{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.769230769230769,
  "eval_steps": 500,
  "global_step": 27,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1,
      "grad_norm": 24.191700961175112,
      "learning_rate": 6.666666666666667e-06,
      "loss": 0.5581,
      "step": 1
    },
    {
      "epoch": 0.51,
      "grad_norm": 12.576350745768869,
      "learning_rate": 1.9659258262890683e-05,
      "loss": 0.8334,
      "step": 5
    },
    {
      "epoch": 0.92,
      "eval_loss": 1.4131343364715576,
      "eval_runtime": 38.4185,
      "eval_samples_per_second": 8.954,
      "eval_steps_per_second": 0.573,
      "step": 9
    },
    {
      "epoch": 1.03,
      "grad_norm": 17.837211332799143,
      "learning_rate": 1.608761429008721e-05,
      "loss": 0.738,
      "step": 10
    },
    {
      "epoch": 1.54,
      "grad_norm": 3.797877129136344,
      "learning_rate": 1e-05,
      "loss": 0.6303,
      "step": 15
    },
    {
      "epoch": 1.95,
      "eval_loss": 1.4060215950012207,
      "eval_runtime": 39.241,
      "eval_samples_per_second": 8.766,
      "eval_steps_per_second": 0.561,
      "step": 19
    },
    {
      "epoch": 2.05,
      "grad_norm": 7.485764754081964,
      "learning_rate": 3.912385709912794e-06,
      "loss": 0.5348,
      "step": 20
    },
    {
      "epoch": 2.56,
      "grad_norm": 3.0173543034517323,
      "learning_rate": 3.4074173710931804e-07,
      "loss": 0.4647,
      "step": 25
    },
    {
      "epoch": 2.77,
      "eval_loss": 1.4294962882995605,
      "eval_runtime": 38.6474,
      "eval_samples_per_second": 8.901,
      "eval_steps_per_second": 0.569,
      "step": 27
    },
    {
      "epoch": 2.77,
      "step": 27,
      "total_flos": 5600905789440.0,
      "train_loss": 0.6124309632513258,
      "train_runtime": 676.0028,
      "train_samples_per_second": 2.72,
      "train_steps_per_second": 0.04
    }
  ],
  "logging_steps": 5,
  "max_steps": 27,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 5600905789440.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}