[Feature]: Update Model and Config Repository Links (if possible)

[Bebra777228]

Description

Recently, I created two repositories for UVR configs and models:

• Repository for storing models and configs
• Repository for storing configs, model data, and a list of all publicly available models

In the second link, you can find two files: model_list_filenames.json and model_list_links.json

• The first file contains user-friendly model names and model filenames.
• The second file contains the same information but with links to the provided models and configs.
  • Models are sourced from my HuggingFace repository.
  • Configs are sourced from my GitHub repository.

I think this will make it much easier to edit configs and add new models.

Do you think it's possible to replace the old links with the current ones?
I haven't tried it myself yet, but I think it's a good idea since the old repository is rarely updated and doesn't add many new models. It would be better if you could add support for files from my repository. And yes, almost all the models there work fine, except those that require changes in the files bs_roformer.py and mel_band_roformer.py and the addition of new architectures.

[beveradb]

Hey, I appreciate the desire to make it easier to add new models - I would like to make that easier too, without me personally being a bottleneck (especially when I'm busy or AFK for a while!)

However I'm not sure I understand the benefit of adding an external repository into the mix - in my opinion, it would be cleaner to keep everything contained within this git repo, which it already is right now (all models and configs from UVR are mirrored here: https://github.com/nomadkaraoke/python-audio-separator/releases/tag/model-configs )

I've actually got a script which makes it real easy to upload new files to that "Models & Configs" release too, here: https://github.com/nomadkaraoke/python-audio-separator/blob/main/tools/sync-to-github.py

That's what I used just now when I added the models from PR #177 an hour or so ago - I just downloaded all the models from the link provided to a local folder and ran my sync-to-github.py script to upload them to github.

The only thing stopping you from doing the same to upload new models & configs to this repo is access permissions; if I add you as a maintainer of this repo, would that work for you? Then you won't have to wait for me in order to add new models to audio-separator 🙏

[Bebra777228]

I still decided to try to do this and I did it, all models, except for those that require changes in the architecture, work without any problems. The models are downloaded directly from my config file model_list_links.json.

I'll leave the change here:
main...Bebra777228:python-audio-separator:UVR_resources
Perhaps you will still decide to take them for yourself.

I regularly upload new models to this config file, and there are already more models there than in download_checks.json and your models.json combined.