trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9936305732484076,
  "eval_steps": 500,
  "global_step": 39,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.025477707006369428,
      "grad_norm": 8.785225798788353,
      "learning_rate": 1.25e-07,
      "logits/chosen": -2.7347123622894287,
      "logits/rejected": -2.7646822929382324,
      "logps/chosen": -233.65045166015625,
      "logps/pi_response": -163.58248901367188,
      "logps/ref_response": -163.58248901367188,
      "logps/rejected": -281.40289306640625,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.25477707006369427,
      "grad_norm": 8.320152802039228,
      "learning_rate": 4.646121984004665e-07,
      "logits/chosen": -2.6830270290374756,
      "logits/rejected": -2.645677328109741,
      "logps/chosen": -206.8861541748047,
      "logps/pi_response": -136.25582885742188,
      "logps/ref_response": -138.557861328125,
      "logps/rejected": -228.14938354492188,
      "loss": 0.6858,
      "rewards/accuracies": 0.5729166865348816,
      "rewards/chosen": 0.0012997923186048865,
      "rewards/margins": 0.017590109258890152,
      "rewards/rejected": -0.016290316358208656,
      "step": 10
    },
    {
      "epoch": 0.5095541401273885,
      "grad_norm": 10.901858428989968,
      "learning_rate": 2.8355831645441387e-07,
      "logits/chosen": -2.7159199714660645,
      "logits/rejected": -2.6807026863098145,
      "logps/chosen": -240.71728515625,
      "logps/pi_response": -133.70616149902344,
      "logps/ref_response": -140.27857971191406,
      "logps/rejected": -249.140869140625,
      "loss": 0.629,
      "rewards/accuracies": 0.671875,
      "rewards/chosen": -0.08579646050930023,
      "rewards/margins": 0.17118123173713684,
      "rewards/rejected": -0.25697770714759827,
      "step": 20
    },
    {
      "epoch": 0.7643312101910829,
      "grad_norm": 12.11154150001121,
      "learning_rate": 7.723433775328384e-08,
      "logits/chosen": -2.727022409439087,
      "logits/rejected": -2.7074286937713623,
      "logps/chosen": -257.46820068359375,
      "logps/pi_response": -148.09524536132812,
      "logps/ref_response": -146.3412628173828,
      "logps/rejected": -303.86480712890625,
      "loss": 0.6083,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.29648229479789734,
      "rewards/margins": 0.3200515806674957,
      "rewards/rejected": -0.6165338754653931,
      "step": 30
    },
    {
      "epoch": 0.9936305732484076,
      "step": 39,
      "total_flos": 0.0,
      "train_loss": 0.6268080564645621,
      "train_runtime": 1801.8213,
      "train_samples_per_second": 5.55,
      "train_steps_per_second": 0.022
    }
  ],
  "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": 8,
  "trial_name": null,
  "trial_params": null
}