trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.992,
  "eval_steps": 500,
  "global_step": 31,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.03,
      "learning_rate": 1.25e-07,
      "logits/chosen": -2.4519619941711426,
      "logits/rejected": -2.4373748302459717,
      "logps/chosen": -313.0509033203125,
      "logps/pi_response": -118.84771728515625,
      "logps/ref_response": -118.84771728515625,
      "logps/rejected": -805.4427490234375,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.32,
      "learning_rate": 4.415111107797445e-07,
      "logits/chosen": -2.326218843460083,
      "logits/rejected": -2.284130811691284,
      "logps/chosen": -410.3197326660156,
      "logps/pi_response": -151.0199432373047,
      "logps/ref_response": -117.56155395507812,
      "logps/rejected": -908.8378295898438,
      "loss": 0.6621,
      "rewards/accuracies": 0.6493055820465088,
      "rewards/chosen": -0.6402885913848877,
      "rewards/margins": 0.8867037892341614,
      "rewards/rejected": -1.5269925594329834,
      "step": 10
    },
    {
      "epoch": 0.64,
      "learning_rate": 1.782991918222275e-07,
      "logits/chosen": -2.226139545440674,
      "logits/rejected": -2.1810803413391113,
      "logps/chosen": -529.3506469726562,
      "logps/pi_response": -162.92213439941406,
      "logps/ref_response": -109.0821533203125,
      "logps/rejected": -1107.5057373046875,
      "loss": 0.6638,
      "rewards/accuracies": 0.737500011920929,
      "rewards/chosen": -1.5208525657653809,
      "rewards/margins": 1.4993822574615479,
      "rewards/rejected": -3.0202348232269287,
      "step": 20
    },
    {
      "epoch": 0.96,
      "learning_rate": 1.690410564514244e-09,
      "logits/chosen": -2.2474887371063232,
      "logits/rejected": -2.211569309234619,
      "logps/chosen": -406.5466613769531,
      "logps/pi_response": -142.61892700195312,
      "logps/ref_response": -118.6996078491211,
      "logps/rejected": -982.9817504882812,
      "loss": 0.5018,
      "rewards/accuracies": 0.7906249761581421,
      "rewards/chosen": -0.5995205044746399,
      "rewards/margins": 0.9062793850898743,
      "rewards/rejected": -1.5057997703552246,
      "step": 30
    },
    {
      "epoch": 0.99,
      "step": 31,
      "total_flos": 0.0,
      "train_loss": 0.6071660441737021,
      "train_runtime": 1788.9782,
      "train_samples_per_second": 4.472,
      "train_steps_per_second": 0.017
    }
  ],
  "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": 8,
  "trial_name": null,
  "trial_params": null
}