llama.cpp

py : handle byte tokens in `get_token_type`

#5341

Merged