{
  "best_metric": 0.98828125,
  "best_model_checkpoint": "./model/checkpoint-140",
  "epoch": 20.0,
  "eval_steps": 500,
  "global_step": 400,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5,
      "learning_rate": 5e-06,
      "loss": 0.0465,
      "step": 10
    },
    {
      "epoch": 1.0,
      "learning_rate": 1e-05,
      "loss": 0.041,
      "step": 20
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.984375,
      "eval_loss": 0.11365039646625519,
      "eval_runtime": 4.8919,
      "eval_samples_per_second": 52.332,
      "eval_steps_per_second": 6.541,
      "step": 20
    },
    {
      "epoch": 1.5,
      "learning_rate": 1.5000000000000002e-05,
      "loss": 0.0333,
      "step": 30
    },
    {
      "epoch": 2.0,
      "learning_rate": 2e-05,
      "loss": 0.037,
      "step": 40
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.98046875,
      "eval_loss": 0.12528091669082642,
      "eval_runtime": 3.9769,
      "eval_samples_per_second": 64.372,
      "eval_steps_per_second": 8.046,
      "step": 40
    },
    {
      "epoch": 2.5,
      "learning_rate": 1.9444444444444445e-05,
      "loss": 0.0342,
      "step": 50
    },
    {
      "epoch": 3.0,
      "learning_rate": 1.888888888888889e-05,
      "loss": 0.0367,
      "step": 60
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.98046875,
      "eval_loss": 0.12140656262636185,
      "eval_runtime": 4.3164,
      "eval_samples_per_second": 59.309,
      "eval_steps_per_second": 7.414,
      "step": 60
    },
    {
      "epoch": 3.5,
      "learning_rate": 1.8333333333333333e-05,
      "loss": 0.0301,
      "step": 70
    },
    {
      "epoch": 4.0,
      "learning_rate": 1.7777777777777777e-05,
      "loss": 0.0298,
      "step": 80
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.98046875,
      "eval_loss": 0.08230899274349213,
      "eval_runtime": 4.114,
      "eval_samples_per_second": 62.227,
      "eval_steps_per_second": 7.778,
      "step": 80
    },
    {
      "epoch": 4.5,
      "learning_rate": 1.7222222222222224e-05,
      "loss": 0.0186,
      "step": 90
    },
    {
      "epoch": 5.0,
      "learning_rate": 1.6666666666666667e-05,
      "loss": 0.0165,
      "step": 100
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.98046875,
      "eval_loss": 0.08846013247966766,
      "eval_runtime": 4.7826,
      "eval_samples_per_second": 53.527,
      "eval_steps_per_second": 6.691,
      "step": 100
    },
    {
      "epoch": 5.5,
      "learning_rate": 1.6111111111111115e-05,
      "loss": 0.0212,
      "step": 110
    },
    {
      "epoch": 6.0,
      "learning_rate": 1.555555555555556e-05,
      "loss": 0.0142,
      "step": 120
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9765625,
      "eval_loss": 0.07800224423408508,
      "eval_runtime": 3.9449,
      "eval_samples_per_second": 64.894,
      "eval_steps_per_second": 8.112,
      "step": 120
    },
    {
      "epoch": 6.5,
      "learning_rate": 1.5000000000000002e-05,
      "loss": 0.0122,
      "step": 130
    },
    {
      "epoch": 7.0,
      "learning_rate": 1.4444444444444446e-05,
      "loss": 0.0145,
      "step": 140
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.98828125,
      "eval_loss": 0.06642253696918488,
      "eval_runtime": 4.2638,
      "eval_samples_per_second": 60.04,
      "eval_steps_per_second": 7.505,
      "step": 140
    },
    {
      "epoch": 7.5,
      "learning_rate": 1.388888888888889e-05,
      "loss": 0.0165,
      "step": 150
    },
    {
      "epoch": 8.0,
      "learning_rate": 1.3333333333333333e-05,
      "loss": 0.0105,
      "step": 160
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.984375,
      "eval_loss": 0.05820595473051071,
      "eval_runtime": 4.7261,
      "eval_samples_per_second": 54.167,
      "eval_steps_per_second": 6.771,
      "step": 160
    },
    {
      "epoch": 8.5,
      "learning_rate": 1.2777777777777777e-05,
      "loss": 0.009,
      "step": 170
    },
    {
      "epoch": 9.0,
      "learning_rate": 1.2222222222222224e-05,
      "loss": 0.0136,
      "step": 180
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.984375,
      "eval_loss": 0.06151523441076279,
      "eval_runtime": 4.7568,
      "eval_samples_per_second": 53.818,
      "eval_steps_per_second": 6.727,
      "step": 180
    },
    {
      "epoch": 9.5,
      "learning_rate": 1.1666666666666668e-05,
      "loss": 0.0095,
      "step": 190
    },
    {
      "epoch": 10.0,
      "learning_rate": 1.1111111111111113e-05,
      "loss": 0.0088,
      "step": 200
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.984375,
      "eval_loss": 0.051664501428604126,
      "eval_runtime": 4.0071,
      "eval_samples_per_second": 63.886,
      "eval_steps_per_second": 7.986,
      "step": 200
    },
    {
      "epoch": 10.5,
      "learning_rate": 1.0555555555555557e-05,
      "loss": 0.0087,
      "step": 210
    },
    {
      "epoch": 11.0,
      "learning_rate": 1e-05,
      "loss": 0.0048,
      "step": 220
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.98828125,
      "eval_loss": 0.04740528017282486,
      "eval_runtime": 4.776,
      "eval_samples_per_second": 53.601,
      "eval_steps_per_second": 6.7,
      "step": 220
    },
    {
      "epoch": 11.5,
      "learning_rate": 9.444444444444445e-06,
      "loss": 0.0056,
      "step": 230
    },
    {
      "epoch": 12.0,
      "learning_rate": 8.888888888888888e-06,
      "loss": 0.004,
      "step": 240
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.984375,
      "eval_loss": 0.04649289697408676,
      "eval_runtime": 3.9995,
      "eval_samples_per_second": 64.008,
      "eval_steps_per_second": 8.001,
      "step": 240
    },
    {
      "epoch": 12.5,
      "learning_rate": 8.333333333333334e-06,
      "loss": 0.007,
      "step": 250
    },
    {
      "epoch": 13.0,
      "learning_rate": 7.77777777777778e-06,
      "loss": 0.004,
      "step": 260
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.984375,
      "eval_loss": 0.05019224062561989,
      "eval_runtime": 4.2397,
      "eval_samples_per_second": 60.382,
      "eval_steps_per_second": 7.548,
      "step": 260
    },
    {
      "epoch": 13.5,
      "learning_rate": 7.222222222222223e-06,
      "loss": 0.0039,
      "step": 270
    },
    {
      "epoch": 14.0,
      "learning_rate": 6.666666666666667e-06,
      "loss": 0.0041,
      "step": 280
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.98828125,
      "eval_loss": 0.03817906230688095,
      "eval_runtime": 4.303,
      "eval_samples_per_second": 59.493,
      "eval_steps_per_second": 7.437,
      "step": 280
    },
    {
      "epoch": 14.5,
      "learning_rate": 6.111111111111112e-06,
      "loss": 0.0041,
      "step": 290
    },
    {
      "epoch": 15.0,
      "learning_rate": 5.555555555555557e-06,
      "loss": 0.0031,
      "step": 300
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.984375,
      "eval_loss": 0.04823793098330498,
      "eval_runtime": 4.2376,
      "eval_samples_per_second": 60.411,
      "eval_steps_per_second": 7.551,
      "step": 300
    },
    {
      "epoch": 15.5,
      "learning_rate": 5e-06,
      "loss": 0.0061,
      "step": 310
    },
    {
      "epoch": 16.0,
      "learning_rate": 4.444444444444444e-06,
      "loss": 0.0072,
      "step": 320
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.98828125,
      "eval_loss": 0.04395627975463867,
      "eval_runtime": 4.863,
      "eval_samples_per_second": 52.642,
      "eval_steps_per_second": 6.58,
      "step": 320
    },
    {
      "epoch": 16.5,
      "learning_rate": 3.88888888888889e-06,
      "loss": 0.0032,
      "step": 330
    },
    {
      "epoch": 17.0,
      "learning_rate": 3.3333333333333333e-06,
      "loss": 0.004,
      "step": 340
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.984375,
      "eval_loss": 0.04199722409248352,
      "eval_runtime": 4.0297,
      "eval_samples_per_second": 63.528,
      "eval_steps_per_second": 7.941,
      "step": 340
    },
    {
      "epoch": 17.5,
      "learning_rate": 2.7777777777777783e-06,
      "loss": 0.0033,
      "step": 350
    },
    {
      "epoch": 18.0,
      "learning_rate": 2.222222222222222e-06,
      "loss": 0.0026,
      "step": 360
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.98828125,
      "eval_loss": 0.03974057734012604,
      "eval_runtime": 4.1185,
      "eval_samples_per_second": 62.158,
      "eval_steps_per_second": 7.77,
      "step": 360
    },
    {
      "epoch": 18.5,
      "learning_rate": 1.6666666666666667e-06,
      "loss": 0.0043,
      "step": 370
    },
    {
      "epoch": 19.0,
      "learning_rate": 1.111111111111111e-06,
      "loss": 0.0041,
      "step": 380
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.98828125,
      "eval_loss": 0.04067450016736984,
      "eval_runtime": 4.7682,
      "eval_samples_per_second": 53.688,
      "eval_steps_per_second": 6.711,
      "step": 380
    },
    {
      "epoch": 19.5,
      "learning_rate": 5.555555555555555e-07,
      "loss": 0.0025,
      "step": 390
    },
    {
      "epoch": 20.0,
      "learning_rate": 0.0,
      "loss": 0.004,
      "step": 400
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.98828125,
      "eval_loss": 0.040177565068006516,
      "eval_runtime": 3.9732,
      "eval_samples_per_second": 64.432,
      "eval_steps_per_second": 8.054,
      "step": 400
    }
  ],
  "logging_steps": 10,
  "max_steps": 400,
  "num_train_epochs": 20,
  "save_steps": 500,
  "total_flos": 2.1859568857379635e+18,
  "trial_name": null,
  "trial_params": null
}