{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 5.361930294906166,
  "eval_steps": 500,
  "global_step": 2000,
  "is_hyper_param_search": true,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "eval_accuracy": 0.7106939554214478,
      "eval_loss": 0.538670003414154,
      "eval_runtime": 8.938,
      "eval_samples_per_second": 333.743,
      "eval_steps_per_second": 20.922,
      "step": 373
    },
    {
      "epoch": 1.3404825737265416,
      "grad_norm": 6.848696231842041,
      "learning_rate": 2.0560750028839876e-06,
      "loss": 0.5981,
      "step": 500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.7321488261222839,
      "eval_loss": 0.5092849135398865,
      "eval_runtime": 8.9404,
      "eval_samples_per_second": 333.653,
      "eval_steps_per_second": 20.916,
      "step": 746
    },
    {
      "epoch": 2.680965147453083,
      "grad_norm": 9.747313499450684,
      "learning_rate": 1.7377971386604603e-06,
      "loss": 0.4883,
      "step": 1000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 0.7412001490592957,
      "eval_loss": 0.4964694678783417,
      "eval_runtime": 8.8911,
      "eval_samples_per_second": 335.504,
      "eval_steps_per_second": 21.032,
      "step": 1119
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.7412001490592957,
      "eval_loss": 0.5169216394424438,
      "eval_runtime": 8.9728,
      "eval_samples_per_second": 332.448,
      "eval_steps_per_second": 20.841,
      "step": 1492
    },
    {
      "epoch": 4.021447721179625,
      "grad_norm": 6.287171363830566,
      "learning_rate": 1.4195192744369326e-06,
      "loss": 0.4333,
      "step": 1500
    },
    {
      "epoch": 5.0,
      "eval_accuracy": 0.7422058582305908,
      "eval_loss": 0.5148842930793762,
      "eval_runtime": 8.9557,
      "eval_samples_per_second": 333.084,
      "eval_steps_per_second": 20.881,
      "step": 1865
    },
    {
      "epoch": 5.361930294906166,
      "grad_norm": 10.289645195007324,
      "learning_rate": 1.101241410213405e-06,
      "loss": 0.3746,
      "step": 2000
    }
  ],
  "logging_steps": 500,
  "max_steps": 3730,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 1.018246245226362e+16,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": {
    "learning_rate": 2.374352867107515e-06,
    "per_device_train_batch_size": 32
  }
}