Dataset · psdn-ai
Multilingual crowdsourced speech corpus in 16 languages
psdn-ai/CROWS-multilingual
A large audio corpus with text-reading recordings and annotations for training ASR and speech classification systems.
What's inside
The corpus from Poseidon AI includes 821,361 text-reading audio recordings from 46,298 speakers across 16 languages (metadata notes Arabic, German, English, Spanish, French, and Hindi). Each recording is accompanied by source text, automatic transcription, English translation, as well as audio quality and voice consistency ratings.
What you can do with it
- Training and fine-tuning multilingual ASR models
- Audio classification and filtering recordings by speech quality and speaker
- Tasks
- Automatic speech recognition, Audio classification
- Languages
- Arabic, German, English, Spanish, French, Hindi, Indonesian, Japanese and 4 more
- Size
- 100K–1M rows
- Downloads this month
- 14K
- Likes
- 5
How to load it
from datasets import load_dataset
ds = load_dataset("psdn-ai/CROWS-multilingual")