We're training a text-to-image model from scratch and open-sourcing it

3 points | by ea016 4 hours ago