Good Token Hunting: A Hitchhiker's Guide to Token Selection for Visual Geometry Transformers

Hugging Face Daily Papers Papers

Summary

This paper introduces a two-stage token selection framework for visual geometry transformers that reduces computational costs by restricting key/value tokens during global attention, achieving over 85% acceleration on scenes with 500 images while maintaining baseline performance.

Visual geometry transformers have become powerful architectures for multi-view 3D reconstruction, enabling joint prediction of multiple 3D attributes in a feed-forward manner. However, their computational cost grows quadratically with the input sequence length due to the global attention layers inside these models. This limits both their scalability and efficiency. In this work, we address this challenge with a simple yet general strategy: restricting the number of key/value tokens that each query interacts with during global attention. To achieve effective token selection, we introduce a two-stage framework. First, an inter-frame selection step operates at the frame level to identify frames that should be preserved. Second, an intra-frame selection step further discards more redundant tokens within the selected frames. Our analysis highlights the advantage of a diversity-based strategy for inter-frame selection, which ensures broad coverage of the scene. For intra-frame selection, we show that layer-aware sparsification is necessary, with the selection process guided by the entropy of the global attention pattern. Our approach offers a superior speed-accuracy trade-off compared to existing solutions. Extensive experiments show that it accelerates visual geometry transformers by over 85% for scenes with 500 images while maintaining, or even improving, baseline performance, which hints that how our token selection strategy can play a crucial role in future applications of visual geometry transformers. Our project website is available at https://zsh2000.github.io/good-token-hunting.github.io.
Original Article
View Cached Full Text

Cached at: 05/25/26, 06:36 AM

Paper page - Good Token Hunting: A Hitchhiker’s Guide to Token Selection for Visual Geometry Transformers

Source: https://huggingface.co/papers/2605.23892

Abstract

Visual geometry transformers are accelerated through a two-stage token selection framework that reduces computational costs while maintaining performance.

Visual geometry transformershave become powerful architectures formulti-view 3D reconstruction, enabling joint prediction of multiple 3D attributes in a feed-forward manner. However, their computational cost grows quadratically with the input sequence length due to theglobal attention layersinside these models. This limits both their scalability and efficiency. In this work, we address this challenge with a simple yet general strategy: restricting the number of key/value tokens that each query interacts with during global attention. To achieve effectivetoken selection, we introduce a two-stage framework. First, aninter-frame selectionstep operates at the frame level to identify frames that should be preserved. Second, anintra-frame selectionstep further discards more redundant tokens within the selected frames. Our analysis highlights the advantage of a diversity-based strategy forinter-frame selection, which ensures broad coverage of the scene. Forintra-frame selection, we show thatlayer-aware sparsificationis necessary, with the selection process guided by the entropy of the global attention pattern. Our approach offers a superior speed-accuracy trade-off compared to existing solutions. Extensive experiments show that it acceleratesvisual geometry transformersby over 85% for scenes with 500 images while maintaining, or even improving, baseline performance, which hints that how ourtoken selectionstrategy can play a crucial role in future applications ofvisual geometry transformers. Our project website is available at https://zsh2000.github.io/good-token-hunting.github.io.

View arXiv pageView PDFProject pageGitHub4Add to collection

Get this paper in your agent:

hf papers read 2605\.23892

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2605.23892 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2605.23892 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2605.23892 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

The Galaxy's Guide to the Tokenizer: A Benchmark for Scientific Foundation Models

Hugging Face Daily Papers

This paper compares four tokenization methods (Affine, AIM, JetFormer, VQ-VAE) for astronomical images within a unified transformer framework, using 640,000 galaxy images to evaluate reconstruction quality, physical property prediction, and morphological preservation. It finds that no single method excels across all tasks, highlighting trade-offs in representation learning.

ReToken: One Token to Improve Vision-Language Models for Visual Retrieval

Hugging Face Daily Papers

ReToken introduces a single learnable retrieval token that selects sparse query-relevant visual tokens from a pre-filled visual KV cache, improving long-context vision-language model performance on visual retrieval tasks. It yields consistent gains across image and video benchmarks while enabling efficient long-video inference on a single H100.

From 2D Grids to 1D Tokens: Reforming Shared Representations for Multimodal Image Fusion

Hugging Face Daily Papers

This paper introduces a multimodal image fusion method that uses a 1D token interface from a pretrained image tokenizer to enhance global appearance coherence while preserving local details through selective token editing (STE). Experiments on four benchmarks show state-of-the-art performance in both global coherence and local fidelity.