荔枝解码：基于混合注意力头稀疏解码的长上下文大语言模型推理加速
LycheeDecode: Accelerating Long-Context LLM Inference via Hybrid-Head Sparse Decoding
February 4, 2026
作者: Gang Lin, Dongfang Li, Zhuoen Chen, Yukun Shi, Xuhui Chen, Baotian Hu, Min Zhang
cs.AI
摘要
长上下文大语言模型（LLMs）的普及暴露了一个关键瓶颈：解码过程中快速扩张的键值缓存导致内存与延迟成本急剧上升。现有方法尝试通过跨层共享单一关键令牌集来缓解此问题，但此类粗粒度共享方案因忽视注意力头的功能多样性而损害模型性能。为此，我们提出LycheeDecode——一种以细粒度混合注意力头机制为核心的高效解码方法，该方法采用硬件友好的top-k选择策略。具体而言，基于HardKuma的新颖机制将注意力头划分为两类：小规模检索头（动态识别关键令牌）与占多数的稀疏头（复用关键令牌以实现高效计算）。通过对Llama3、Qwen3等主流模型在长文本理解（如LongBench、RULER）和复杂推理（如AIME24、OlympiadBench）多项基准上的广泛实验，我们证明LycheeDecode在生成质量上可比拟甚至有时超越全注意力基线，同时在128K上下文长度下实现最高2.7倍的加速。通过保留注意力头的功能多样性，我们的细粒度策略突破了现有方法的性能瓶颈，为高效且高质量的长上下文LLM推理提供了经过验证的有效路径。
English
The proliferation of long-context large language models (LLMs) exposes a key bottleneck: the rapidly expanding key-value cache during decoding, which imposes heavy memory and latency costs. While recent approaches attempt to alleviate this by sharing a single set of crucial tokens across layers, such coarse-grained sharing undermines model performance by neglecting the functional diversity of attention heads. To address this, we propose LycheeDecode, an efficient decoding method centered on a fine-grained hybrid-head attention mechanism that employs a hardware-efficient top-k selection strategy. Specifically, the novel HardKuma-based mechanism partitions attention heads into a small subset of retrieval heads that dynamically identify crucial tokens and a majority of sparse heads that reuse them for efficient computation. Through extensive experiments on leading models like Llama3 and Qwen3 across diverse benchmarks for long-context understanding (e.g., LongBench, RULER) and complex reasoning (e.g., AIME24, OlympiadBench), we demonstrate that LycheeDecode achieves generative quality comparable to, and at times surpassing even the full-attention baseline. Crucially, this is accomplished with up to a 2.7x speedup at a 128K context length. By preserving the functional diversity of attention heads, our fine-grained strategy overcomes the performance bottlenecks of existing methods, providing a powerful and validated pathway to both efficient and high-quality long-context LLM inference.