{
  "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": 0.9110807180404663,
      "eval_loss": 0.2603251338005066,
      "eval_runtime": 3.7587,
      "eval_samples_per_second": 194.484,
      "eval_steps_per_second": 24.477,
      "step": 195
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9179206490516663,
      "eval_loss": 0.2944748103618622,
      "eval_runtime": 3.8108,
      "eval_samples_per_second": 191.821,
      "eval_steps_per_second": 24.142,
      "step": 390
    },
    {
      "epoch": 2.56,
      "learning_rate": 3.282996735800617e-05,
      "loss": 0.1807,
      "step": 500
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9083447456359863,
      "eval_loss": 0.43401873111724854,
      "eval_runtime": 3.7691,
      "eval_samples_per_second": 193.947,
      "eval_steps_per_second": 24.409,
      "step": 585
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9124487042427063,
      "eval_loss": 0.39779114723205566,
      "eval_runtime": 3.7531,
      "eval_samples_per_second": 194.771,
      "eval_steps_per_second": 24.513,
      "step": 780
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9042407870292664,
      "eval_loss": 0.6419674754142761,
      "eval_runtime": 3.6686,
      "eval_samples_per_second": 199.26,
      "eval_steps_per_second": 25.078,
      "step": 975
    },
    {
      "epoch": 5.13,
      "learning_rate": 2.800203098182879e-05,
      "loss": 0.0523,
      "step": 1000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.9069767594337463,
      "eval_loss": 0.6490611433982849,
      "eval_runtime": 3.596,
      "eval_samples_per_second": 203.28,
      "eval_steps_per_second": 25.584,
      "step": 1170
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9124487042427063,
      "eval_loss": 0.6302383542060852,
      "eval_runtime": 3.7728,
      "eval_samples_per_second": 193.754,
      "eval_steps_per_second": 24.385,
      "step": 1365
    },
    {
      "epoch": 7.69,
      "learning_rate": 2.3174094605651413e-05,
      "loss": 0.0146,
      "step": 1500
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9097127318382263,
      "eval_loss": 0.728517472743988,
      "eval_runtime": 3.574,
      "eval_samples_per_second": 204.533,
      "eval_steps_per_second": 25.742,
      "step": 1560
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9206566214561462,
      "eval_loss": 0.6614168882369995,
      "eval_runtime": 3.5326,
      "eval_samples_per_second": 206.93,
      "eval_steps_per_second": 26.043,
      "step": 1755
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9151846766471863,
      "eval_loss": 0.7393134236335754,
      "eval_runtime": 3.6341,
      "eval_samples_per_second": 201.15,
      "eval_steps_per_second": 25.316,
      "step": 1950
    },
    {
      "epoch": 10.26,
      "learning_rate": 1.8346158229474035e-05,
      "loss": 0.0063,
      "step": 2000
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.9165526628494263,
      "eval_loss": 0.7556606531143188,
      "eval_runtime": 3.6289,
      "eval_samples_per_second": 201.437,
      "eval_steps_per_second": 25.352,
      "step": 2145
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9151846766471863,
      "eval_loss": 0.7709615230560303,
      "eval_runtime": 3.5866,
      "eval_samples_per_second": 203.815,
      "eval_steps_per_second": 25.651,
      "step": 2340
    },
    {
      "epoch": 12.82,
      "learning_rate": 1.3518221853296657e-05,
      "loss": 0.0037,
      "step": 2500
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.9192886352539062,
      "eval_loss": 0.7879533171653748,
      "eval_runtime": 3.5996,
      "eval_samples_per_second": 203.079,
      "eval_steps_per_second": 25.558,
      "step": 2535
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.9124487042427063,
      "eval_loss": 0.8315412402153015,
      "eval_runtime": 3.763,
      "eval_samples_per_second": 194.257,
      "eval_steps_per_second": 24.448,
      "step": 2730
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.9179206490516663,
      "eval_loss": 0.8191958069801331,
      "eval_runtime": 3.6714,
      "eval_samples_per_second": 199.108,
      "eval_steps_per_second": 25.059,
      "step": 2925
    },
    {
      "epoch": 15.38,
      "learning_rate": 8.690285477119281e-06,
      "loss": 0.0001,
      "step": 3000
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.9192886352539062,
      "eval_loss": 0.8122023940086365,
      "eval_runtime": 3.7565,
      "eval_samples_per_second": 194.598,
      "eval_steps_per_second": 24.491,
      "step": 3120
    },
    {
      "epoch": 17.0,
      "eval_accuracy": 0.9179206490516663,
      "eval_loss": 0.8226170539855957,
      "eval_runtime": 3.6589,
      "eval_samples_per_second": 199.788,
      "eval_steps_per_second": 25.144,
      "step": 3315
    },
    {
      "epoch": 17.95,
      "learning_rate": 3.862349100941902e-06,
      "loss": 0.0001,
      "step": 3500
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.9179206490516663,
      "eval_loss": 0.8273718953132629,
      "eval_runtime": 3.7025,
      "eval_samples_per_second": 197.433,
      "eval_steps_per_second": 24.848,
      "step": 3510
    },
    {
      "epoch": 19.0,
      "eval_accuracy": 0.9179206490516663,
      "eval_loss": 0.8303136825561523,
      "eval_runtime": 3.6587,
      "eval_samples_per_second": 199.797,
      "eval_steps_per_second": 25.145,
      "step": 3705
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.9179206490516663,
      "eval_loss": 0.8313552737236023,
      "eval_runtime": 3.6572,
      "eval_samples_per_second": 199.879,
      "eval_steps_per_second": 25.156,
      "step": 3900
    },
    {
      "epoch": 20.0,
      "step": 3900,
      "total_flos": 4114530683719680.0,
      "train_loss": 0.033037299305582656,
      "train_runtime": 1549.9583,
      "train_samples_per_second": 40.362,
      "train_steps_per_second": 2.516
    }
  ],
  "max_steps": 3900,
  "num_train_epochs": 20,
  "total_flos": 4114530683719680.0,
  "trial_name": null,
  "trial_params": null
}