This project is under CreativeML Open RAIL-M license. Playground: https://huggingface.co/spaces/kakaobrain/karlo https://huggingface.co/spaces/kakaobrain/karlo

This project is under CreativeML Open RAIL-M license. Playground: https://huggingface.co/spaces/kakaobrain/karlo https://huggingface.co/spaces/kakaobrain/karlo My impressions: out of the box the coherency is superior than a SD model, also the model knows more about popular culture. Unfortunately the model can only generate 256x256 images and it's much bigger