{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 15.0,
  "global_step": 15315,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.49,
      "learning_rate": 1.9347045380346067e-05,
      "loss": 0.1842,
      "step": 500
    },
    {
      "epoch": 0.98,
      "learning_rate": 1.8694090760692135e-05,
      "loss": 0.1537,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.9574627876281738,
      "eval_loss": 0.13882458209991455,
      "eval_runtime": 3.6569,
      "eval_samples_per_second": 2552.167,
      "eval_steps_per_second": 79.849,
      "step": 1021
    },
    {
      "epoch": 1.47,
      "learning_rate": 1.80411361410382e-05,
      "loss": 0.1102,
      "step": 1500
    },
    {
      "epoch": 1.96,
      "learning_rate": 1.7388181521384265e-05,
      "loss": 0.1123,
      "step": 2000
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.9614272117614746,
      "eval_loss": 0.12761715054512024,
      "eval_runtime": 3.6293,
      "eval_samples_per_second": 2571.601,
      "eval_steps_per_second": 80.457,
      "step": 2042
    },
    {
      "epoch": 2.45,
      "learning_rate": 1.673522690173033e-05,
      "loss": 0.0822,
      "step": 2500
    },
    {
      "epoch": 2.94,
      "learning_rate": 1.6082272282076395e-05,
      "loss": 0.0811,
      "step": 3000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.9639987349510193,
      "eval_loss": 0.130422443151474,
      "eval_runtime": 3.6605,
      "eval_samples_per_second": 2549.638,
      "eval_steps_per_second": 79.77,
      "step": 3063
    },
    {
      "epoch": 3.43,
      "learning_rate": 1.5429317662422463e-05,
      "loss": 0.0602,
      "step": 3500
    },
    {
      "epoch": 3.92,
      "learning_rate": 1.4776363042768528e-05,
      "loss": 0.0479,
      "step": 4000
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.9649630188941956,
      "eval_loss": 0.17105060815811157,
      "eval_runtime": 3.6351,
      "eval_samples_per_second": 2567.485,
      "eval_steps_per_second": 80.328,
      "step": 4084
    },
    {
      "epoch": 4.41,
      "learning_rate": 1.4123408423114595e-05,
      "loss": 0.032,
      "step": 4500
    },
    {
      "epoch": 4.9,
      "learning_rate": 1.3470453803460662e-05,
      "loss": 0.03,
      "step": 5000
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.9616414904594421,
      "eval_loss": 0.19338741898536682,
      "eval_runtime": 3.6341,
      "eval_samples_per_second": 2568.197,
      "eval_steps_per_second": 80.351,
      "step": 5105
    },
    {
      "epoch": 5.39,
      "learning_rate": 1.2817499183806728e-05,
      "loss": 0.0243,
      "step": 5500
    },
    {
      "epoch": 5.88,
      "learning_rate": 1.2164544564152792e-05,
      "loss": 0.0172,
      "step": 6000
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.962498664855957,
      "eval_loss": 0.23469704389572144,
      "eval_runtime": 3.639,
      "eval_samples_per_second": 2564.695,
      "eval_steps_per_second": 80.241,
      "step": 6126
    },
    {
      "epoch": 6.37,
      "learning_rate": 1.1511589944498858e-05,
      "loss": 0.0139,
      "step": 6500
    },
    {
      "epoch": 6.86,
      "learning_rate": 1.0858635324844923e-05,
      "loss": 0.0102,
      "step": 7000
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9654987454414368,
      "eval_loss": 0.27059507369995117,
      "eval_runtime": 3.6202,
      "eval_samples_per_second": 2578.034,
      "eval_steps_per_second": 80.659,
      "step": 7147
    },
    {
      "epoch": 7.35,
      "learning_rate": 1.020568070519099e-05,
      "loss": 0.0131,
      "step": 7500
    },
    {
      "epoch": 7.84,
      "learning_rate": 9.552726085537057e-06,
      "loss": 0.0094,
      "step": 8000
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9667845368385315,
      "eval_loss": 0.2712595760822296,
      "eval_runtime": 3.6231,
      "eval_samples_per_second": 2575.978,
      "eval_steps_per_second": 80.594,
      "step": 8168
    },
    {
      "epoch": 8.33,
      "learning_rate": 8.899771465883122e-06,
      "loss": 0.0051,
      "step": 8500
    },
    {
      "epoch": 8.81,
      "learning_rate": 8.246816846229188e-06,
      "loss": 0.0059,
      "step": 9000
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9651773571968079,
      "eval_loss": 0.31283998489379883,
      "eval_runtime": 3.6304,
      "eval_samples_per_second": 2570.776,
      "eval_steps_per_second": 80.431,
      "step": 9189
    },
    {
      "epoch": 9.3,
      "learning_rate": 7.593862226575253e-06,
      "loss": 0.0038,
      "step": 9500
    },
    {
      "epoch": 9.79,
      "learning_rate": 6.94090760692132e-06,
      "loss": 0.0045,
      "step": 10000
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.9661416411399841,
      "eval_loss": 0.31859999895095825,
      "eval_runtime": 3.6286,
      "eval_samples_per_second": 2572.061,
      "eval_steps_per_second": 80.472,
      "step": 10210
    },
    {
      "epoch": 10.28,
      "learning_rate": 6.287952987267385e-06,
      "loss": 0.0008,
      "step": 10500
    },
    {
      "epoch": 10.77,
      "learning_rate": 5.634998367613451e-06,
      "loss": 0.005,
      "step": 11000
    },
    {
      "epoch": 11.0,
      "eval_accuracy": 0.9667845368385315,
      "eval_loss": 0.32250431180000305,
      "eval_runtime": 3.623,
      "eval_samples_per_second": 2576.056,
      "eval_steps_per_second": 80.597,
      "step": 11231
    },
    {
      "epoch": 11.26,
      "learning_rate": 4.9820437479595175e-06,
      "loss": 0.0024,
      "step": 11500
    },
    {
      "epoch": 11.75,
      "learning_rate": 4.329089128305583e-06,
      "loss": 0.0031,
      "step": 12000
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.9653916358947754,
      "eval_loss": 0.32603543996810913,
      "eval_runtime": 3.6195,
      "eval_samples_per_second": 2578.552,
      "eval_steps_per_second": 80.675,
      "step": 12252
    },
    {
      "epoch": 12.24,
      "learning_rate": 3.676134508651649e-06,
      "loss": 0.0024,
      "step": 12500
    },
    {
      "epoch": 12.73,
      "learning_rate": 3.023179888997715e-06,
      "loss": 0.0016,
      "step": 13000
    },
    {
      "epoch": 13.0,
      "eval_accuracy": 0.9679631590843201,
      "eval_loss": 0.33604705333709717,
      "eval_runtime": 3.6293,
      "eval_samples_per_second": 2571.596,
      "eval_steps_per_second": 80.457,
      "step": 13273
    },
    {
      "epoch": 13.22,
      "learning_rate": 2.370225269343781e-06,
      "loss": 0.0029,
      "step": 13500
    },
    {
      "epoch": 13.71,
      "learning_rate": 1.7172706496898467e-06,
      "loss": 0.0016,
      "step": 14000
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.9681774377822876,
      "eval_loss": 0.3418048024177551,
      "eval_runtime": 3.6597,
      "eval_samples_per_second": 2550.205,
      "eval_steps_per_second": 79.788,
      "step": 14294
    },
    {
      "epoch": 14.2,
      "learning_rate": 1.0643160300359125e-06,
      "loss": 0.0002,
      "step": 14500
    },
    {
      "epoch": 14.69,
      "learning_rate": 4.1136141038197843e-07,
      "loss": 0.0007,
      "step": 15000
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.9683917164802551,
      "eval_loss": 0.33835601806640625,
      "eval_runtime": 3.6312,
      "eval_samples_per_second": 2570.225,
      "eval_steps_per_second": 80.414,
      "step": 15315
    }
  ],
  "max_steps": 15315,
  "num_train_epochs": 15,
  "total_flos": 3.222847324631859e+16,
  "trial_name": null,
  "trial_params": null
}