fad-pytorch 0.0.4

Frechet Audio Distance evaluation in PyTorch

fad_pytorch

Original FAD paper (PDF)

Install

pip install fad_pytorch

About

(Intended) Features:

• runs in parallel on multiple GPUs
• supports 48kHz sample rates and stereo when possible
• supports CLAP embeddings, in addition to VGGish and PANN
• favors ops in PyTorch instead of numpy
• allows dataset access via WebDataset (over s3://)
• runs on CPU, CUDA, or MPS

This is designed to be run as 3 command-line scripts in succession. The latter 2 (fad_embed and fad_score) are probably what most people will want:

1. fad_gen: produces directories of real & fake audio
2. fad_embed <real_audio_dir> <fake_audio_dir>: produces directories of embeddings of real & fake audio
3. fad_score <real_emb_dir> <fake_emb_dir>: reads the embeddings & generates FAD score, for real (“$r$”) and fake (“$f$”):

$$ FAD = || \mu_r - \mu_f ||^2 + tr\left(\Sigma_r + \Sigma_f - 2 \sqrt{\Sigma_r \Sigma_f}\right)$$

FAQ / Troubleshooting

• “RuntimeError: CUDA error: invalid device ordinal”: This happens when you have a “bad node” on an AWS cluster. Haven’t yet figured out what causes it or how to fix it. Workaround: Just add the current node to your SLURM --exclude list, exit and retry. Note: it may take as many as 5 to 7 retries before you get a “good node”.

Contributing

This repo is still fairly “bare bones” and will benefit from more documentation and features as time goes on. Note that it is written using nbdev, so the things to are:

1. fork this repo
2. clone your fork to your (local) machine
3. Install nbdev: python3 -m pip install -U nbdev
4. Make changes by editing the notebooks in nbs/, not the .py files in fad_pytorch/.
5. Run nbdev_export to export notebook changes to .py files
6. For good measure, run nbdev_install_hooks and nbdev_clean - especially if you’ve added any notebooks.
7. Do a git status to see all the .ipynb and .py files that need to be added & committed
8. git add those files and then git commit, and then git push
9. Take a look in your GitHub Actions tab, and see if the “test” and “deploy” CI runs finish properly (green light) or fail (red light)
10. One you get green lights, send in a Pull Request!

Feel free to ask me for tips with nbdev, it has quite a learning curve. You can also ask on fast.ai forums and/or fast.ai Discord

Related Repos

There are [several] others, but this one is mine. These repos didn’t have all the features I wanted, but I used them for inspiration:

• https://github.com/gudgud96/frechet-audio-distance
• https://github.com/google-research/google-research/tree/master/frechet_audio_distance: Goes with Original FAD paper
• https://github.com/AndreevP/speech_distances