why settle for a few images from #stablediffusion when you can slowly walk your way around the sample space and create hyponotic videos you can't look away from? In this 2min video (~1hr to render on A100) I'm smoothly interpolating between random noise inputs into the model. https://t.co/A4Ue1pqoMo
— Andrej Karpathy (@karpathy) Aug 16, 2022
from Twitter https://twitter.com/karpathy
August 15, 2022 at 05:57PM
via IFTTT
No comments:
Post a Comment