# Audio Datasets

Speech and audio datasets for ML experiments.

## Contents

### LibriSpeech (train-clean-100)
- train-clean-100.tar.gz (~6 GB)
- ~100 hours of clean read English speech
- 16 kHz, multiple speakers
- Useful for ASR (automatic speech recognition) experiments

## License
CC BY 4.0 (refer to LibriSpeech original release)