What models are used for img2prompt and LLAVA

[kz29]

Hello,
Thank you for the provided code. I was reading the paper and checking the GitHub as well but there are no implementation details regarding img2prompt and LLAVA models. Can you please elaborate more and share the details of how can we reproduce these models?
Thank you in advance.

[sftsgly]

bro，do u solve it？

[yukodada]

bro, solved it?

[sftsgly]

I check the lavis and find that img2prompt only have base type now , so I use this model type and code some function by myself to sove it.