Dataset · MikhailT
English audiobook corpus for speech synthesis
MikhailT/hifi-tts
Large multi-speaker dataset of English audio recordings and texts for building high-quality speech synthesis systems.
What's inside
The set includes speech audio recordings from multiple speakers and corresponding English text fragments. Collection size is 100k to 1M examples. Data compiled from LibriVox public domain audiobooks and Project Gutenberg texts.
What you can do with it
- Training multi-speaker text-to-speech models with natural speaker intonation
- Voice cloning and adapting speech synthesis systems for reading long texts
- Tasks
- Text → speech, Text → audio
- Languages
- English
- Size
- 100K–1M rows
- Downloads this month
- 4.4K
- Likes
- 41
Models trained on it
- 11mlabs/indri-0.1-124m-ttsText → speech 982
- mradermacher/indri-0.1-124m-tts-GGUF 300
- mradermacher/indri-0.1-350m-tts-GGUF 262
- Lwasinam/voiceraText → speech 256
- mmmhusto/TTS-EnglishText → speech 143
- Backup-bdg/Xoron-Dev-MultiMoeMultimodal 87
- TuringLogic/little-ttsText → speech 57
- 11mlabs/indri-0.1-350m-ttsText → speech 41
How to load it
from datasets import load_dataset
ds = load_dataset("MikhailT/hifi-tts")