Token-Driven GammaTune: Adaptive Calibration for Enhanced Speculative Decoding

Abstract

Speculative decoding accelerates large language model (LLM) inference by using a smaller draft model to propose tokens, which are then verified by a larger target model. However, selecting an optimal speculation length is critical for maximizing speedup while minimizing wasted computation. We introduce \textit{GammaTune} and \textit{GammaTune+}, training-free adaptive algorithms that dynamically adjust speculation length based on token acceptance rates using a heuristic-based switching mechanism. Evaluated on SpecBench across multiple tasks and model pairs, our method outperforms other heuristic-based approaches and fixed-length speculative decoding, achieving an average speedup of 15\% ($\pm$5\%) with \textit{GammaTune} and 16\% ($\pm$3\%) with \textit{GammaTune+}, while reducing performance variance. This makes \textit{GammaTune} a robust and efficient solution for real-world deployment.

Abstract: arXiv descriptive metadata (CC0). Thank you to arXiv for use of its open access interoperability. This service was not reviewed or approved by, nor does it necessarily express or reflect the policies or opinions of, arXiv.