BASS: Batched Attention-optimized Speculative Sampling
arxiv(2024)
摘要
Speculative decoding has emerged as a powerful method to improve latency and
throughput in hosting large language models. However, most existing
implementations focus on generating a single sequence. Real-world generative AI
applications often require multiple responses and how to perform speculative
decoding in a batched setting while preserving its latency benefits poses
non-trivial challenges. This paper describes a system of batched speculative
decoding that sets a new state of the art in multi-sequence generation latency
and that demonstrates superior GPU utilization as well as quality of
generations within a time budget. For example, for a 7.8B-size model on a
single A100 GPU and with a batch size of 8, each sequence is generated at an
average speed of 5.8ms per token, the overall throughput being 1.1K tokens per
second. These results represent state-of-the-art latency and a 2.15X speed-up
over optimized regular decoding. Within a time budget that regular decoding
does not finish, our system is able to generate sequences with HumanEval
Pass@First of 43
single-sequence speculative decoding. Our peak GPU utilization during decoding
reaches as high as 15.8
and around 10X of single-sequence speculative decoding.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要