token pruning
- Attribution-Driven Adaptive Token Pruning for Transformers
- Balanced Token Pruning: Accelerating Vision Language Models Beyond Local Optimization
- Don't Just Chase “Highlighted Tokens” in MLLMs: Revisiting Visual Holistic Context Retention
- HoliTom: Holistic Token Merging for Fast Video Large Language Models
- Twilight: Adaptive Attention Sparsity with Hierarchical Top-$p$ Pruning
- VisionThink: Smart and Efficient Vision Language Model via Reinforcement Learning