Unlocking Efficiency in Large Language Model Inference: A Comprehensive Survey of Speculative Decoding
This is an incremental survey paper that addresses efficiency issues for users of LLMs by summarizing existing methods.
The paper tackles the problem of high inference latency in Large Language Models (LLMs) by surveying Speculative Decoding, a method that drafts and verifies multiple tokens in parallel to accelerate inference, though no concrete performance numbers are provided.
To mitigate the high inference latency stemming from autoregressive decoding in Large Language Models (LLMs), Speculative Decoding has emerged as a novel decoding paradigm for LLM inference. In each decoding step, this method first drafts several future tokens efficiently and then verifies them in parallel. Unlike autoregressive decoding, Speculative Decoding facilitates the simultaneous decoding of multiple tokens per step, thereby accelerating inference. This paper presents a comprehensive overview and analysis of this promising decoding paradigm. We begin by providing a formal definition and formulation of Speculative Decoding. Then, we organize in-depth discussions on its key facets, such as drafter selection and verification strategies. Furthermore, we present a comparative analysis of leading methods under third-party testing environments. We aim for this work to serve as a catalyst for further research on Speculative Decoding, ultimately contributing to more efficient LLM inference.