LexiconCode

No matter which method I use there seems to be an issue with the transformers. 1. zphang transformers 68d640f7c368bcaaaecfc678f11908ebbd3d6176 ``` pip uninstall transformers pip install git+https://github.com/zphang/transformers.git@68d640f7c368bcaaaecfc678f11908ebbd3d6176 ``` ``` Found existing...

Changing `LLaMATokenizer` to `LlamaTokenizer` works `tokenizer_config.json` Thank you

-removed-

> PR welcome. I apologize my issue ended up being related to the standalone python build not the virtual environment.

OS agnostic temp path for `whisper_server.py` ```py import tempfile temp_dir = tempfile.TemporaryDirectory().name audio_filename = os.path.join(temp_dir,"whisper.wav") temp_dir.cleanup() # place near the end of `die()` function ```

Dragonfly definitely used by a broad range of community. I would say main repository has become stagnant which puts the community at risk. I think there's a lot that could...

If there was a function to activate and deactivate multiple grammars this could allow for context/application specific grammars. Use case would be multiple programs each with their own grammars. Client-side...

> Looking back over this issue today, it occurs to me that Dragonfly users have been able to do this for a long time via the special `Grammar.process_recognition()` callback documented...

It does provide a different type of exposure as people browse these lists looking to contribute. I've not had a ton of experience with it myself but I seen some...

This is reproducible with dragonfly's text engine. The issue is either with short integer ref or a ccr bug. Further investigation is needed.