{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "global_step": 4371,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.34,
      "learning_rate": 1.77121940059483e-05,
      "loss": 4.5055,
      "step": 500
    },
    {
      "epoch": 0.69,
      "learning_rate": 1.5424388011896593e-05,
      "loss": 3.5521,
      "step": 1000
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.6000000238418579,
      "eval_loss": 2.099963903427124,
      "eval_runtime": 0.0837,
      "eval_samples_per_second": 59.702,
      "eval_steps_per_second": 11.94,
      "step": 1457
    },
    {
      "epoch": 1.03,
      "learning_rate": 1.3136582017844888e-05,
      "loss": 2.8365,
      "step": 1500
    },
    {
      "epoch": 1.37,
      "learning_rate": 1.0848776023793184e-05,
      "loss": 2.1971,
      "step": 2000
    },
    {
      "epoch": 1.72,
      "learning_rate": 8.56097002974148e-06,
      "loss": 1.8933,
      "step": 2500
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 1.0,
      "eval_loss": 0.9956113696098328,
      "eval_runtime": 0.0198,
      "eval_samples_per_second": 252.666,
      "eval_steps_per_second": 50.533,
      "step": 2914
    },
    {
      "epoch": 2.06,
      "learning_rate": 6.273164035689774e-06,
      "loss": 1.6154,
      "step": 3000
    },
    {
      "epoch": 2.4,
      "learning_rate": 3.98535804163807e-06,
      "loss": 1.2959,
      "step": 3500
    },
    {
      "epoch": 2.75,
      "learning_rate": 1.697552047586365e-06,
      "loss": 1.2182,
      "step": 4000
    },
    {
      "epoch": 3.0,
      "eval_accuracy": 1.0,
      "eval_loss": 0.6986985802650452,
      "eval_runtime": 0.0153,
      "eval_samples_per_second": 326.486,
      "eval_steps_per_second": 65.297,
      "step": 4371
    },
    {
      "epoch": 3.0,
      "step": 4371,
      "total_flos": 1151412634864512.0,
      "train_loss": 2.285688607541627,
      "train_runtime": 404.3347,
      "train_samples_per_second": 86.446,
      "train_steps_per_second": 10.81
    }
  ],
  "max_steps": 4371,
  "num_train_epochs": 3,
  "total_flos": 1151412634864512.0,
  "trial_name": null,
  "trial_params": null
}