Closed simon-mo closed 1 month ago
@simon-mo I think this could be solved by switching this line https://github.com/huggingface/huggingface.js/blob/cc01ed508cfcc49da56f4a7dd761f18180956692/packages/tasks/src/local-apps.ts#L218
from isGgufModel
to isLlamaCppGgufModel
WDYT?
cc @Vaibhavs10 too
@simon-mo please review @Vaibhavs10's PR^ – or if you want to author the change please open a similar PR! 🙏
While vLLM does support gguf with models that have the full
config.json
this model is not supported. https://huggingface.co/kyutai/moshika-candle-q8