VoxZip: Semantic-Anchored Temporal KV Cache Compression for Long-Context Audio Inference
Wenxu Jia, Dongjie Fu, Xize Cheng, Fangming Feng, Linjun Li, Wenshi Chen, Yingming Li, Zhou Zhao, Tao Jin
cs.AI, cs.SD
Submitted: 2026-08-09
Updated: 2026-08-11
Code: https://github.com/MM-Speech/VoxZip
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- BLAB: Brutally Long Audio Bench
- PyramidKV: Dynamic KV Cache Compression based on Pyramidal Information Funneling
- X-OPD: Cross-Modal On-Policy Distillation for Capability Alignment in Speech LLMs
- An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models
- Character Beyond Speech: Leveraging Role-Playing Evaluation in Audio Large Language Models via Reinforcement Learning
- Audio Flamingo 2: An Audio-Language Model with Long-Audio Understanding and Expert Reasoning Abilities
- AudioMarathon: A Comprehensive Benchmark for Long-Context Audio Understanding and Efficiency in Audio LLMs
- WavChat: A Survey of Spoken Dialogue Models
- Kimi-Audio Technical Report
- A Survey on Large Language Model Acceleration based on KV Cache Management
- Prompt Compression for Large Language Models: A Survey
- ChunkKV: Semantic-Preserving KV Cache Compression for Efficient Long-Context LLM Inference
- KV Cache Compression for Inference Efficiency in LLMs: A Review
- Efficiently Scaling Transformer Inference
- Robust Speech Recognition via Large-Scale Weak Supervision
- D2O: Dynamic Discriminative Operations for Efficient Long-Context Inference of Large Language Models
- Qwen2.5-Omni Technical Report
- Qwen3-Omni Technical Report
- GLM-4-Voice: Towards Intelligent and Human-Like End-to-End Spoken Chatbot
- A Survey on Efficient Inference for Large Language Models
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems