8000 GitHub - jinjinw/VLMs-Compositionality-Game-Theory: The author's implementation for the ICML 2024 paper.
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content

jinjinw/VLMs-Compositionality-Game-Theory

Repository files navigation

Diagnosing the Compositional Knowledge of Vision Language Models from a Game-Theoretic View

This repo includes the authors' Pytorch implementation of the paper:

[arxiv]

Introduction

Compositional reasoning capabilities are usually considered as fundamental skills to characterize human perception. Recent studies show that current Vision Language Models (VLMs) surprisingly lack sufficient knowledge with respect to such capabilities. To this end, we propose to thoroughly diagnose the composition representations encoded by VLMs, systematically revealing the potential cause for this weakness. Specifically, we propose evaluation methods from a novel game-theoretic view to assess the vulnerability of VLMs on different aspects of compositional understanding, e.g., relations and attributes. Extensive experimental results demonstrate and validate several insights to understand the incapabilities of VLMs on compositional reasoning, which provide useful and reliable guidance for future studies.

overview

Updates

  • [09/2024] release the text-part code for CLIP.
  • text-part code for other VLMs
  • image-part code for other VLMs
  • text&image-part code for other VLMs

Dependencies

  • Python 3 >= 3.8
  • Pytorch >= 2.1.0
  • OpenCV >= 4.8.1.78
  • Scipy >= 1.10.1
  • NumPy >= 1.24.3

Data Preparation

Download the VG_relation dataset to ./data from here.

Evaluations

To evaluate the model's compositional knowledge, please run:

python3 inference.py

Acknowledgements

Contact

If you have any questions, please feel free to contact me via wj0529@connect.hku.hk.

About

The author's implementation for the ICML 2024 paper.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

0