JoeyS2T: Minimalistic Speech-to-Text Modeling with JoeyNMT

TL;DR

JoeyS2T is a minimalist, easy-to-use speech-to-text toolkit built on JoeyNMT, offering competitive performance with a simple, integrated workflow for speech recognition and translation tasks.

Contribution

It extends JoeyNMT with speech-specific components, creating a unified, accessible toolkit for speech-to-text modeling that maintains simplicity and competitive accuracy.

Findings

Performs competitively on speech recognition benchmarks.

Provides an integrated, easy-to-use pipeline from data to evaluation.

Maintains simplicity while including key speech modeling features.

Abstract

JoeyS2T is a JoeyNMT extension for speech-to-text tasks such as automatic speech recognition and end-to-end speech translation. It inherits the core philosophy of JoeyNMT, a minimalist NMT toolkit built on PyTorch, seeking simplicity and accessibility. JoeyS2T's workflow is self-contained, starting from data pre-processing, over model training and prediction to evaluation, and is seamlessly integrated into JoeyNMT's compact and simple code base. On top of JoeyNMT's state-of-the-art Transformer-based encoder-decoder architecture, JoeyS2T provides speech-oriented components such as convolutional layers, SpecAugment, CTC-loss, and WER evaluation. Despite its simplicity compared to prior implementations, JoeyS2T performs competitively on English speech recognition and English-to-German speech translation benchmarks. The implementation is accompanied by a walk-through tutorial and available on https://github.com/may-/joeys2t.