dpr-question_encoder-multiset-base/tokenizer_config.json

{
  "do_lower_case": true
}