12 articles tagged with #token-pruning. AI-curated summaries with sentiment analysis and key takeaways from 50+ sources.
Researchers introduce OccamToken, a training-free method for compressing vision-language models by pruning unnecessary visual tokens while maintaining accuracy. The approach reduces visual token sequences by 98.6% (from 2,880 to 40 tokens) on LLaVA-NeXT while preserving over 93% accuracy, addressing computational bottlenecks in VLM inference.
Researchers propose VLA-Pruner, a novel token pruning method that accelerates Vision-Language-Action models for embodied AI by addressing the mismatch between semantic and action-critical visual processing. The method achieves up to 1.99x speedup while maintaining manipulation performance by considering both semantic context and temporal action relevance, unlike existing VLM pruning approaches.
SVD-Prune introduces a training-free token pruning method for Vision-Language Models using Singular Value Decomposition to reduce computational overhead. The approach maintains model performance while drastically reducing vision tokens to 16-32, addressing efficiency challenges in multimodal AI systems without requiring retraining.
Researchers developed EvoPrune, a new method that prunes visual tokens during the encoding stage of Multimodal Large Language Models (MLLMs) rather than after encoding. The technique achieves 2x inference speedup with less than 1% performance loss on video datasets, addressing efficiency bottlenecks in AI models processing high-resolution images and videos.
Researchers have developed a lightweight token pruning framework that reduces computational costs for vision-language models in document understanding tasks by filtering out non-informative background regions before processing. The approach uses a binary patch-level classifier and max-pooling refinement to maintain accuracy while substantially lowering compute demands.
Researchers introduce OC-VTP, a lightweight vision token pruning method for Vision Language Models that reduces computational overhead by selectively retaining the most representative visual tokens without requiring model fine-tuning. The approach maintains inference accuracy across all pruning ratios while providing computational efficiency gains and interpretability benefits.
Researchers introduce COAST, a novel pruning framework for vision-language models that reduces visual tokens by 77.8% while maintaining 98.64% performance and achieving 2.15x speedup. Unlike existing methods that discard low-attention tokens, COAST uses adaptive semantic routing to preserve contextually essential information, preventing 'Visual Aphasia'—a failure mode where models lose visual grounding.
Researchers introduce CLASP, a token reduction framework that optimizes Multimodal Large Language Models by intelligently pruning visual tokens through class-adaptive layer fusion and dual-stage pruning. The approach addresses computational inefficiency in MLLMs while maintaining performance across diverse benchmarks and architectures.
Researchers have developed Efficient3D, a framework that accelerates 3D Multimodal Large Language Models (MLLMs) while maintaining accuracy through adaptive token pruning. The system uses a Debiased Visual Token Importance Estimator and Adaptive Token Rebalancing to reduce computational overhead without sacrificing performance, showing +2.57% CIDEr improvement on benchmarks.
Researchers developed QAPruner, a new framework that simultaneously optimizes vision token pruning and post-training quantization for Multimodal Large Language Models (MLLMs). The method addresses the problem where traditional token pruning can discard important activation outliers needed for quantization stability, achieving 2.24% accuracy improvement over baselines while retaining only 12.5% of visual tokens.
Researchers developed E-AdaPrune, an energy-driven adaptive pruning framework that optimizes Vision-Language Models by dynamically allocating visual tokens based on image information density. The method shows up to 0.6% average improvement across benchmarks, with a notable 5.1% boost on reasoning tasks, while adding only 8ms latency per image.
Researchers developed EmbedLens, a tool to analyze how multimodal large language models process visual information, finding that only 60% of visual tokens carry meaningful image-specific information. The study reveals significant inefficiencies in current MLLM architectures and proposes optimizations through selective token pruning and mid-layer injection.
