Conformer-based Hybrid ASR System for Switchboard Dataset

TL;DR

This paper explores the use of conformer architecture in hybrid ASR systems, demonstrating improved performance and training efficiency on the Switchboard dataset compared to traditional models.

Contribution

It introduces a conformer-based hybrid ASR training recipe, including time downsampling and transposed convolutions, and evaluates its effectiveness on the Switchboard dataset.

Findings

Achieves competitive word-error-rate on Switchboard 300h dataset.

Outperforms BLSTM-based hybrid models significantly.

Generalizes well on Hub5'01 test set.

Abstract

The recently proposed conformer architecture has been successfully used for end-to-end automatic speech recognition (ASR) architectures achieving state-of-the-art performance on different datasets. To our best knowledge, the impact of using conformer acoustic model for hybrid ASR is not investigated. In this paper, we present and evaluate a competitive conformer-based hybrid model training recipe. We study different training aspects and methods to improve word-error-rate as well as to increase training speed. We apply time downsampling methods for efficient training and use transposed convolutions to upsample the output sequence again. We conduct experiments on Switchboard 300h dataset and our conformer-based hybrid model achieves competitive results compared to other architectures. It generalizes very well on Hub5'01 test set and outperforms the BLSTM-based hybrid model significantly.