ViQ: Text-Aligned Visual Quantized Representations at Any Resolution
2026-06-25 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors present ViQ, a new way to represent images that balances both detailed visuals and meaningful content by using discrete tokens, similar to text. They train their model in two steps: first aligning visual features with language understanding, then carefully compressing these features while preserving detail. This approach allows ViQ to handle images at their original resolution and work well across different tasks. The authors show that ViQ performs competitively with other vision models but is more efficient, speeding up multimodal training significantly.
multimodal representationdiscrete visual tokenstext-aligned pre-trainingfeature quantizationnative-resolution inputsemantic featureslow-level reconstructionmultimodal vision encoderlarge language models (LLMs)training efficiency
Authors
Xumin Yu, Zuyan Liu, Zhenyu Yang, Yuhao Dong, Shengsheng Qian, Jiwen Lu, Han Hu, Yongming Rao
Abstract
A unified representation for text and vision is a natural pursuit, as it enables simpler multimodal modeling and more efficient training. However, representing images as discrete signals in the same way as text inevitably introduces severe information loss. Existing work struggles to balance low-level details and high-level semantics in discrete representations: reconstruction-oriented representations often lack semantic information, whereas semantically stronger features typically suffer from severe loss of detail. We present ViQ, a Visual Quantized Representations framework, which is designed to balance semantics and details in discrete representations while supporting inputs at native resolutions, thereby enabling it to serve as a unified and general discrete representation for arbitrary visual inputs. Our approach structures quantization learning into two stages: text-aligned pre-training and feature discretization. With text-aligned pre-training, we enhance the visual encoder semantic-rich supervision from the pretrained language model and enable it to process native-resolution visual inputs. During discretization, we propose a proximal representation learning strategy to progressively compact the feature space, along with a position-aware head-wise quantization mechanism that enables flexible processing of arbitrary resolutions. Extensive experiments on multimodal tasks demonstrate that ViQ achieves competitive performance compared to state-of-the-art multimodal vision encoders with continuous and high-dimensional visual features, while maintaining high precision in low-level reconstruction. We also show that multimodal training with visual quantized representations largely improves efficiency, yielding up to 20\%-70\% acceleration with different base LLMs and training recipes.