ActiView: Evaluating Active Perception Ability for Multimodal Large Language Models

🌐 Homepage| 📖 arXiv

Coming soon ⏳🔜

• 🚧 We will release our benchmark and evaluation toolkit ASAP.

• 📨 If you have any questions regarding the data and evaluation scripts when they are temporarily unavailable, please contact [email protected].

Project Overview

This repository contains all the necessary materials, including:

• Datasets for active perception evaluation
• Toolkits for model evaluation
• Results and analysis from the paper