{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 15.0,
  "global_step": 19800,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.38,
      "learning_rate": 1.9494949494949496e-05,
      "loss": 4.4759,
      "step": 500
    },
    {
      "epoch": 0.76,
      "learning_rate": 1.8989898989898993e-05,
      "loss": 3.4666,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5767824649810791,
      "eval_loss": 2.3355202674865723,
      "eval_runtime": 2.6806,
      "eval_samples_per_second": 408.112,
      "eval_steps_per_second": 51.107,
      "step": 1320
    },
    {
      "epoch": 1.14,
      "learning_rate": 1.8484848484848487e-05,
      "loss": 2.6641,
      "step": 1500
    },
    {
      "epoch": 1.52,
      "learning_rate": 1.797979797979798e-05,
      "loss": 1.9312,
      "step": 2000
    },
    {
      "epoch": 1.89,
      "learning_rate": 1.7474747474747475e-05,
      "loss": 1.5293,
      "step": 2500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.8144423961639404,
      "eval_loss": 1.1118409633636475,
      "eval_runtime": 2.7222,
      "eval_samples_per_second": 401.874,
      "eval_steps_per_second": 50.326,
      "step": 2640
    },
    {
      "epoch": 2.27,
      "learning_rate": 1.6969696969696972e-05,
      "loss": 1.0511,
      "step": 3000
    },
    {
      "epoch": 2.65,
      "learning_rate": 1.6464646464646466e-05,
      "loss": 0.8031,
      "step": 3500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.880255937576294,
      "eval_loss": 0.6362389922142029,
      "eval_runtime": 2.657,
      "eval_samples_per_second": 411.738,
      "eval_steps_per_second": 51.561,
      "step": 3960
    },
    {
      "epoch": 3.03,
      "learning_rate": 1.595959595959596e-05,
      "loss": 0.5864,
      "step": 4000
    },
    {
      "epoch": 3.41,
      "learning_rate": 1.5454545454545454e-05,
      "loss": 0.3632,
      "step": 4500
    },
    {
      "epoch": 3.79,
      "learning_rate": 1.4949494949494952e-05,
      "loss": 0.2985,
      "step": 5000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.8957952260971069,
      "eval_loss": 0.5118863582611084,
      "eval_runtime": 2.6864,
      "eval_samples_per_second": 407.231,
      "eval_steps_per_second": 50.997,
      "step": 5280
    },
    {
      "epoch": 4.17,
      "learning_rate": 1.4444444444444446e-05,
      "loss": 0.2188,
      "step": 5500
    },
    {
      "epoch": 4.55,
      "learning_rate": 1.3939393939393942e-05,
      "loss": 0.1478,
      "step": 6000
    },
    {
      "epoch": 4.92,
      "learning_rate": 1.3434343434343436e-05,
      "loss": 0.1284,
      "step": 6500
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.8930529952049255,
      "eval_loss": 0.5023291707038879,
      "eval_runtime": 2.7918,
      "eval_samples_per_second": 391.858,
      "eval_steps_per_second": 49.072,
      "step": 6600
    },
    {
      "epoch": 5.3,
      "learning_rate": 1.2929292929292931e-05,
      "loss": 0.0608,
      "step": 7000
    },
    {
      "epoch": 5.68,
      "learning_rate": 1.2424242424242425e-05,
      "loss": 0.0842,
      "step": 7500
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9021937847137451,
      "eval_loss": 0.5246109962463379,
      "eval_runtime": 2.6464,
      "eval_samples_per_second": 413.391,
      "eval_steps_per_second": 51.768,
      "step": 7920
    },
    {
      "epoch": 6.06,
      "learning_rate": 1.191919191919192e-05,
      "loss": 0.064,
      "step": 8000
    },
    {
      "epoch": 6.44,
      "learning_rate": 1.1414141414141415e-05,
      "loss": 0.0461,
      "step": 8500
    },
    {
      "epoch": 6.82,
      "learning_rate": 1.0909090909090909e-05,
      "loss": 0.0414,
      "step": 9000
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9012796878814697,
      "eval_loss": 0.5580916404724121,
      "eval_runtime": 2.6062,
      "eval_samples_per_second": 419.767,
      "eval_steps_per_second": 52.567,
      "step": 9240
    },
    {
      "epoch": 7.2,
      "learning_rate": 1.0404040404040405e-05,
      "loss": 0.0476,
      "step": 9500
    },
    {
      "epoch": 7.58,
      "learning_rate": 9.8989898989899e-06,
      "loss": 0.0333,
      "step": 10000
    },
    {
      "epoch": 7.95,
      "learning_rate": 9.393939393939396e-06,
      "loss": 0.0372,
      "step": 10500
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9003656506538391,
      "eval_loss": 0.5721055269241333,
      "eval_runtime": 3.0748,
      "eval_samples_per_second": 355.794,
      "eval_steps_per_second": 44.556,
      "step": 10560
    },
    {
      "epoch": 8.33,
      "learning_rate": 8.888888888888888e-06,
      "loss": 0.0276,
      "step": 11000
    },
    {
      "epoch": 8.71,
      "learning_rate": 8.383838383838384e-06,
      "loss": 0.0292,
      "step": 11500
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9140768051147461,
      "eval_loss": 0.5468941926956177,
      "eval_runtime": 3.0391,
      "eval_samples_per_second": 359.977,
      "eval_steps_per_second": 45.079,
      "step": 11880
    },
    {
      "epoch": 9.09,
      "learning_rate": 7.87878787878788e-06,
      "loss": 0.0298,
      "step": 12000
    },
    {
      "epoch": 9.47,
      "learning_rate": 7.373737373737374e-06,
      "loss": 0.0209,
      "step": 12500
    },
    {
      "epoch": 9.85,
      "learning_rate": 6.868686868686869e-06,
      "loss": 0.0257,
      "step": 13000
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9058501124382019,
      "eval_loss": 0.5871404409408569,
      "eval_runtime": 2.8242,
      "eval_samples_per_second": 387.373,
      "eval_steps_per_second": 48.51,
      "step": 13200
    },
    {
      "epoch": 10.23,
      "learning_rate": 6.363636363636364e-06,
      "loss": 0.0205,
      "step": 13500
    },
    {
      "epoch": 10.61,
      "learning_rate": 5.858585858585859e-06,
      "loss": 0.0145,
      "step": 14000
    },
    {
      "epoch": 10.98,
      "learning_rate": 5.353535353535354e-06,
      "loss": 0.0189,
      "step": 14500
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.9049360156059265,
      "eval_loss": 0.6180957555770874,
      "eval_runtime": 2.9518,
      "eval_samples_per_second": 370.627,
      "eval_steps_per_second": 46.413,
      "step": 14520
    },
    {
      "epoch": 11.36,
      "learning_rate": 4.848484848484849e-06,
      "loss": 0.0117,
      "step": 15000
    },
    {
      "epoch": 11.74,
      "learning_rate": 4.343434343434344e-06,
      "loss": 0.0104,
      "step": 15500
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9067641496658325,
      "eval_loss": 0.618419349193573,
      "eval_runtime": 2.8761,
      "eval_samples_per_second": 380.381,
      "eval_steps_per_second": 47.635,
      "step": 15840
    },
    {
      "epoch": 12.12,
      "learning_rate": 3.8383838383838385e-06,
      "loss": 0.0155,
      "step": 16000
    },
    {
      "epoch": 12.5,
      "learning_rate": 3.3333333333333333e-06,
      "loss": 0.0087,
      "step": 16500
    },
    {
      "epoch": 12.88,
      "learning_rate": 2.8282828282828286e-06,
      "loss": 0.009,
      "step": 17000
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.9049360156059265,
      "eval_loss": 0.6013244986534119,
      "eval_runtime": 2.7727,
      "eval_samples_per_second": 394.554,
      "eval_steps_per_second": 49.409,
      "step": 17160
    },
    {
      "epoch": 13.26,
      "learning_rate": 2.3232323232323234e-06,
      "loss": 0.0078,
      "step": 17500
    },
    {
      "epoch": 13.64,
      "learning_rate": 1.8181818181818183e-06,
      "loss": 0.0051,
      "step": 18000
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.9058501124382019,
      "eval_loss": 0.620483934879303,
      "eval_runtime": 3.0663,
      "eval_samples_per_second": 356.786,
      "eval_steps_per_second": 44.68,
      "step": 18480
    },
    {
      "epoch": 14.02,
      "learning_rate": 1.3131313131313134e-06,
      "loss": 0.0091,
      "step": 18500
    },
    {
      "epoch": 14.39,
      "learning_rate": 8.080808080808082e-07,
      "loss": 0.0046,
      "step": 19000
    },
    {
      "epoch": 14.77,
      "learning_rate": 3.0303030303030305e-07,
      "loss": 0.0035,
      "step": 19500
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.9067641496658325,
      "eval_loss": 0.6223405599594116,
      "eval_runtime": 3.0619,
      "eval_samples_per_second": 357.298,
      "eval_steps_per_second": 44.744,
      "step": 19800
    },
    {
      "epoch": 15.0,
      "step": 19800,
      "total_flos": 5215995096399360.0,
      "train_loss": 0.4635289347472817,
      "train_runtime": 2578.0344,
      "train_samples_per_second": 61.419,
      "train_steps_per_second": 7.68
    }
  ],
  "max_steps": 19800,
  "num_train_epochs": 15,
  "total_flos": 5215995096399360.0,
  "trial_name": null,
  "trial_params": null
}