paper-with-me

Papers

RoPE-Aware Bit Allocation for KV-Cache Quantization

2026-06-23 · Fengfeng Liang, Yuechen Zhang, Jiaya Jia arxiv

Existing low-bit KV-cache quantizers often treat each cached key as a flat vector. Under RoPE, however, a key's contribution to a future attention logit decomposes into a position-dependent sum over two-dimensional frequency blocks. This makes key-cache quantization a block-wise bit-allocation problem: high-energy RoPE blocks are more sensitive to quantization error and should receive more bits. We introduce Block-GTQ, a RoPE-aware bit allocator for key-cache quantization built on TurboQuant-MSE(TQ-MSE). For each layer and KV head, Block-GTQ computes a label-free energy score for each RoPE block and greedily allocates integer bit widths by marginal gain. Under matched K/V bit budgets, Block-GTQ better preserves RoPE query-key logits on a ten-model diagnostic panel, cutting per-layer MAE by 32-80% at 2 and 3 b/dim K-only quantization and winning all 367/367 layer comparisons against uniform TQ-MSE. These fidelity gains translate to stronger downstream long-context retrieval, understanding, and reasoning. At K2V2 on Llama-3.1-8B-Instruct, Block-GTQ raises the six-task NIAH average from 70.6 to 97.4, and the LongBench-EN average from 36.87 to 53.31. On AIME 2024/2025 with DeepSeek-R1-Distill-Qwen-7B, without an fp16 recent-key buffer, Block-GTQ at K3V2 scores 51.7/37.5, close to fp16's 54.2/37.9, whereas uniform TQ-MSE collapses to 0.0/0.0. We further implement a packed-cache serving path. On a single H800 GPU with Qwen2.5-3B-Instruct, packed K3V3 achieves 3.24x KV-cache compression with fp16-comparable quality, runs 1.34x faster than fp16 FlashAttention2 at 128K context, reduces peak memory from 56.31 GB to 19.85 GB, and remains feasible at 256K and 512K where fp16 OOMs. Code is available at https://github.com/JIA-Lab-research/blockgtq.

📄 PDF Abstract BibTeX arXiv:2606.24033

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

KVmix: Gradient-Based Layer Importance-Aware Mixed-Precision Quantization for KV Cache

2025-05-18 · Fei Li, Song Liu, Weiguo Wu, Shiqiang Nie 외

The high memory demands of the Key-Value (KV) Cache during the inference of Large Language Models (LLMs) severely restrict their deployment in resource-constrained platforms. Quantization can effectively alleviate the me…

Quantization

SVDq: 1.25-bit and 410x Key Cache Compression for LLM Attention

2025-02-21 · Hong Yankun, Li Xing, Zhen Hui-Ling, Yu Xianzhi 외

For the efficient inference of Large Language Models (LLMs), the effective compression of key-value (KV) cache is essential. Three main types of KV cache compression techniques, namely sparsity, channel compression, and …

Quantization

RotateKV: Accurate and Robust 2-Bit KV Cache Quantization for LLMs via Outlier-Aware Adaptive Rotations

2025-01-25 · Zunhai Su, Zhe Chen, Wang Shen, Hanyu Wei 외

Key-Value (KV) cache facilitates efficient large language models (LLMs) inference by avoiding recomputation of past KVs. As the batch size and context length increase, the oversized KV caches become a significant memory …

Computational EfficiencyGSM8KQuantization

PM-KVQ: Progressive Mixed-precision KV Cache Quantization for Long-CoT LLMs

2025-05-24 · Tengxuan Liu, Shiyao Li, Jiayi Yang, Tianchen Zhao 외

Recently, significant progress has been made in developing reasoning-capable Large Language Models (LLMs) through long Chain-of-Thought (CoT) techniques. However, this long-CoT reasoning process imposes substantial memor…

Quantization

RDKV: Rate-Distortion Bit Allocation for Joint Eviction and Quantization of the KV Cache

2026-05-08 · Junkai Zhang, Hang Guo, Luca Benini, Yawei Li arxiv

Large language models (LLMs) have shown strong performance across diverse tasks, but their inference with long input contexts is bottlenecked by memory size and bandwidth. The Key-Value (KV) cache size grows linearly wit…