trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 48,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "learning_rate": 1e-07,
      "logits/chosen": -2.714864492416382,
      "logits/rejected": -2.751317024230957,
      "logps/chosen": -236.86740112304688,
      "logps/pi_response": -114.0628662109375,
      "logps/ref_response": -114.0628662109375,
      "logps/rejected": -478.6082763671875,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.21,
      "learning_rate": 4.83504027183137e-07,
      "logits/chosen": -2.657212257385254,
      "logits/rejected": -2.586080312728882,
      "logps/chosen": -296.349365234375,
      "logps/pi_response": -145.713623046875,
      "logps/ref_response": -136.1793212890625,
      "logps/rejected": -501.9734802246094,
      "loss": 0.643,
      "rewards/accuracies": 0.6006944179534912,
      "rewards/chosen": -0.26413747668266296,
      "rewards/margins": 0.26705610752105713,
      "rewards/rejected": -0.5311936140060425,
      "step": 10
    },
    {
      "epoch": 0.42,
      "learning_rate": 3.643105808261596e-07,
      "logits/chosen": -2.5398590564727783,
      "logits/rejected": -2.4759039878845215,
      "logps/chosen": -350.80548095703125,
      "logps/pi_response": -158.80862426757812,
      "logps/ref_response": -116.90655517578125,
      "logps/rejected": -726.4019775390625,
      "loss": 0.5769,
      "rewards/accuracies": 0.7593749761581421,
      "rewards/chosen": -1.0708297491073608,
      "rewards/margins": 1.4764198064804077,
      "rewards/rejected": -2.5472495555877686,
      "step": 20
    },
    {
      "epoch": 0.62,
      "learning_rate": 1.8676665440207977e-07,
      "logits/chosen": -2.436058282852173,
      "logits/rejected": -2.3596558570861816,
      "logps/chosen": -376.2472229003906,
      "logps/pi_response": -170.85630798339844,
      "logps/ref_response": -132.64312744140625,
      "logps/rejected": -646.6350708007812,
      "loss": 0.5035,
      "rewards/accuracies": 0.7281249761581421,
      "rewards/chosen": -0.9769057035446167,
      "rewards/margins": 0.9189132452011108,
      "rewards/rejected": -1.895819067955017,
      "step": 30
    },
    {
      "epoch": 0.83,
      "learning_rate": 4.1500545527530544e-08,
      "logits/chosen": -2.4626142978668213,
      "logits/rejected": -2.351714849472046,
      "logps/chosen": -363.10650634765625,
      "logps/pi_response": -163.067626953125,
      "logps/ref_response": -126.86649322509766,
      "logps/rejected": -648.2885131835938,
      "loss": 0.4795,
      "rewards/accuracies": 0.793749988079071,
      "rewards/chosen": -0.845011830329895,
      "rewards/margins": 0.9437265396118164,
      "rewards/rejected": -1.788738489151001,
      "step": 40
    },
    {
      "epoch": 1.0,
      "step": 48,
      "total_flos": 0.0,
      "train_loss": 0.5470561385154724,
      "train_runtime": 2831.8023,
      "train_samples_per_second": 4.318,
      "train_steps_per_second": 0.017
    }
  ],
  "logging_steps": 10,
  "max_steps": 48,
  "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
}