I want dall.e to include a text in french in an image. I prompt the text “le jardin sous les etoiles” and tell dall.e to include it in the design, but he never got it right. He writes a text like “le jarden suus lees etoiiiles”. How to force it to write the exact text ?
Hey there and welcome to the community!
So, with image generation, language models are notoriously bad at making correct text. This does not currently seem to be within the scope of Dall-E’s capabilities. So, unfortunately, if you require correct text in an image, it must be edited manually to include the desired text.
Image generation still has a ways to go until it it can get text right in images.
Ok, thank you, so I will edit manually the text, which is not so hard to do after all.