can lumina-mgpt support inference speed up project such as vllm

[yyyouy]

thank you for you great work. i met a question that the generation of a image is too slow, generate a 512x512 image cost almost one minute. so i want to know that can lumina-mgpt support inference speed up project such as vllm

[ChrisLiu6]

We're working on it