": 0.6601, |
| "rewards/accuracies": 0.765625, |
| "rewards/chosen": 0.045443661510944366, |
| "rewards/margins": 0.06504254043102264, |
| "rewards/rejected": -0.019598882645368576, |
| "step": 20 |
| }, |
| { |
| "epoch": 0.5, |
| "learning_rate": 2.8691164100062034e-07, |
| "logits/chosen": -2.8319926261901855, |
| "logits/rejected": -2.7600412368774414, |
| "logps/chosen": -287.2323303222656, |
| "logps/pi_response": -224.0599365234375, |
| "logps/ref_response": -200.99618530273438, |
| "logps/rejected": -287.2210998535156, |
| "loss": 0.6115, |
| "rewards/accuracies": 0.824999988079071, |
| "rewards/chosen": -0.010728769935667515, |
| "rewards/margins": 0.23644748330116272, |
| "rewards/rejected": -0.24717621505260468, |
| "step": 30 |
| }, |
| { |
| "epoch": 0.67, |
| "learning_rate": 1.4248369943086995e-07, |
| "logits/chosen": -2.7870166301727295, |
| "logits/rejected": -2.7307822704315186, |
| "logps/chosen": -258.9184265136719, |
| "logps/pi_response": -226.122314453125, |
| "logps/ref_response": -185.12936401367188, |
| "logps/rejected": -311.3656311035156, |
| "loss": 0.5706, |
| "rewards/accuracies": 0.809374988079071, |
| "rewards/chosen": -0.15448638796806335, |
| "rewards/margins": 0.28851866722106934, |
| "rewards/rejected": -0.4430050253868103, |
| "step": 40 |
| }, |
| { |
| "epoch": 0.84, |
| "learning_rate": 3.473909705816111e-08, |
| "logits/chosen": -2.794811964035034, |
| "logits/rejected": -2.7504866123199463, |
| "logps/chosen": -283.52069091796875, |
| "logps/pi_response": -245.2189483642578, |
| "logps/ref_response": -190.3854522705078, |
| "logps/rejected": -315.20977783203125, |
| "loss": 0.5375, |
| "rewards/accuracies": 0.8031250238418579, |
| "rewards/chosen": -0.2589899003505707, |
| "rewards/margins": 0.3519485890865326, |
| "rewards/rejected": -0.6109384298324585, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.99, |
| "step": 59, |
| "total_flos": 0.0, |
| "train_loss": 0.5997810929508532, |
| "train_runtime": 3304.1575, |
| "train_samples_per_second": 4.625, |
| "train_steps_per_second": 0.018 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 59, |
| "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 |
| } |
|
|
