ET-Prune: Evidence-Aware Dynamic Budgeting for Visual Token Pruning in Text-Rich MLLMs
cs.CV, cs.CL
Submitted: 2026-08-03
Updated: 2026-08-27
Comments: Code and supplementary material is at https://github.com/Labyrinth0419/ET-Prune
Code: https://github.com/Labyrinth0419/ET-Prune
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Qwen3-VL Technical Report
- MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models
- OCRBench v2: An Improved Benchmark for Evaluating Large Multimodal Models on Visual Text Localization and Reasoning
- ZOO-Prune: Training-Free Token Pruning via Zeroth-Order Gradient Estimation in Vision-Language Models
- G squared TR: Generation-Guided Visual Token Reduction for Separate-Encoder Unified Multimodal Models
- Mixture-of-Depths: Dynamically allocating compute in transformer-based language models
- RoboVQA: Multimodal Long-Horizon Reasoning for Robotics
- InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency
- PyramidDrop: Accelerating Your Large Vision-Language Models via Pyramid Visual Redundancy Reduction
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