Browse Datasets

AfriSpeech-200

Pan-African Speech Corpus

asr speech-recognition benchmarking 200 hours CC BY-NC-SA 4.0
Access Dataset

Description

200-hour Pan-African speech corpus. Useful for benchmarking ASR model efficiency on mobile devices with 120 African accents.

Quick Start

Python 
from datasets import load_dataset

dataset = load_dataset("intronhealth/afrispeech-200", split="test")

Project Ideas

Raspberry Pi ASR

Deploy quantized ASR model on Raspberry Pi with AfriSpeech

Advanced

Ethical Considerations

  • Clinical data privacy (curated)