trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9984,
  "eval_steps": 100,
  "global_step": 39,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.03,
      "learning_rate": 1.25e-06,
      "logits/chosen": -2.6728761196136475,
      "logits/rejected": -2.6764090061187744,
      "logps/chosen": -261.5552978515625,
      "logps/rejected": -274.96466064453125,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.26,
      "learning_rate": 4.646121984004666e-06,
      "logits/chosen": -2.7910807132720947,
      "logits/rejected": -2.7170495986938477,
      "logps/chosen": -285.13818359375,
      "logps/rejected": -261.8572998046875,
      "loss": 0.6908,
      "rewards/accuracies": 0.5572916865348816,
      "rewards/chosen": 0.002809329889714718,
      "rewards/margins": 0.004932739306241274,
      "rewards/rejected": -0.002123409416526556,
      "step": 10
    },
    {
      "epoch": 0.51,
      "learning_rate": 2.835583164544139e-06,
      "logits/chosen": -2.7387890815734863,
      "logits/rejected": -2.6834359169006348,
      "logps/chosen": -278.64666748046875,
      "logps/rejected": -251.53060913085938,
      "loss": 0.6812,
      "rewards/accuracies": 0.7250000238418579,
      "rewards/chosen": 0.020782727748155594,
      "rewards/margins": 0.027223270386457443,
      "rewards/rejected": -0.006440544966608286,
      "step": 20
    },
    {
      "epoch": 0.77,
      "learning_rate": 7.723433775328385e-07,
      "logits/chosen": -2.7434194087982178,
      "logits/rejected": -2.673462152481079,
      "logps/chosen": -294.9715881347656,
      "logps/rejected": -267.701904296875,
      "loss": 0.6738,
      "rewards/accuracies": 0.7124999761581421,
      "rewards/chosen": 0.03247123584151268,
      "rewards/margins": 0.04708124324679375,
      "rewards/rejected": -0.014610009267926216,
      "step": 30
    },
    {
      "epoch": 1.0,
      "step": 39,
      "total_flos": 0.0,
      "train_loss": 0.6793185824002975,
      "train_runtime": 3271.4615,
      "train_samples_per_second": 3.057,
      "train_steps_per_second": 0.012
    }
  ],
  "logging_steps": 10,
  "max_steps": 39,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 100,
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}