Feature request: Let user choose local whisper model to use

[pulimento]

There are several whisper models available to use locally, perhaps the local env has a beefier GPU that can run a bigger model
A switch pointing to a model will probably do
Thanks for the awesome work!

[vivekuppal]

This request is implemented in the fork
https://github.com/vivekuppal/transcribe

It allows the use of tiny, base, small models as long as the models have been downloaded to the appropriate location.