🛰️ Daily AI Frontier
‹ back to 2026-09-11

OmniKVQuant: KV Cache Quantization for Omni-LLMs

Research Efficiency & Systems

Ranking

Overall 81
Content 100
Popularity 37

Observed public metrics from 1 member.

Representative image for OmniKVQuant: KV Cache Quantization for Omni-LLMs

Merged summary

TL;DR - OmniKVQuant is a training-free KV-cache quantization framework for multimodal LLMs that accounts for differences across audio, video, and text representations. It enables 2-bit caches while substantially preserving performance, reducing memory demands without reconstructing a dense FP16 cache during decoding.

  • Identifies temporal key drift and heterogeneous value geometry as key limitations of applying text-oriented TurboQuant to multimodal caches.
  • Sets key quantization ranges over short input-stream windows and applies separate value rotations for each modality.
  • Demonstrates 2-bit KV caches on Qwen2.5-Omni and Qwen3-Omni across seven audio-visual benchmarks.
  • Provides a fused Triton decoding kernel that unpacks quantized values directly during attention.

Sources (1)

OmniKVQuant: KV Cache Quantization for Omni-LLMs

arXiv cs.CV Suho Yoo, Hyunjong Ok, Jongmin Choi, Jihoo Jung, Joon Son Chung 2026-09-10 arXiv:2609.11582
Public signals Semantic Scholar citations 0 · Semantic Scholar influential citations 0
Providers: Hugging Face · N/A OpenAlex · N/A Publisher · N/A Semantic Scholar · Citations 0 · Influential citations 0 X · N/A Fetched 2026-09-24 14:22:39.158601 UTC

TL;DR - OmniKVQuant is a training-free KV-cache quantization framework for multimodal LLMs that accounts for differences across audio, video, and text representations. It enables 2-bit caches while substantially preserving performance, reducing memory demands without reconstructing a dense FP16 cache during decoding.

  • Identifies temporal key drift and heterogeneous value geometry as key limitations of applying text-oriented TurboQuant to multimodal caches.
  • Sets key quantization ranges over short input-stream windows and applies separate value rotations for each modality.
  • Demonstrates 2-bit KV caches on Qwen2.5-Omni and Qwen3-Omni across seven audio-visual benchmarks.
  • Provides a fused Triton decoding kernel that unpacks quantized values directly during attention.
item →