TACT-KV: Tri-Axis Cosine Transform for Compressing Volumetric KV Caches in Medical VLMs
cs.CV, cs.LG
Submitted: 2026-11-05
Updated: 2026-11-05
Code: https://github.com/Chongyu1117/TACT-KV
Terminology
Sources
- M3D: Advancing 3D Medical Image Analysis with Multi-Modal Large Language Models
- Qwen3-VL Technical Report
- GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers
- Better Tokens for Better 3D: Advancing Vision-Language Modeling in 3D Medical Imaging
- Generalist Foundation Models from a Multimodal Dataset for 3D Computed Tomography
- Hulu-Med: A Transparent Generalist Model towards Holistic Medical Vision-Language Understanding
- FreqKV: Key-Value Compression in Frequency Domain for Context Window Extension
- A Study of BFLOAT16 for Deep Learning Training
- From Performance to Practice: Knowledge-Distilled Segmentator for On-Premises Clinical Workflows
- LLaVA-OneVision: Easy Visual Task Transfer
- KIVI: A Tuning-Free Asymmetric 2bit Quantization for KV Cache
- BioMedGPT: Open Multimodal Generative Pre-trained Transformer for BioMedicine
- Dual-Signal Adaptive KV-Cache Optimization for Long-Form Video Understanding in Vision-Language Models
- MedGemma Technical Report
- KV Cache Transform Coding for Compact Storage in LLM Inference
- Plug-and-Play 1.x-Bit KV Cache Quantization for Video Large Language Models
- Gemma 3 Technical Report
- VL-Cache: Sparsity and Modality-Aware KV Cache Compression for Vision-Language Model Inference Acceleration
- Fourier Compressor: Frequency-Domain Visual Token Compression for Vision-Language Models
- Revisiting Multimodal KV Cache Compression: A Frequency-Domain-Guided Outlier-KV-Aware Approach
Related papers
- TEA: Text Encoder Alignment for Robust Concept Erasure in Text-to-Image Models
- Do Egocentric Video-Language Models Capture Both Hand- and Object-Centric Cues?
- Aura: Consistent Multi-Subject Video Generation via VLM-Grounded Semantic Alignment
- PVRF: All-in-one Adverse Weather Removal via Prior-modulated and Velocity-constrained Rectified Flow
- DLWM: Diverse Latent World Models for Efficient Multimodal Reasoning
- Out of Sight, Still in Mind: Token Compression for Omni-LLMs