Skip to content

Dataset · HKUSTAudio

Contextual TTS Dataset for ISCSLP 2026

HKUSTAudio/ISCSLP2026-CoT-TTS

Audio data from movies and TV shows for training context-dependent and expressive speech synthesis.

What's inside

Created for the ISCSLP 2026 CoT-TTS Challenge based on dialogs from movies, radio, and television dramas in Chinese and English. Contains examples with emotional variations and scene changes.

What you can do with it

  • Training chain-of-thought speech generation models (CoT-TTS)
  • Developing emotional and context-oriented voice synthesis systems
Tasks
Text → speech
Languages
Chinese, English
Downloads this month
5.3K
Likes
22

How to load it

from datasets import load_dataset

ds = load_dataset("HKUSTAudio/ISCSLP2026-CoT-TTS")