ExKaldi-RT: A Real-Time Automatic Speech Recognition Extension Toolkit of Kaldi

TL;DR

ExKaldi-RT is a Python-based real-time speech recognition toolkit built on Kaldi, enabling easy development of online ASR systems with neural network integration and demonstrating competitive performance on LibriSpeech.

Contribution

It introduces a Python-compatible extension of Kaldi for online ASR, facilitating research and development with neural network models in real-time recognition.

Findings

Achieved competitive real-time ASR performance on LibriSpeech

Provides an easy-to-use Python interface for Kaldi-based online recognition

Supports neural network-based signal processing and decoding models

Abstract

This paper describes the ExKaldi-RT online automatic speech recognition (ASR) toolkit that is implemented based on the Kaldi ASR toolkit and Python language. ExKaldi-RT provides tools for building online recognition pipelines. While similar tools are available built on Kaldi, a key feature of ExKaldi-RT that it works on Python, which has an easy-to-use interface that allows online ASR system developers to develop original research, such as by applying neural network-based signal processing and by decoding model trained with deep learning frameworks. We performed benchmark experiments on the minimum LibriSpeech corpus, and it showed that ExKaldi-RT could achieve competitive ASR performance in real-time recognition.