Skip to content

feat: download models from our mirror first and verify SHA-256 (v0.48.0) #317

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
beveradb merged 2 commits into from
Oct 29, 2026
Merged

feat: download models from our mirror first and verify SHA-256 (v0.48.0) #317

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
73799f8
feat: download models from our mirror first and verify SHA-256 (v0.48.0)
beveradb Oct 29, 2026
185328e
fix: per-process partial download path; retry transient errors in che…
beveradb Oct 29, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 6 additions & 0 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -381,6 +381,12 @@ For programmatic use, you can output the model list in JSON format:
audio-separator -l --list_format=json
```

### Where Models Are Downloaded From

Model files are downloaded on first use into `model_file_dir` (or `AUDIO_SEPARATOR_MODEL_DIR`). Each file is fetched from the [audio-separator models release](https://github.com/nomadkaraoke/python-audio-separator/releases/tag/model-configs) first, with the original source (the UVR model repo, or Meta's CDN for Demucs) as a fallback. UVR VIP models are still fetched from Anjok07's repo first.

Every model file must match the SHA-256 published in [`audio_separator/model-checksums.json`](audio_separator/model-checksums.json). A download that doesn't match is discarded and the next source is tried. A file already in the model directory is checked too, and replaced if it doesn't match. Maintainers who upload new model files should regenerate the manifest with `tools/update-model-checksums.py`.

### Processing Large Files

For very long audio files (>1 hour), you may encounter out-of-memory errors. The `--chunk_duration` option automatically splits large files into smaller chunks, processes them separately, and merges the results:
Expand Down
Loading
Loading