{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.0,
  "eval_steps": 100,
  "global_step": 40,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 3.4379498958587646,
      "learning_rate": 0.00015000000000000001,
      "loss": 0.9538,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.1374919414520264,
      "learning_rate": 0.0001,
      "loss": 0.3279,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.5991754531860352,
      "learning_rate": 5e-05,
      "loss": 0.1614,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.4659822881221771,
      "learning_rate": 0.0,
      "loss": 0.1422,
      "step": 40
    },
    {
      "epoch": 4.0,
      "step": 40,
      "total_flos": 4.587608018268979e+16,
      "train_loss": 0.39633166790008545,
      "train_runtime": 48.3343,
      "train_samples_per_second": 12.248,
      "train_steps_per_second": 0.828
    }
  ],
  "logging_steps": 10,
  "max_steps": 40,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 4,
  "save_steps": 100,
  "total_flos": 4.587608018268979e+16,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}