Swedish Whispers; Leveraging a Massive Speech Corpus for Swedish Speech Recognition

TL;DR

This paper introduces fine-tuned Whisper models for Swedish speech recognition, trained on a large, diverse dataset, achieving significant improvements over existing models in reducing word error rates.

Contribution

It presents a new Swedish speech recognition model trained on an unprecedented dataset, demonstrating substantial performance gains over prior multilingual models.

Findings

47% reduction in WER compared to OpenAI's whisper-large-v3

Improved performance across multiple Swedish speech datasets

Effective fine-tuning of multilingual models for underrepresented languages

Abstract

This work presents a suite of fine-tuned Whisper models for Swedish, trained on a dataset of unprecedented size and variability for this mid-resourced language. As languages of smaller sizes are often underrepresented in multilingual training datasets, substantial improvements in performance can be achieved by fine-tuning existing multilingual models, as shown in this work. This work reports an overall improvement across model sizes compared to OpenAI's Whisper evaluated on Swedish. Most notably, we report an average 47% reduction in WER comparing our best performing model to OpenAI's whisper-large-v3, in evaluations across FLEURS, Common Voice, and NST.