Saved in:
Bibliographic Details
Main Authors: Huang, Chien-yu, Shih, Min-Han, Lu, Ke-Han, Hsiao, Chi-Yuan, Lee, Hung-yi
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2408.13891
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910576191471616
author Huang, Chien-yu
Shih, Min-Han
Lu, Ke-Han
Hsiao, Chi-Yuan
Lee, Hung-yi
author_facet Huang, Chien-yu
Shih, Min-Han
Lu, Ke-Han
Hsiao, Chi-Yuan
Lee, Hung-yi
contents Instruction-based speech processing is becoming popular. Studies show that training with multiple tasks boosts performance, but collecting diverse, large-scale tasks and datasets is expensive. Thus, it is highly desirable to design a fundamental task that benefits other downstream tasks. This paper introduces a multi-talker speaking style captioning task to enhance the understanding of speaker and prosodic information. We used large language models to generate descriptions for multi-talker speech. Then, we trained our model with pre-training on this captioning task followed by instruction tuning. Evaluation on Dynamic-SUPERB shows our model outperforming the baseline pre-trained only on single-talker tasks, particularly in speaker and emotion recognition. Additionally, tests on a multi-talker QA task reveal that current models struggle with attributes such as gender, pitch, and speaking rate. The code and dataset are available at https://github.com/cyhuang-tw/speechcaps.
format Preprint
id arxiv_https___arxiv_org_abs_2408_13891
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle SpeechCaps: Advancing Instruction-Based Universal Speech Models with Multi-Talker Speaking Style Captioning
Huang, Chien-yu
Shih, Min-Han
Lu, Ke-Han
Hsiao, Chi-Yuan
Lee, Hung-yi
Computation and Language
Audio and Speech Processing
Instruction-based speech processing is becoming popular. Studies show that training with multiple tasks boosts performance, but collecting diverse, large-scale tasks and datasets is expensive. Thus, it is highly desirable to design a fundamental task that benefits other downstream tasks. This paper introduces a multi-talker speaking style captioning task to enhance the understanding of speaker and prosodic information. We used large language models to generate descriptions for multi-talker speech. Then, we trained our model with pre-training on this captioning task followed by instruction tuning. Evaluation on Dynamic-SUPERB shows our model outperforming the baseline pre-trained only on single-talker tasks, particularly in speaker and emotion recognition. Additionally, tests on a multi-talker QA task reveal that current models struggle with attributes such as gender, pitch, and speaking rate. The code and dataset are available at https://github.com/cyhuang-tw/speechcaps.
title SpeechCaps: Advancing Instruction-Based Universal Speech Models with Multi-Talker Speaking Style Captioning
topic Computation and Language
Audio and Speech Processing
url https://arxiv.org/abs/2408.13891