Make Some Noise: Unlocking Language Model Parallel Inference Capability through Noisy Training

TL;DR

This paper introduces MSN, a noise-based training framework that enhances language model parallel inference speed by 2.3-2.7x without altering core capabilities, using simple noise addition and a new decoding strategy.

Contribution

The paper presents MSN, a novel noise-injection training method that improves parallel decoding in language models without requiring structural modifications.

Findings

MSN improves inference speed by 2.3-2.7x times.

MSN maintains original task performance.

Comparable acceleration to state-of-the-art models with additional structure.

Abstract

Existing speculative decoding methods typically require additional model structure and training processes to assist the model for draft token generation. This makes the migration of acceleration methods to the new model more costly and more demanding on device memory. To address this problem, we propose the Make Some Noise (MSN) training framework as a replacement for the supervised fine-tuning stage of the large language model. The training method simply introduces some noise at the input for the model to learn the denoising task. It significantly enhances the parallel decoding capability of the model without affecting the original task capability. In addition, we propose a tree-based retrieval-augmented Jacobi (TR-Jacobi) decoding strategy to further improve the inference speed of MSN models. Experiments in both the general and code domains have shown that MSN can improve inference speed by 2.3-2.7x times without compromising model performance. The MSN model also achieves comparable acceleration ratios to the SOTA model with additional model structure on Spec-Bench.