{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 30,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.1,
      "grad_norm": 6.293304377818269,
      "learning_rate": 6.666666666666667e-06,
      "loss": 0.6271,
      "step": 1
    },
    {
      "epoch": 0.5,
      "grad_norm": 25.185235706709786,
      "learning_rate": 1.973044870579824e-05,
      "loss": 0.792,
      "step": 5
    },
    {
      "epoch": 1.0,
      "grad_norm": 4.228931412669131,
      "learning_rate": 1.686241637868734e-05,
      "loss": 0.7945,
      "step": 10
    },
    {
      "epoch": 1.0,
      "eval_loss": 0.6700347661972046,
      "eval_runtime": 11.8455,
      "eval_samples_per_second": 7.936,
      "eval_steps_per_second": 0.507,
      "step": 10
    },
    {
      "epoch": 1.5,
      "grad_norm": 2.0210374934042243,
      "learning_rate": 1.1736481776669307e-05,
      "loss": 0.5254,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.6038521999778248,
      "learning_rate": 6.039202339608432e-06,
      "loss": 0.4388,
      "step": 20
    },
    {
      "epoch": 2.0,
      "eval_loss": 0.556283175945282,
      "eval_runtime": 11.7263,
      "eval_samples_per_second": 8.016,
      "eval_steps_per_second": 0.512,
      "step": 20
    },
    {
      "epoch": 2.5,
      "grad_norm": 1.3024784212416203,
      "learning_rate": 1.6451218858706374e-06,
      "loss": 0.371,
      "step": 25
    },
    {
      "epoch": 3.0,
      "grad_norm": 1.1334201202802174,
      "learning_rate": 0.0,
      "loss": 0.3518,
      "step": 30
    },
    {
      "epoch": 3.0,
      "eval_loss": 0.5362299084663391,
      "eval_runtime": 11.5046,
      "eval_samples_per_second": 8.171,
      "eval_steps_per_second": 0.522,
      "step": 30
    },
    {
      "epoch": 3.0,
      "step": 30,
      "total_flos": 6229044756480.0,
      "train_loss": 0.5401023705800374,
      "train_runtime": 730.1019,
      "train_samples_per_second": 2.597,
      "train_steps_per_second": 0.041
    }
  ],
  "logging_steps": 5,
  "max_steps": 30,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 6229044756480.0,
  "train_batch_size": 4,
  "trial_name": null,
  "trial_params": null
}