descriptinc / audiotools

Object-oriented handling of audio data, with GPU-powered augmentations, and more.
https://descriptinc.github.io/audiotools/
MIT License
233 stars 39 forks source link

Adding a model registry #51

Closed pseeth closed 2 years ago

pseeth commented 2 years ago

This PR adds a model registry that is backed by either a local filestore, or a GCP bucket. The idea is to enforce some kind of versioning system for wav2wav like models. Some example output from the list_models functionality:

image

It's hopefully written in such a way that you can do more flexibly use other backends.