Why have diffusion models displaced GANs so quickly? Consider the tale of the (very strange) first DALLE model. In 2021, diffusions were almost unheard of, yet the creators of DALLE had already rejected the GAN approach. Here’s why. 🧵
18
369
37
2,031
Replying to @tomgoldsteincs
I am convinced that autoregressive and diffusion models just scale better than GANs (w.r.t. compute 📈). Also they have less problems with mode dropping and recall. In the paper "Diffusion Models Beat GANs on Image Synthesis" they convinced the skeptics in it ✅

Aug 19, 2022 · 2:40 AM UTC

1
2
55