Train model when I have image only without any bbox info

[Decalogue]

Hi, Thanks for your great work! I want to know whether the model can be trained without regions. In other words, I have caption and image only without any bbox info, how can I make the model work? Thank you so much!

[LuoweiZhou]

@Decalogue You will need to adjust the image input (e.g., feature vectors and positional encodings) accordingly. If you intend to take CNN activations as the input, you might want to refer to our recent work ClipBERT: https://github.com/jayleicn/ClipBERT