News for hackers. Stuff that matters.

A comment on Imagen, a text-to-image diffusion model. Parent comment

Re:Imagen, a text-to-image diffusion model by thorum on Monday May 23, 2022 @06:46PM (#31485800)
That only lasts until the community copies the paper and catches up. For example the open source DALLE-2 implementation is coming along great: https://github.com/lucidrains/DALLE2-pytorch
[ Parent | on HN ]
Re:Imagen, a text-to-image diffusion model by lucidrains on Monday May 23, 2022 @08:44PM (#31486693)
Imagen actually shows some of the components in DALLE2 is unnecessary, so Imagen will end up being easier to build. I'll definitely add the dynamic thresholding trick from Imagen to DALLE2 repository though; that is a finding that should boost any DDPMs using classifier free guidance.
[ Parent | on HN ]
Re:Imagen, a text-to-image diffusion model by forgingahead on Monday May 23, 2022 @11:55PM (#31487806)
Thanks for all your work on these projects!
[ Parent | on HN ]