Not All Tasks Quantize Equally: Fisher-Guided Quantization for Visual Geometry Transformer

arXiv:2605.15828 · cs.CV · Submitted 2026-05-15 · Read on arXiv

cs.CV

Submitted: 2026-05-15

Updated: 2026-09-27

Code: https://github.com/ypzhng/FGQ

Terminology

• Quantization — 89× in this paper · explained in 12 episodes
• Calibration — 69× in this paper · explained in 7 episodes
• AUC — 26× in this paper · explained in 2 episodes
• Transformer — 21× in this paper · explained in 2 episodes
• Fisher Information — 15× in this paper · explained in 3 episodes
• Accuracy — 14× in this paper · explained in 1 episode
• Fisher-Guided Quantization — 7× in this paper
• Post-Training Quantization — 6× in this paper
• Visual Geometry Grounded Transformer — 5× in this paper
• Consistency — 4× in this paper · explained in 3 episodes
• Benchmark — 3× in this paper · explained in 2 episodes
• Ablation — 2× in this paper · explained in 1 episode
• Ablation study — 2× in this paper · explained in 1 episode
• Pruning — 2× in this paper · explained in 4 episodes
• Aggregation — 1× in this paper · explained in 1 episode
• Area Under the Curve — 1× in this paper
• Area Under the Curve (AUC) — 1× in this paper · explained in 1 episode
• Computational overhead — 1× in this paper · explained in 2 episodes
• Elias — 1× in this paper · explained in 2 episodes
• Fidelity — 1× in this paper · explained in 2 episodes
• Gaussian Splatting — 1× in this paper · explained in 4 episodes
• Generalization — 1× in this paper · explained in 13 episodes
• KV cache — 1× in this paper · explained in 3 episodes
• L Alternating-Attention — 1× in this paper
• Normalized Performance Degradation — 1× in this paper
• Reconstruction Error — 1× in this paper · explained in 2 episodes
• Redundancy — 1× in this paper · explained in 2 episodes

Sources

• Quantized Visual Geometry Grounded Transformer
• GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers
• G$^2$VLM: Geometry Grounded Vision Language Model with Unified 3D Reconstruction and Spatial Reasoning
• VGD: Visual Geometry Gaussian Splatting for Feed-Forward Surround-view Driving Reconstruction
• SpinQuant: LLM quantization with learned rotations
• DINOv2: Learning Robust Visual Features without Supervision
• Tail-Aware Post-Training Quantization for 3D Geometry Models
• FastVGGT: Training-Free Acceleration of Visual Geometry Transformer
• LiteVGGT: Boosting Vanilla VGGT via Geometry-aware Cached Token Merging
• AVGGT: Rethinking Global Attention for Accelerating VGGT
• $\pi^3$: Permutation-Equivariant Visual Geometry Learning
• FlashVGGT: Efficient and Scalable Visual Geometry Transformers with Compressed Descriptor Attention
• Fisher-aware Quantization for DETR Detectors with Critical-category Objectives
• Robo3R: Enhancing Robotic Manipulation with Accurate Feed-Forward 3D Reconstruction
• VersaQ-3D: Architecture Support for Visual Geometry Grounded Transformers via Versatile Quantization
• Stereo Magnification: Learning View Synthesis using Multiplane Images

Related papers

• Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
• AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
• Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
• MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
• TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
• A Survey on Efficient Vision-Language-Action Models