{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 18,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.16666666666666666,
      "grad_norm": 222.82342420922615,
      "learning_rate": 5e-06,
      "loss": 2.4489,
      "step": 1
    },
    {
      "epoch": 0.3333333333333333,
      "grad_norm": 191.2213317228195,
      "learning_rate": 1e-05,
      "loss": 2.0952,
      "step": 2
    },
    {
      "epoch": 0.5,
      "grad_norm": 44.239187858987805,
      "learning_rate": 9.903926402016153e-06,
      "loss": 0.998,
      "step": 3
    },
    {
      "epoch": 0.6666666666666666,
      "grad_norm": 14.348099429987915,
      "learning_rate": 9.619397662556434e-06,
      "loss": 0.727,
      "step": 4
    },
    {
      "epoch": 0.8333333333333334,
      "grad_norm": 7.562754528894627,
      "learning_rate": 9.157348061512728e-06,
      "loss": 0.4583,
      "step": 5
    },
    {
      "epoch": 1.0,
      "grad_norm": 24.45138003404346,
      "learning_rate": 8.535533905932739e-06,
      "loss": 0.4519,
      "step": 6
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.3483733534812927,
      "eval_runtime": 1.0034,
      "eval_samples_per_second": 8.97,
      "eval_steps_per_second": 2.99,
      "step": 6
    },
    {
      "epoch": 1.1666666666666667,
      "grad_norm": 3.517997247609117,
      "learning_rate": 7.777851165098012e-06,
      "loss": 0.2755,
      "step": 7
    },
    {
      "epoch": 1.3333333333333333,
      "grad_norm": 2.744037064490017,
      "learning_rate": 6.913417161825449e-06,
      "loss": 0.2597,
      "step": 8
    },
    {
      "epoch": 1.5,
      "grad_norm": 1.2169458930476553,
      "learning_rate": 5.975451610080643e-06,
      "loss": 0.247,
      "step": 9
    },
    {
      "epoch": 1.6666666666666665,
      "grad_norm": 3.819506951093238,
      "learning_rate": 5e-06,
      "loss": 0.2465,
      "step": 10
    },
    {
      "epoch": 1.8333333333333335,
      "grad_norm": 1.4747090067383897,
      "learning_rate": 4.02454838991936e-06,
      "loss": 0.2356,
      "step": 11
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.6543936276406785,
      "learning_rate": 3.0865828381745515e-06,
      "loss": 0.1999,
      "step": 12
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.2685082256793976,
      "eval_runtime": 0.9843,
      "eval_samples_per_second": 9.143,
      "eval_steps_per_second": 3.048,
      "step": 12
    },
    {
      "epoch": 2.1666666666666665,
      "grad_norm": 1.0165748777270285,
      "learning_rate": 2.2221488349019903e-06,
      "loss": 0.1942,
      "step": 13
    },
    {
      "epoch": 2.3333333333333335,
      "grad_norm": 1.1243172670894819,
      "learning_rate": 1.4644660940672628e-06,
      "loss": 0.1792,
      "step": 14
    },
    {
      "epoch": 2.5,
      "grad_norm": 0.9298098825188076,
      "learning_rate": 8.426519384872733e-07,
      "loss": 0.2014,
      "step": 15
    },
    {
      "epoch": 2.6666666666666665,
      "grad_norm": 0.8728951381474357,
      "learning_rate": 3.8060233744356634e-07,
      "loss": 0.2173,
      "step": 16
    },
    {
      "epoch": 2.8333333333333335,
      "grad_norm": 0.6744276876185934,
      "learning_rate": 9.607359798384785e-08,
      "loss": 0.1815,
      "step": 17
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.6844816911131187,
      "learning_rate": 0.0,
      "loss": 0.2005,
      "step": 18
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.25620439648628235,
      "eval_runtime": 1.0135,
      "eval_samples_per_second": 8.88,
      "eval_steps_per_second": 2.96,
      "step": 18
    },
    {
      "epoch": 3.0,
      "step": 18,
      "total_flos": 3943670415360.0,
      "train_loss": 0.545412374039491,
      "train_runtime": 944.2174,
      "train_samples_per_second": 0.537,
      "train_steps_per_second": 0.019
    }
  ],
  "logging_steps": 1,
  "max_steps": 18,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "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": 3943670415360.0,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}