Skip to content

Can EgoLifeQA Benchmark Be Integrated into LMMs-Eval? #18

Description

@Artessay

Hi,

Thank you for the great work on EgoLife. It has been very helpful for benchmarking and evaluating large multimodal models on very long video understanding.
I am writing to ask whether it would be possible to integrate the EgoLifeQA Benchmark into the LMMs-Eval framework, which is also developed in your Lab.

Specifically, I would like to ask:

  • Is there interest in supporting EgoLifeQA as an evaluation task within LMMs-Eval?
  • Are there any technical or design constraints we should be aware of (e.g., data format, evaluation protocol, task abstraction)?
  • Would you recommend submitting a PR, or is there a preferred process for proposing new benchmark integrations?

Thank you very much for your time and guidance. I look forward to your thoughts.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions