When Text Matters: Design Principles for Visual Token Pruning in Vision-Language Model

arXiv:2609.34861 · cs.CV, cs.LG · Submitted 2026-09-28 · Read on arXiv

cs.CV, cs.LG

Submitted: 2026-09-28

Updated: 2026-09-28

Code: https://github.com/kmc3661/DeFT

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Pruning — 91× in this paper · explained in 4 episodes
• Benchmark — 21× in this paper · explained in 2 episodes
• Conference on Computer Vision and Pattern Recognition — 8× in this paper
• Ablation — 6× in this paper · explained in 1 episode
• Quantization — 4× in this paper · explained in 12 episodes
• Alignment — 3× in this paper · explained in 1 episode
• CVF International Conference on Computer Vision — 3× in this paper
• Multimodal Reasoning — 3× in this paper · explained in 5 episodes
• Redundancy — 3× in this paper · explained in 2 episodes
• Transformer — 3× in this paper · explained in 2 episodes
• Accuracy — 2× in this paper · explained in 1 episode
• Distillation — 2× in this paper · explained in 2 episodes
• Instruction Tuning — 2× in this paper · explained in 2 episodes
• Knowledge Distillation — 2× in this paper · explained in 11 episodes
• Ablation Studies — 1× in this paper · explained in 1 episode
• Adaptation — 1× in this paper · explained in 1 episode
• Foundation Model — 1× in this paper · explained in 3 episodes
• Foundation Models — 1× in this paper · explained in 7 episodes
• Korea Advanced Institute of Science and Technology — 1× in this paper
• Multimodal Learning — 1× in this paper · explained in 2 episodes
• Vision-Language Models (VLMs) — 1× in this paper · explained in 6 episodes

Sources

• LLaVA-OneVision-1.5: Fully Open Framework for Democratized Multimodal Training
• Qwen3-VL Technical Report
• Token Merging: Your ViT But Faster
• Improving Visual Token Reduction via Rectifying Distortions for Efficient Multimodal LLM Inference
• The Llama 3 Herd of Models
• Distilling the Knowledge in a Neural Network
• LLaVA-OneVision: Easy Visual Task Transfer
• Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution
• Fine-Grained Post-Training Quantization for Large Vision Language Models with Quantization-Aware Integrated Gradients
• PyramidDrop: Accelerating Your Large Vision-Language Models via Pyramid Visual Redundancy Reduction
• SparseVLM: Visual Token Sparsification for Efficient Vision-Language Model Inference

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