trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.010141490144761593,
  "eval_steps": 100,
  "global_step": 31,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "grad_norm": 2.421875,
      "learning_rate": 1.25e-06,
      "logits/chosen": -2.3689165115356445,
      "logits/rejected": -2.3419089317321777,
      "logps/chosen": -304.96429443359375,
      "logps/rejected": -224.31954956054688,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.0,
      "grad_norm": 1.9296875,
      "learning_rate": 4.415111107797445e-06,
      "logits/chosen": -2.3774471282958984,
      "logits/rejected": -2.358837127685547,
      "logps/chosen": -267.6408386230469,
      "logps/rejected": -221.9726104736328,
      "loss": 0.6921,
      "rewards/accuracies": 0.5166666507720947,
      "rewards/chosen": 0.008927525021135807,
      "rewards/margins": 0.002250629710033536,
      "rewards/rejected": 0.006676895078271627,
      "step": 10
    },
    {
      "epoch": 0.01,
      "grad_norm": 2.203125,
      "learning_rate": 1.7829919182222752e-06,
      "logits/chosen": -2.4560706615448,
      "logits/rejected": -2.402303695678711,
      "logps/chosen": -265.12762451171875,
      "logps/rejected": -272.61566162109375,
      "loss": 0.6889,
      "rewards/accuracies": 0.5900000333786011,
      "rewards/chosen": 0.028245043009519577,
      "rewards/margins": 0.009032377041876316,
      "rewards/rejected": 0.019212666898965836,
      "step": 20
    },
    {
      "epoch": 0.01,
      "grad_norm": 2.375,
      "learning_rate": 1.6904105645142443e-08,
      "logits/chosen": -2.3814165592193604,
      "logits/rejected": -2.3470723628997803,
      "logps/chosen": -304.08697509765625,
      "logps/rejected": -281.0203552246094,
      "loss": 0.6844,
      "rewards/accuracies": 0.6299999952316284,
      "rewards/chosen": 0.03472686558961868,
      "rewards/margins": 0.019056813791394234,
      "rewards/rejected": 0.015670055523514748,
      "step": 30
    },
    {
      "epoch": 0.01,
      "step": 31,
      "total_flos": 0.0,
      "train_loss": 0.6888245363389293,
      "train_runtime": 439.9957,
      "train_samples_per_second": 1.389,
      "train_steps_per_second": 0.07
    }
  ],
  "logging_steps": 10,
  "max_steps": 31,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 5,
  "trial_name": null,
  "trial_params": null
}