{
  "best_metric": 0.033053942024707794,
  "best_model_checkpoint": "./results/checkpoint-3520",
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 5280,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.28,
      "learning_rate": 4.5265151515151514e-05,
      "loss": 0.4808,
      "step": 500
    },
    {
      "epoch": 0.57,
      "learning_rate": 4.053030303030303e-05,
      "loss": 0.0427,
      "step": 1000
    },
    {
      "epoch": 0.85,
      "learning_rate": 3.579545454545455e-05,
      "loss": 0.0385,
      "step": 1500
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.03568091243505478,
      "eval_runtime": 219.2181,
      "eval_samples_per_second": 51.843,
      "eval_steps_per_second": 3.243,
      "step": 1760
    },
    {
      "epoch": 1.14,
      "learning_rate": 3.106060606060606e-05,
      "loss": 0.0309,
      "step": 2000
    },
    {
      "epoch": 1.42,
      "learning_rate": 2.6325757575757575e-05,
      "loss": 0.0265,
      "step": 2500
    },
    {
      "epoch": 1.7,
      "learning_rate": 2.1590909090909093e-05,
      "loss": 0.0246,
      "step": 3000
    },
    {
      "epoch": 1.99,
      "learning_rate": 1.6856060606060605e-05,
      "loss": 0.0244,
      "step": 3500
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.033053942024707794,
      "eval_runtime": 217.1125,
      "eval_samples_per_second": 52.346,
      "eval_steps_per_second": 3.275,
      "step": 3520
    },
    {
      "epoch": 2.27,
      "learning_rate": 1.2121212121212122e-05,
      "loss": 0.0145,
      "step": 4000
    },
    {
      "epoch": 2.56,
      "learning_rate": 7.386363636363637e-06,
      "loss": 0.014,
      "step": 4500
    },
    {
      "epoch": 2.84,
      "learning_rate": 2.651515151515152e-06,
      "loss": 0.0135,
      "step": 5000
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.035611141473054886,
      "eval_runtime": 217.9599,
      "eval_samples_per_second": 52.143,
      "eval_steps_per_second": 3.262,
      "step": 5280
    }
  ],
  "logging_steps": 500,
  "max_steps": 5280,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 2.2877576286437376e+16,
  "trial_name": null,
  "trial_params": null
}