[Model]: Add support for qwen3-VL

[anxiaohuai]

Model Name

Qwen/Qwen3-VL-2B-Instruct

Model Source

https://huggingface.co/Qwen/Qwen3-VL-2B-Instruct

Model Description

Meet Qwen3-VL — the most powerful vision-language model in the Qwen series to date.

Paper or Documentation

No response

Known Challenges or Special Requirements

No response

Contribution

• I am willing to help implement and submit a Pull Request for this model.

[chenghuaWang]

We are planning to support Qwen3.5 instead of Qwen3. Qwen3.5 has better performace than Qwen3VL