checkpoint-4000/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 4.0,
  "eval_steps": 500,
  "global_step": 4000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5,
      "grad_norm": 28.553239822387695,
      "learning_rate": 0.18000000000000002,
      "loss": 26.996,
      "step": 500
    },
    {
      "epoch": 1.0,
      "grad_norm": 15.302693367004395,
      "learning_rate": 0.16000000000000003,
      "loss": 27.5789,
      "step": 1000
    },
    {
      "epoch": 1.5,
      "grad_norm": 3.7751123905181885,
      "learning_rate": 0.13999999999999999,
      "loss": 21.6633,
      "step": 1500
    },
    {
      "epoch": 2.0,
      "grad_norm": 5.127112865447998,
      "learning_rate": 0.12,
      "loss": 20.7534,
      "step": 2000
    },
    {
      "epoch": 2.5,
      "grad_norm": 22.80483055114746,
      "learning_rate": 0.1,
      "loss": 18.4192,
      "step": 2500
    },
    {
      "epoch": 3.0,
      "grad_norm": 30.512271881103516,
      "learning_rate": 0.08000000000000002,
      "loss": 16.2861,
      "step": 3000
    },
    {
      "epoch": 3.5,
      "grad_norm": 23.139978408813477,
      "learning_rate": 0.06,
      "loss": 14.2996,
      "step": 3500
    },
    {
      "epoch": 4.0,
      "grad_norm": 23.255733489990234,
      "learning_rate": 0.04000000000000001,
      "loss": 9.7249,
      "step": 4000
    }
  ],
  "logging_steps": 500,
  "max_steps": 5000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 5,
  "save_steps": 500,
  "total_flos": 2.9814348175589376e+16,
  "train_batch_size": 8,
  "trial_name": null,
  "trial_params": null
}