{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 10.0,
  "eval_steps": 500,
  "global_step": 100,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 1.0,
      "grad_norm": 0.5378170013427734,
      "learning_rate": 0.00019510565162951537,
      "loss": 0.8601,
      "step": 10
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.21907852590084076,
      "learning_rate": 0.00018090169943749476,
      "loss": 0.6658,
      "step": 20
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.2130054235458374,
      "learning_rate": 0.00015877852522924732,
      "loss": 0.6321,
      "step": 30
    },
    {
      "epoch": 4.0,
      "grad_norm": 0.2033412605524063,
      "learning_rate": 0.00013090169943749476,
      "loss": 0.6132,
      "step": 40
    },
    {
      "epoch": 5.0,
      "grad_norm": 0.23296789824962616,
      "learning_rate": 0.0001,
      "loss": 0.5994,
      "step": 50
    },
    {
      "epoch": 6.0,
      "grad_norm": 0.23752418160438538,
      "learning_rate": 6.909830056250527e-05,
      "loss": 0.588,
      "step": 60
    },
    {
      "epoch": 7.0,
      "grad_norm": 0.24876683950424194,
      "learning_rate": 4.12214747707527e-05,
      "loss": 0.5797,
      "step": 70
    },
    {
      "epoch": 8.0,
      "grad_norm": 0.2501850128173828,
      "learning_rate": 1.9098300562505266e-05,
      "loss": 0.5734,
      "step": 80
    },
    {
      "epoch": 9.0,
      "grad_norm": 0.2533394396305084,
      "learning_rate": 4.8943483704846475e-06,
      "loss": 0.5706,
      "step": 90
    },
    {
      "epoch": 10.0,
      "grad_norm": 0.24403364956378937,
      "learning_rate": 0.0,
      "loss": 0.5691,
      "step": 100
    },
    {
      "epoch": 10.0,
      "step": 100,
      "total_flos": 2.023641298501632e+16,
      "train_loss": 0.625129919052124,
      "train_runtime": 1025.9433,
      "train_samples_per_second": 3.09,
      "train_steps_per_second": 0.097
    }
  ],
  "logging_steps": 10,
  "max_steps": 100,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 10,
  "save_steps": 500,
  "total_flos": 2.023641298501632e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}