An Effective Training Framework for Light-Weight Automatic Speech Recognition Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Hannan, Abdul, Brutti, Alessio, Nawaz, Shah, Noman, Mubashir
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916763898216448
author Hannan, Abdul
Brutti, Alessio
Nawaz, Shah
Noman, Mubashir
author_facet Hannan, Abdul
Brutti, Alessio
Nawaz, Shah
Noman, Mubashir
contents Recent advancement in deep learning encouraged developing large automatic speech recognition (ASR) models that achieve promising results while ignoring computational and memory constraints. However, deploying such models on low resource devices is impractical despite of their favorable performance. Existing approaches (pruning, distillation, layer skip etc.) transform the large models into smaller ones at the cost of significant performance degradation or require prolonged training of smaller models for better performance. To address these issues, we introduce an efficacious two-step representation learning based approach capable of producing several small sized models from a single large model ensuring considerably better performance in limited number of epochs. Comprehensive experimentation on ASR benchmarks reveals the efficacy of our approach, achieving three-fold training speed-up and up to 12.54% word error rate improvement.
format Preprint
id arxiv_https___arxiv_org_abs_2505_16991
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle An Effective Training Framework for Light-Weight Automatic Speech Recognition Models
Hannan, Abdul
Brutti, Alessio
Nawaz, Shah
Noman, Mubashir
Computer Vision and Pattern Recognition
Recent advancement in deep learning encouraged developing large automatic speech recognition (ASR) models that achieve promising results while ignoring computational and memory constraints. However, deploying such models on low resource devices is impractical despite of their favorable performance. Existing approaches (pruning, distillation, layer skip etc.) transform the large models into smaller ones at the cost of significant performance degradation or require prolonged training of smaller models for better performance. To address these issues, we introduce an efficacious two-step representation learning based approach capable of producing several small sized models from a single large model ensuring considerably better performance in limited number of epochs. Comprehensive experimentation on ASR benchmarks reveals the efficacy of our approach, achieving three-fold training speed-up and up to 12.54% word error rate improvement.
title An Effective Training Framework for Light-Weight Automatic Speech Recognition Models
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2505.16991