Update tokenization_arcade100k.py

#1

When converting to gguf in llama.cpp, special_tokens is required because it is converted using the qwen like approach, so you need to add this.

Ready to merge
This branch is ready to get merged automatically.

Sign up or log in to comment