{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 19.997442455242968,
  "global_step": 3900,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9274965524673462,
      "eval_loss": 0.23106738924980164,
      "eval_runtime": 10.0851,
      "eval_samples_per_second": 72.483,
      "eval_steps_per_second": 9.122,
      "step": 195
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9165526628494263,
      "eval_loss": 0.3205892741680145,
      "eval_runtime": 10.086,
      "eval_samples_per_second": 72.477,
      "eval_steps_per_second": 9.122,
      "step": 390
    },
    {
      "epoch": 2.56,
      "learning_rate": 1.2754443842729812e-05,
      "loss": 0.1729,
      "step": 500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9274965524673462,
      "eval_loss": 0.360751748085022,
      "eval_runtime": 9.9443,
      "eval_samples_per_second": 73.509,
      "eval_steps_per_second": 9.251,
      "step": 585
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9233925938606262,
      "eval_loss": 0.4383067190647125,
      "eval_runtime": 10.0596,
      "eval_samples_per_second": 72.667,
      "eval_steps_per_second": 9.145,
      "step": 780
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9206566214561462,
      "eval_loss": 0.5189540982246399,
      "eval_runtime": 10.0775,
      "eval_samples_per_second": 72.538,
      "eval_steps_per_second": 9.129,
      "step": 975
    },
    {
      "epoch": 5.13,
      "learning_rate": 1.0878790336446016e-05,
      "loss": 0.0589,
      "step": 1000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9233925938606262,
      "eval_loss": 0.5728073716163635,
      "eval_runtime": 9.9637,
      "eval_samples_per_second": 73.366,
      "eval_steps_per_second": 9.233,
      "step": 1170
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9274965524673462,
      "eval_loss": 0.5684310793876648,
      "eval_runtime": 10.2744,
      "eval_samples_per_second": 71.148,
      "eval_steps_per_second": 8.954,
      "step": 1365
    },
    {
      "epoch": 7.69,
      "learning_rate": 9.003136830162221e-06,
      "loss": 0.0171,
      "step": 1500
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9261285662651062,
      "eval_loss": 0.5847771167755127,
      "eval_runtime": 10.1261,
      "eval_samples_per_second": 72.189,
      "eval_steps_per_second": 9.085,
      "step": 1560
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.928864598274231,
      "eval_loss": 0.5692142248153687,
      "eval_runtime": 10.1226,
      "eval_samples_per_second": 72.214,
      "eval_steps_per_second": 9.089,
      "step": 1755
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9261285662651062,
      "eval_loss": 0.6423346400260925,
      "eval_runtime": 10.1278,
      "eval_samples_per_second": 72.178,
      "eval_steps_per_second": 9.084,
      "step": 1950
    },
    {
      "epoch": 10.26,
      "learning_rate": 7.1274833238784245e-06,
      "loss": 0.0094,
      "step": 2000
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.9274965524673462,
      "eval_loss": 0.6332233548164368,
      "eval_runtime": 10.2523,
      "eval_samples_per_second": 71.301,
      "eval_steps_per_second": 8.974,
      "step": 2145
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9316005706787109,
      "eval_loss": 0.6691190600395203,
      "eval_runtime": 10.205,
      "eval_samples_per_second": 71.631,
      "eval_steps_per_second": 9.015,
      "step": 2340
    },
    {
      "epoch": 12.82,
      "learning_rate": 5.2518298175946285e-06,
      "loss": 0.0015,
      "step": 2500
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.930232584476471,
      "eval_loss": 0.6589073538780212,
      "eval_runtime": 10.1851,
      "eval_samples_per_second": 71.771,
      "eval_steps_per_second": 9.033,
      "step": 2535
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.928864598274231,
      "eval_loss": 0.6745939254760742,
      "eval_runtime": 10.1553,
      "eval_samples_per_second": 71.982,
      "eval_steps_per_second": 9.059,
      "step": 2730
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.9357045292854309,
      "eval_loss": 0.6391911506652832,
      "eval_runtime": 10.1864,
      "eval_samples_per_second": 71.762,
      "eval_steps_per_second": 9.032,
      "step": 2925
    },
    {
      "epoch": 15.38,
      "learning_rate": 3.376176311310833e-06,
      "loss": 0.0002,
      "step": 3000
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.928864598274231,
      "eval_loss": 0.6711709499359131,
      "eval_runtime": 10.0882,
      "eval_samples_per_second": 72.461,
      "eval_steps_per_second": 9.12,
      "step": 3120
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.9343365430831909,
      "eval_loss": 0.6648762226104736,
      "eval_runtime": 9.9723,
      "eval_samples_per_second": 73.303,
      "eval_steps_per_second": 9.226,
      "step": 3315
    },
    {
      "epoch": 17.95,
      "learning_rate": 1.5005228050270366e-06,
      "loss": 0.0,
      "step": 3500
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.930232584476471,
      "eval_loss": 0.6720365881919861,
      "eval_runtime": 10.1302,
      "eval_samples_per_second": 72.16,
      "eval_steps_per_second": 9.082,
      "step": 3510
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.9316005706787109,
      "eval_loss": 0.6750090718269348,
      "eval_runtime": 10.1546,
      "eval_samples_per_second": 71.987,
      "eval_steps_per_second": 9.06,
      "step": 3705
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.9316005706787109,
      "eval_loss": 0.6705575585365295,
      "eval_runtime": 10.275,
      "eval_samples_per_second": 71.144,
      "eval_steps_per_second": 8.954,
      "step": 3900
    },
    {
      "epoch": 20.0,
      "step": 3900,
      "total_flos": 1.4573542659723264e+16,
      "train_loss": 0.03345794048399115,
      "train_runtime": 4631.396,
      "train_samples_per_second": 13.508,
      "train_steps_per_second": 0.842
    }
  ],
  "max_steps": 3900,
  "num_train_epochs": 20,
  "total_flos": 1.4573542659723264e+16,
  "trial_name": null,
  "trial_params": null
}