Skip to content

Dataset · KRAFTON

Massive English speech corpus for text-to-speech (TTS)

KRAFTON/Raon-OpenTTS-Pool

615 thousand hours of annotated English speech for training voice synthesis models.

What's inside

A large-scale open text-audio pair corpus in English created by KRAFTON to train the Raon-OpenTTS model. It includes 615 thousand hours of speech (239.7 million segments) collected from 8 open speech datasets and internet recordings. All data is formatted uniformly.

What you can do with it

  • Training from scratch and fine-tuning SOTA text-to-speech (TTS) models
  • Training multi-speaker voice models and voice cloning systems
Tasks
Text → speech
Languages
English
Size
100M–1B rows
Downloads this month
3.4K
Likes
44

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("KRAFTON/Raon-OpenTTS-Pool")