Skip to content

MMVet results are low #947

@Xnhyacinth

Description

@Xnhyacinth

Hi, it was found that the scores obtained during the MMVet test were unusually low. The official Qwen-VL model also only had a score of less than 5. I used vllm and vllm_generate.

Image

same as dvlab-research/VisionThink#17 (comment)

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions