Karlo is a text-conditional image generation model based on OpenAI’s unCLIP architecture with the improvement over the standard super-resolution model from 64px to 256px, recovering high-frequency details only in the small number of denoising steps.
One thing that I’m still waiting for in the AI generation image space is higher resolution outputs; I understand why the image outputs are always capped by some upper bound width/height, but 256px is a huge limit for doing anything practical.
The images look really cool though! Impressive.
1 Like