[zaptrem]

Including the text positioning generation part? What’s the source on that?

[alextheparrot]

The comment was directed at “doesn't this method add another cost and overhead for calling Text-to-Image models”

[wahnfrieden]

No