trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 1.0,
  "eval_steps": 500,
  "global_step": 274,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.18248175182481752,
      "grad_norm": 4.739284992218018,
      "learning_rate": 9.803954791481239e-08,
      "logits/chosen": 8.744650840759277,
      "logits/rejected": 8.797840118408203,
      "logps/chosen": -107.8392562866211,
      "logps/rejected": -133.73532104492188,
      "loss": 0.0907,
      "rewards/accuracies": 0.9800000190734863,
      "rewards/chosen": 2.932006597518921,
      "rewards/margins": 3.9326796531677246,
      "rewards/rejected": -1.0006736516952515,
      "step": 50
    },
    {
      "epoch": 0.36496350364963503,
      "grad_norm": 3.2815303802490234,
      "learning_rate": 8.03112705483319e-08,
      "logits/chosen": 8.751543998718262,
      "logits/rejected": 8.796612739562988,
      "logps/chosen": -105.77054595947266,
      "logps/rejected": -136.58460998535156,
      "loss": 0.0648,
      "rewards/accuracies": 0.9900000095367432,
      "rewards/chosen": 2.9739089012145996,
      "rewards/margins": 4.341518402099609,
      "rewards/rejected": -1.3676095008850098,
      "step": 100
    },
    {
      "epoch": 0.5474452554744526,
      "grad_norm": 4.354165554046631,
      "learning_rate": 5.063851773579869e-08,
      "logits/chosen": 8.725605964660645,
      "logits/rejected": 8.783352851867676,
      "logps/chosen": -103.66476440429688,
      "logps/rejected": -137.76138305664062,
      "loss": 0.0504,
      "rewards/accuracies": 0.9950000047683716,
      "rewards/chosen": 3.048506736755371,
      "rewards/margins": 4.554016590118408,
      "rewards/rejected": -1.505509614944458,
      "step": 150
    },
    {
      "epoch": 0.7299270072992701,
      "grad_norm": 0.6422143578529358,
      "learning_rate": 2.0714150283590258e-08,
      "logits/chosen": 8.768350601196289,
      "logits/rejected": 8.77940559387207,
      "logps/chosen": -103.38182830810547,
      "logps/rejected": -142.48721313476562,
      "loss": 0.0406,
      "rewards/accuracies": 1.0,
      "rewards/chosen": 3.227271795272827,
      "rewards/margins": 4.983362197875977,
      "rewards/rejected": -1.7560904026031494,
      "step": 200
    },
    {
      "epoch": 0.9124087591240876,
      "grad_norm": 0.6735481023788452,
      "learning_rate": 2.330180397253473e-09,
      "logits/chosen": 8.723614692687988,
      "logits/rejected": 8.806389808654785,
      "logps/chosen": -101.47320556640625,
      "logps/rejected": -134.9160919189453,
      "loss": 0.0573,
      "rewards/accuracies": 0.9850000143051147,
      "rewards/chosen": 3.1267037391662598,
      "rewards/margins": 4.904569625854492,
      "rewards/rejected": -1.7778658866882324,
      "step": 250
    }
  ],
  "logging_steps": 50,
  "max_steps": 274,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 2,
  "trial_name": null,
  "trial_params": null
}