[Feature] support qwen2 vl

[zhyncs]

Checklist

• 1. If the issue you raised is not a feature but a question, please raise a discussion at https://github.com/sgl-project/sglang/discussions/new/choose Otherwise, it will be closed.
• 2. Please use English, otherwise it will be closed.

Motivation

https://huggingface.co/collections/Qwen/qwen2-vl-66cee7455501d7126940800d

Related resources

No response

[merrymercy]

cc @JustinLin610

[HiDostoyevskiy]

I eagerly look forward to the support for the qwen2-vl series models by sglang as soon as possible. Thanks!

[zhyncs]

I eagerly look forward to the support for the qwen2-vl series models by sglang as soon as possible. Thanks!

@ispobock's partner is currently working on it.

[bash99]

I would like to know will sglang support qwen2 vl gptq-int4 release with more than one GPU?

https://qwen.readthedocs.io/en/latest/quantization/gptq.html said vllm don't work and need padding and quantizated again.

[zhyncs]

ref #1546