Text+ Incomplete image prompting

[cachett-ML]

Hi,

In the OpenAI blogpost they show that they can provide a partial image in addition to a piece of text and that Dall-E will complete the image. Can we do the same here? If yes, I would really appreciate someone explaining how to do it!

Best.

[borisdayma]

So I've not implemented it yet but the auto-completion this model can do would be left to right, top to bottom. So you would for example hide bottom half and it would complete it.

[cachett-ML]

I see, thank you ! Do you think it would be hard to implement? If you have clear enough directions, I may be able to help a bit. Best