👁 FlashVGGT Logo

FlashVGGT

Efficient and Scalable Visual Geometry Transformers with Compressed Descriptor Attention

3D reconstruction from multi-view images is a core challenge in computer vision. Recently, feed-forward methods have emerged as efficient and robust alternatives to traditional per-scene optimization techniques. Among them, state-of-the-art models like the Visual Geometry Grounding Transformer (VGGT) leverage full self-attention over all image tokens to capture global relationships. However, this approach suffers from poor scalability due to the quadratic complexity of self-attention and the large number of tokens generated in long image sequences. In this work, we introduce FlashVGGT, an efficient alternative that addresses this bottleneck through a descriptor-based attention mechanism. Instead of applying dense global attention across all tokens, FlashVGGT compresses spatial information from each frame into a compact set of descriptor tokens. Global attention is then computed as cross-attention between the full set of image tokens and this smaller descriptor set, significantly reducing computational overhead. Moreover, the compactness of the descriptors enables online inference over long sequences via a chunk-recursive mechanism that reuses cached descriptors from previous chunks. Experimental results show that FlashVGGT achieves reconstruction accuracy competitive with VGGT while reducing inference time to just 9.3% of VGGT for 1,000 images, and scaling to sequences exceeding 3,000 images.

Efficiently compress dense image tokens into a compact set of spatial descriptors via resampling.

Replace quadratic self-attention with linear cross-attention between image queries and descriptor keys/values.

Cache and reuse descriptors across temporal chunks to scale seamlessly to sequences of thousands of frames.

● Scroll to zoom in/out

● Drag to rotate

● Press "shift" and drag to pan

● Scroll to zoom in/out

● Drag to rotate

● Press "shift" and drag to pan