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": 0.032,
      "grad_norm": 8.94253801855711,
      "learning_rate": 1.25e-07,
      "logits/chosen": -2.838801383972168,
      "logits/rejected": -2.8086488246917725,
      "logps/chosen": -257.9438171386719,
      "logps/pi_response": -93.52229309082031,
      "logps/ref_response": -93.52229309082031,
      "logps/rejected": -216.83636474609375,
      "loss": 0.6931,
      "rewards/accuracies": 0.0,
      "rewards/chosen": 0.0,
      "rewards/margins": 0.0,
      "rewards/rejected": 0.0,
      "step": 1
    },
    {
      "epoch": 0.32,
      "grad_norm": 7.275195524114833,
      "learning_rate": 4.415111107797445e-07,
      "logits/chosen": -2.8140869140625,
      "logits/rejected": -2.8060314655303955,
      "logps/chosen": -246.06365966796875,
      "logps/pi_response": -74.96839141845703,
      "logps/ref_response": -74.20439147949219,
      "logps/rejected": -178.8218536376953,
      "loss": 0.6863,
      "rewards/accuracies": 0.5972222089767456,
      "rewards/chosen": 0.016777267679572105,
      "rewards/margins": 0.01401255652308464,
      "rewards/rejected": 0.0027647078968584538,
      "step": 10
    },
    {
      "epoch": 0.64,
      "grad_norm": 6.738990682860812,
      "learning_rate": 1.782991918222275e-07,
      "logits/chosen": -2.693695545196533,
      "logits/rejected": -2.6687941551208496,
      "logps/chosen": -225.8972625732422,
      "logps/pi_response": -78.36871337890625,
      "logps/ref_response": -68.63687896728516,
      "logps/rejected": -179.7605743408203,
      "loss": 0.6544,
      "rewards/accuracies": 0.731249988079071,
      "rewards/chosen": 0.040796808898448944,
      "rewards/margins": 0.08692635595798492,
      "rewards/rejected": -0.04612954705953598,
      "step": 20
    },
    {
      "epoch": 0.96,
      "grad_norm": 6.623989399704757,
      "learning_rate": 1.690410564514244e-09,
      "logits/chosen": -2.707422971725464,
      "logits/rejected": -2.6639552116394043,
      "logps/chosen": -243.7229461669922,
      "logps/pi_response": -95.37244415283203,
      "logps/ref_response": -70.27430725097656,
      "logps/rejected": -177.5935821533203,
      "loss": 0.6407,
      "rewards/accuracies": 0.715624988079071,
      "rewards/chosen": -0.027013743296265602,
      "rewards/margins": 0.13096943497657776,
      "rewards/rejected": -0.15798316895961761,
      "step": 30
    },
    {
      "epoch": 0.992,
      "step": 31,
      "total_flos": 0.0,
      "train_loss": 0.6603157808703761,
      "train_runtime": 1409.3993,
      "train_samples_per_second": 5.676,
      "train_steps_per_second": 0.022
    }
  ],
  "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
}