tried to fix transcribing

..

the model tries to cache on startup to boost performance

offline transcribe

..

fixes for VAD. after good deployment

changed transcriber

changed dockerfile

updated code

Add Spaces deployment requirements without PyAudio and remove PortAudio from Docker build

Docker: remove build-essential/cmake/git to reduce OOM during build

Docker: use Python 3.11 and prefer binary wheels to reduce build OOM

Docker: remove separate llama-cpp-python install (let requirements handle it)

Add Hugging Face Spaces frontmatter to README

Configure for CPU-only zero-gpu deployment

Add Hugging Face Spaces deployment files

Track models folder with Git LFS

changed to moonshine tiny

made changes to allow transcribe

added readme.md

Merge branch 'master'

added pipeline file

Initial commit