Interactive Vector-Quantized Variational Autoencoder (VQ-VAE) running live in the browser:

https://xnought.github.io/vq-vae-explainer/

trimmed-vq-vae-explainer-demo.mov

Dev

pnpm install
pnpm dev

Cite

@misc{bertucci2024vq,
  title = {{VQ-VAE Explainer}},
  author = {Donald Bertucci and Duen Horng (Polo) Chau},
  url = {https://xnought.github.io/vq-vae-explainer/},
  year = {2024}
}

References

• https://keras.io/examples/generative/vq_vae/ (used for training)
• https://arxiv.org/abs/1711.00937 (original VQ-VAE paper)
• https://github.com/xnought/vae-explainer