We're training a text-to-image model from scratch and open-sourcing itphotoroom.com3 points·ea016··0 commentsOpen articleSaveView on HN