arXiv：cs.AI· Chence Yang, Ningxi Cheng, Arash Akbari, Qitao Tan, Qingchan Zhu, Ci Zhang, Changdi Yang, Yanzhi Wang, Wei Niu, Jinhui Wang, Jin Lu, Geng Yuan·· 9 小时前AI 评分35
BitNest：位嵌套投机解码实现内存高效 LLM 推理加速
BitNest: Bit-Nested Speculative Decoding for Memory-Efficient LLM Inference Acceleration
AI 导读
BitNest 是一种位嵌套投机解码框架，将低精度草稿模型直接嵌入高精度目标模型的权重表示中，使两者共享单一物理权重，无需额外草稿模型。在多个 7B–8B 边缘友好 LLM 上，BitNest 平均投机接受率达 95.2%，端到端解码较 FP16 自回归提速 1.48–1.61 倍，并将该渐进精度设计扩展至 KV cache 以支持长上下文推理。
正文
Authors:Chence Yang, Ningxi Cheng, Arash Akbari, Qitao Tan, Qingchan Zhu, Ci Zhang, Changdi Yang, Yanzhi Wang, Wei Niu, Jinhui Wang, Jin Lu, Geng Yuan
Abstract:Speculative decoding accelerates autoregressive generation by using a lightweight draft to propose multiple tokens for parallel verification. However, existing methods often require an additional draft model or weight representation, introducing non-negligible memory overhead on resource-constrained devices. Self-speculative approaches reduce this overhead, yet still face trade-offs between draft quality, target quality, and storage efficiency. We propose BitNest, a bit-nested speculative decoding framework that embeds a low-precision draft directly into the higher-precision target representation. Instead of deriving a draft from a predefined target, BitNest first constructs a strong low-precision base and then recovers the higher-precision target through residual refinement, enabling both models to share a single physical weight representation. BitNest further extends this progressive-precision design to the KV cache for long-context inference. Across multiple 7B--8B edge-friendly LLMs and diverse workloads, BitNest achieves an average speculative acceptance rate of 95.2% while closely preserving higher-precision model quality, and delivers 1.48--1.61x end-to-end speedup over FP16 autoregressive decoding. On the LLaMA models supported by all representative self-speculative baselines, BitNest also achieves consistently competitive or higher decoding speedup.
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.02800 [cs.AI] |
| (or arXiv:2610.02800v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.02800 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Chence Yang [view email]
[v1]
Fri, 2 Oct 2026 04:43:42 UTC (257 KB)
来源：arXiv：cs.AI · arxiv.org