checkpoint-21/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 0.9882352941176471,
  "eval_steps": 500,
  "global_step": 21,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.24,
      "grad_norm": 1.444962501525879,
      "learning_rate": 0.00017777777777777779,
      "loss": 1.9679,
      "step": 5
    },
    {
      "epoch": 0.47,
      "grad_norm": 1.3045181035995483,
      "learning_rate": 0.00012222222222222224,
      "loss": 1.9906,
      "step": 10
    },
    {
      "epoch": 0.71,
      "grad_norm": 1.1084049940109253,
      "learning_rate": 6.666666666666667e-05,
      "loss": 1.927,
      "step": 15
    },
    {
      "epoch": 0.94,
      "grad_norm": 1.062194585800171,
      "learning_rate": 1.1111111111111112e-05,
      "loss": 1.7807,
      "step": 20
    }
  ],
  "logging_steps": 5,
  "max_steps": 21,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 1,
  "save_steps": 500,
  "total_flos": 3671542022012928.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}