Add Multimodal (Vision/Image) Support via libmtmd

[aadric239]

Is there any plan to implement the support for the images?
ref: https://github.com/ggml-org/llama.cpp/blob/master/docs/multimodal.md

[giladgd]

Yes, I'm currently working on it and I'll release it as part of the v4 beta. Subscribe to #88 to follow updates on this.

[aadric239]

Thank you will be waiting for it!