4 ms·
Show HN: LatComp – Compress your image into a small and reversible format
While taking the fast.ai course, I was learning about Variational Autoencoders (VAE) and began to wonder:
Is it possible to represent the latent space as an image, and then reconstruct the original image from that representation?
Well, here's the answer.
- ThatGuySam 2y agoI feel like it’s only a matter of time before we get AI-enabled image formats to supersede AVIF and WEBP. Hopefully we can get some meaningful wins on mobile web performance.
- Shynobi 2y agoDefinitely see that happening