TreeVGR-7B

This repository contains the TreeVGR-7B model, as presented in the paper Traceable Evidence Enhanced Visual Grounded Reasoning: Evaluation and Methodology.

TL; DR: We propose TreeBench, the first benchmark specially designed for evaluating "thinking with images" capabilities with traceable visual evidence, and TreeVGR, the current state-of-the-art open-source visual grounded reasoning models.

Abstract

Models like OpenAI-o3 pioneer visual grounded reasoning by dynamically referencing visual regions, just like human "thinking with images". However, no benchmark exists to evaluate these capabilities holistically. To bridge this gap, we propose TreeBench (Traceable Evidence Evaluation Benchmark), a diagnostic benchmark built on three principles: (1) focused visual perception of subtle targets in complex scenes, (2) traceable evidence via bounding box evaluation, and (3) second-order reasoning to test object interactions and spatial hierarchies beyond simple object localization. Prioritizing images with dense objects, we initially sample 1K high-quality images from SA-1B, and incorporate eight LMM experts to manually annotate questions, candidate options, and answers for each image. After three stages of quality control, TreeBench consists of 405 challenging visual question-answering pairs, even the most advanced models struggle with this benchmark, where none of them reach 60% accuracy, e.g., OpenAI-o3 scores only 54.87. Furthermore, we introduce TreeVGR (Traceable Evidence Enhanced Visual Grounded Reasoning), a training paradigm to supervise localization and reasoning jointly with reinforcement learning, enabling accurate localizations and explainable reasoning pathways. Initialized from Qwen2.5-VL-7B, it improves V* Bench (+16.8), MME-RealWorld (+12.6), and TreeBench (+13.4), proving traceability is key to advancing visual grounded reasoning.

For more details, please refer to the GitHub repository.

Installation

To get started, first clone the repository and install the required dependencies:

git clone https://github.com/Haochen-Wang409/TreeVGR
cd TreeVGR
pip3 install -r requirements.txt
pip3 install flash-attn --no-build-isolation -v

Usage

This repository provides a simple local inference demo of TreeVGR on TreeBench. After installation, you can run the inference script:

python3 inference_treebench.py

This should produce an output similar to:

Perception/Attributes 18/29=62.07
Perception/Material 7/13=53.85
Perception/Physical State 19/23=82.61
Perception/Object Retrieval 10/16=62.5
Perception/OCR 42/68=61.76
Reasoning/Perspective Transform 19/85=22.35
Reasoning/Ordering 20/57=35.09
Reasoning/Contact and Occlusion 25/41=60.98
Reasoning/Spatial Containment 20/29=68.97
Reasoning/Comparison 20/44=45.45
==> Overall 200/405=49.38
==> Mean IoU: 43.3

Note: This result is slightly different from the paper, as we mainly utilized VLMEvalKit for a more comprehensive evaluation.

Citation

If you find this work useful for your research and applications, please cite using the following BibTeX:

@article{wang2025traceable,
  title={Traceable Evidence Enhanced Visual Grounded Reasoning: Evaluation and Methodology},
  author={Haochen Wang and Xiangtai Li and Zilong Huang and Anran Wang and Jiacong Wang and Tao Zhang and Jiani Zheng and Sule Bai and Zijian Kang and Jiashi Feng and Zhuochen Wang and Zhaoxiang Zhang},
  journal={arXiv preprint arXiv:2507.07999},
  year={2025}
}