Accelerating the inference of string generation-based chemical reaction models for industrial applications

TL;DR

This paper introduces a speculative decoding method to accelerate autoregressive SMILES-based chemical reaction models, achieving over three times faster inference without sacrificing accuracy, thereby enhancing industrial applicability.

Contribution

The authors propose a novel speculative decoding approach that significantly speeds up SMILES-to-SMILES translation models for chemical reactions, maintaining high accuracy.

Findings

Over 3X faster inference in reaction prediction

No loss in model accuracy

Applicable to molecular transformer models

Abstract

Template-free SMILES-to-SMILES translation models for reaction prediction and single-step retrosynthesis are of interest for industrial applications in computer-aided synthesis planning systems due to their state-of-the-art accuracy. However, they suffer from slow inference speed. We present a method to accelerate inference in autoregressive SMILES generators through speculative decoding by copying query string subsequences into target strings in the right places. We apply our method to the molecular transformer implemented in Pytorch Lightning and achieve over 3X faster inference in reaction prediction and single-step retrosynthesis, with no loss in accuracy.