Does it able to connect with llama for image understanding?

[lucasjinreal]

Does it able to connect with llama for image understanding?

[cxxgtxy]

Yes. In principle, it can be trained using CLIP and replace the VIT in vision-language models such as MobileVLM(https://github.com/Meituan-AutoML/MobileVLM).