VisionThink: Smart and Efficient Vision Language Model via Reinforcement Learning
Yang, Senqiao, Li, Junyi, Lai, Xin, Yu, Bei, Zhao, Hengshuang, Jia, Jiaya
–arXiv.org Artificial Intelligence
Recent advancements in vision-language models (VLMs) have improved performance by increasing the number of visual tokens, which are often significantly longer than text tokens. However, we observe that most real-world scenarios do not require such an extensive number of visual tokens. While the performance drops significantly in a small subset of OCR-related tasks, models still perform accurately in most other general VQA tasks with only 1/4 resolution. Therefore, we propose to dynamically process distinct samples with different resolutions, and present a new paradigm for visual token compression, namely, VisionThink. It starts with a downsampled image and smartly decides whether it is sufficient for problem solving. Otherwise, the model could output a special token to request the higher-resolution image. Compared to existing Efficient VLM methods that compress tokens using fixed pruning ratios or thresholds, VisionThink autonomously decides whether to compress tokens case by case. As a result, it demonstrates strong fine-grained visual understanding capability on OCR-related tasks, and meanwhile saves substantial visual tokens on simpler tasks. We adopt reinforcement learning and propose the LLM-as-Judge strategy to successfully apply RL to general VQA tasks. Moreover, we carefully design a reward function and penalty mechanism to achieve a stable and reasonable image resize call ratio. Extensive experiments demonstrate the superiority, efficiency, and effectiveness of our method. Our code is available at https://github.com/dvlab-research/VisionThink.
arXiv.org Artificial Intelligence
Jul-18-2025
- Country:
- Europe > Monaco (0.04)
- South America > Chile
- Genre:
- Research Report > New Finding (0.46)
- Technology:
- Information Technology > Artificial Intelligence
- Vision (1.00)
- Representation & Reasoning (1.00)
- Cognitive Science (1.00)
- Natural Language
- Large Language Model (1.00)
- Chatbot (0.93)
- Machine Learning
- Reinforcement Learning (1.00)
- Neural Networks > Deep Learning (0.93)
- Information Technology > Artificial Intelligence