quantized/text_encoder-int8/generation_config.json

{
  "_from_model_config": true,
  "bos_token_id": 151643,
  "eos_token_id": 151645,
  "transformers_version": "5.16.1",
  "use_cache": true
}