generation_config.json

{
  "batch_size": 64,
  "bos_token_id": 128000,
  "do_sample": true,
  "eos_token_id": 128001,
  "max_length": 4096,
  "num_continuous_generation_layers": 0,
  "special_tokens_ids": {
    "begin_continuous": 128015,
    "continuous_rep": 128017,
    "end_continuous": 128016
  },
  "subject_embed_dim": 4096,
  "temperature": 0.6,
  "top_p": 0.9,
  "transformers_version": "4.55.0"
}