Dataset · chimechallenge
CHiME-5 Audio Dataset for Noisy Environments
chimechallenge/chime5
Dinner party audio recordings for speech separation and recognition.
What's inside
Contains English-language audio recordings collected during 20 different real-world home dinner parties. The dataset size is under 1,000 examples (download archives).
What you can do with it
- Training and testing speech separation and ASR systems in noisy environments
- Researching speech recognition tasks in party and domestic settings
- Languages
- English
- Size
- up to 1K rows
- Downloads this month
- 250
- Likes
- 26
How to load it
from datasets import load_dataset
ds = load_dataset("chimechallenge/chime5")