trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.10196078431372549,
  "eval_steps": 100,
  "global_step": 13,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.01,
      "grad_norm": 0.5078125,
      "learning_rate": 2.5e-06,
      "logits/chosen": -2.3996620178222656,
      "logits/rejected": -2.364995002746582,
      "logps/chosen": -288.2684326171875,
      "logps/rejected": -260.847412109375,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.08,
      "grad_norm": 0.48046875,
      "learning_rate": 8.628481651367876e-07,
      "logits/chosen": -2.3491926193237305,
      "logits/rejected": -2.307647705078125,
      "logps/chosen": -264.3670349121094,
      "logps/rejected": -252.78759765625,
      "loss": 0.6887,
      "rewards/accuracies": 0.5314815640449524,
      "rewards/chosen": 0.012712381780147552,
      "rewards/margins": 0.01025653351098299,
      "rewards/rejected": 0.0024558501318097115,
      "step": 10
    },
    {
      "epoch": 0.1,
      "step": 13,
      "total_flos": 0.0,
      "train_loss": 0.6879739944751446,
      "train_runtime": 738.7034,
      "train_samples_per_second": 8.276,
      "train_steps_per_second": 0.018
    }
  ],
  "logging_steps": 10,
  "max_steps": 13,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 15,
  "trial_name": null,
  "trial_params": null
}