any diffusion model support plan?
Hi. At this moment, we don't have plans to support diffusion models. Compared to language model serving, diffusion workloads are quite different and introduce significantly more complexity, and it's non-trivial to keep the codebase both clean and high-performance while supporting them at this stage.
Feel free to open a PR or share your proposal if you have ideas you'd like to discuss.
thanks for the reply。 cause i saw there are sglang diffusion code inside sglang repo。i thought mini sglang might follow this style to support diffusion models。 yes,it hard to combine diffusion and llm based arch together。but for some of the image generation model,like autogressive,should this kind of model support in your repo? OR you will (maybe) init a new repo mini-sglang-diffusion just like vllm omni?