trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9842931937172775,
  "eval_steps": 100,
  "global_step": 47,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "grad_norm": 5.378226862681596,
      "learning_rate": 1e-07,
      "logits/chosen": 0.09814453125,
      "logits/rejected": 1.1875,
      "logps/chosen": -536.0,
      "logps/rejected": -432.0,
      "loss": 1.853,
      "rewards/accuracies": 0.484375,
      "rewards/chosen": -0.06787109375,
      "rewards/margins": -0.08935546875,
      "rewards/rejected": 0.021240234375,
      "step": 1
    },
    {
      "epoch": 0.21,
      "grad_norm": 4.964034862946189,
      "learning_rate": 4.82718437161051e-07,
      "logits/chosen": 0.228515625,
      "logits/rejected": 1.1484375,
      "logps/chosen": -516.0,
      "logps/rejected": -432.0,
      "loss": 1.8421,
      "rewards/accuracies": 0.5008680820465088,
      "rewards/chosen": -0.208984375,
      "rewards/margins": -0.1328125,
      "rewards/rejected": -0.07666015625,
      "step": 10
    },
    {
      "epoch": 0.42,
      "grad_norm": 5.274181848522071,
      "learning_rate": 3.584709347793895e-07,
      "logits/chosen": 0.130859375,
      "logits/rejected": 1.125,
      "logps/chosen": -512.0,
      "logps/rejected": -430.0,
      "loss": 1.7766,
      "rewards/accuracies": 0.48515623807907104,
      "rewards/chosen": -0.251953125,
      "rewards/margins": -0.130859375,
      "rewards/rejected": -0.12060546875,
      "step": 20
    },
    {
      "epoch": 0.63,
      "grad_norm": 4.641365779199991,
      "learning_rate": 1.763112063972739e-07,
      "logits/chosen": 0.2451171875,
      "logits/rejected": 1.1328125,
      "logps/chosen": -508.0,
      "logps/rejected": -428.0,
      "loss": 1.8372,
      "rewards/accuracies": 0.49296873807907104,
      "rewards/chosen": -0.06591796875,
      "rewards/margins": -0.04150390625,
      "rewards/rejected": -0.024169921875,
      "step": 30
    },
    {
      "epoch": 0.84,
      "grad_norm": 4.815119613081631,
      "learning_rate": 3.349364905389032e-08,
      "logits/chosen": 0.2177734375,
      "logits/rejected": 1.078125,
      "logps/chosen": -492.0,
      "logps/rejected": -422.0,
      "loss": 1.6402,
      "rewards/accuracies": 0.5132812261581421,
      "rewards/chosen": 0.185546875,
      "rewards/margins": 0.14453125,
      "rewards/rejected": 0.0419921875,
      "step": 40
    },
    {
      "epoch": 0.98,
      "step": 47,
      "total_flos": 0.0,
      "train_loss": 1.778102144281915,
      "train_runtime": 2602.6066,
      "train_samples_per_second": 2.348,
      "train_steps_per_second": 0.018
    }
  ],
  "logging_steps": 10,
  "max_steps": 47,
  "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
}