checkpoint-108/trainer_state.json

{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 2.958904109589041,
  "eval_steps": 500,
  "global_step": 108,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.68,
      "grad_norm": 0.5096186399459839,
      "learning_rate": 2.5670103092783506e-05,
      "loss": 1.7612,
      "step": 25
    },
    {
      "epoch": 1.37,
      "grad_norm": 0.42532336711883545,
      "learning_rate": 1.7938144329896908e-05,
      "loss": 1.5598,
      "step": 50
    },
    {
      "epoch": 2.05,
      "grad_norm": 0.6212334632873535,
      "learning_rate": 1.020618556701031e-05,
      "loss": 1.414,
      "step": 75
    },
    {
      "epoch": 2.74,
      "grad_norm": 0.4876074194908142,
      "learning_rate": 2.4742268041237115e-06,
      "loss": 1.4442,
      "step": 100
    }
  ],
  "logging_steps": 25,
  "max_steps": 108,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 500,
  "total_flos": 1.7643382323019776e+16,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}