{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 8360,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.23923444976076555,
      "grad_norm": 1.421875,
      "learning_rate": 1.8803827751196172e-05,
      "loss": 0.192,
      "mean_token_accuracy": 0.9506647505760193,
      "step": 500
    },
    {
      "epoch": 0.4784688995215311,
      "grad_norm": 1.0546875,
      "learning_rate": 1.7607655502392345e-05,
      "loss": 0.0681,
      "mean_token_accuracy": 0.9816158946752548,
      "step": 1000
    },
    {
      "epoch": 0.7177033492822966,
      "grad_norm": 1.0546875,
      "learning_rate": 1.641148325358852e-05,
      "loss": 0.0608,
      "mean_token_accuracy": 0.98284204351902,
      "step": 1500
    },
    {
      "epoch": 0.9569377990430622,
      "grad_norm": 0.8515625,
      "learning_rate": 1.5215311004784689e-05,
      "loss": 0.0562,
      "mean_token_accuracy": 0.9840622389316559,
      "step": 2000
    },
    {
      "epoch": 1.1961722488038278,
      "grad_norm": 0.79296875,
      "learning_rate": 1.4019138755980863e-05,
      "loss": 0.0501,
      "mean_token_accuracy": 0.9850809412002564,
      "step": 2500
    },
    {
      "epoch": 1.4354066985645932,
      "grad_norm": 1.03125,
      "learning_rate": 1.2822966507177035e-05,
      "loss": 0.0468,
      "mean_token_accuracy": 0.9858997522592544,
      "step": 3000
    },
    {
      "epoch": 1.674641148325359,
      "grad_norm": 0.51171875,
      "learning_rate": 1.1626794258373206e-05,
      "loss": 0.0491,
      "mean_token_accuracy": 0.9851967771053314,
      "step": 3500
    },
    {
      "epoch": 1.9138755980861244,
      "grad_norm": 1.0859375,
      "learning_rate": 1.0430622009569378e-05,
      "loss": 0.0466,
      "mean_token_accuracy": 0.98594608771801,
      "step": 4000
    },
    {
      "epoch": 2.15311004784689,
      "grad_norm": 1.359375,
      "learning_rate": 9.234449760765552e-06,
      "loss": 0.0445,
      "mean_token_accuracy": 0.9863237723112106,
      "step": 4500
    },
    {
      "epoch": 2.3923444976076556,
      "grad_norm": 0.65234375,
      "learning_rate": 8.038277511961722e-06,
      "loss": 0.0416,
      "mean_token_accuracy": 0.9870117912292481,
      "step": 5000
    },
    {
      "epoch": 2.6315789473684212,
      "grad_norm": 0.765625,
      "learning_rate": 6.842105263157896e-06,
      "loss": 0.041,
      "mean_token_accuracy": 0.9870736453533172,
      "step": 5500
    },
    {
      "epoch": 2.8708133971291865,
      "grad_norm": 0.91015625,
      "learning_rate": 5.645933014354067e-06,
      "loss": 0.0412,
      "mean_token_accuracy": 0.9871674480438233,
      "step": 6000
    },
    {
      "epoch": 3.110047846889952,
      "grad_norm": 0.77734375,
      "learning_rate": 4.449760765550239e-06,
      "loss": 0.0412,
      "mean_token_accuracy": 0.9871292731761933,
      "step": 6500
    },
    {
      "epoch": 3.349282296650718,
      "grad_norm": 1.5078125,
      "learning_rate": 3.2535885167464117e-06,
      "loss": 0.0404,
      "mean_token_accuracy": 0.9873058786392211,
      "step": 7000
    },
    {
      "epoch": 3.588516746411483,
      "grad_norm": 1.15625,
      "learning_rate": 2.057416267942584e-06,
      "loss": 0.0392,
      "mean_token_accuracy": 0.9877065912485122,
      "step": 7500
    },
    {
      "epoch": 3.827751196172249,
      "grad_norm": 0.9921875,
      "learning_rate": 8.612440191387561e-07,
      "loss": 0.0389,
      "mean_token_accuracy": 0.9877618228197098,
      "step": 8000
    }
  ],
  "logging_steps": 500,
  "max_steps": 8360,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "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": 4.668357955416883e+17,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}