trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 100,
  "global_step": 38,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.03,
      "grad_norm": 4.6681179725575745,
      "learning_rate": 1.25e-07,
      "logits/chosen": -0.3981982469558716,
      "logits/rejected": 0.007589429616928101,
      "logps/chosen": -253.3463134765625,
      "logps/rejected": -671.2213745117188,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.26,
      "grad_norm": 4.768585444261873,
      "learning_rate": 4.6255428393240354e-07,
      "logits/chosen": -0.5132235884666443,
      "logits/rejected": -0.01092798262834549,
      "logps/chosen": -188.30332946777344,
      "logps/rejected": -627.010986328125,
      "loss": 0.6914,
      "rewards/accuracies": 0.625,
      "rewards/chosen": 0.0005456734797917306,
      "rewards/margins": 0.004515086766332388,
      "rewards/rejected": -0.003969413228332996,
      "step": 10
    },
    {
      "epoch": 0.53,
      "grad_norm": 4.366842926411694,
      "learning_rate": 2.730670898658255e-07,
      "logits/chosen": -0.6587502360343933,
      "logits/rejected": -0.0813610702753067,
      "logps/chosen": -157.48902893066406,
      "logps/rejected": -626.0028686523438,
      "loss": 0.6775,
      "rewards/accuracies": 0.8812500238418579,
      "rewards/chosen": -0.002243603579699993,
      "rewards/margins": 0.03073681890964508,
      "rewards/rejected": -0.03298041969537735,
      "step": 20
    },
    {
      "epoch": 0.79,
      "grad_norm": 4.881380621263082,
      "learning_rate": 6.524777069483525e-08,
      "logits/chosen": -0.49691787362098694,
      "logits/rejected": -0.0660545602440834,
      "logps/chosen": -177.71261596679688,
      "logps/rejected": -657.294677734375,
      "loss": 0.6583,
      "rewards/accuracies": 0.918749988079071,
      "rewards/chosen": -0.006053019780665636,
      "rewards/margins": 0.07823501527309418,
      "rewards/rejected": -0.08428805321455002,
      "step": 30
    },
    {
      "epoch": 1.0,
      "step": 38,
      "total_flos": 0.0,
      "train_loss": 0.6694252616480777,
      "train_runtime": 502.7814,
      "train_samples_per_second": 4.799,
      "train_steps_per_second": 0.076
    }
  ],
  "logging_steps": 10,
  "max_steps": 38,
  "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
}