Dataset · Ethnic-Voices-Tz
Speech and Texts in Six Tanzanian Languages
Ethnic-Voices-Tz/ethnic-voices-tanzania
A corpus of 462 hours of audio recordings and texts in rare ethnic languages of Tanzania for training ASR and language models.
What's inside
It contains 91,872 speech audio recordings totaling 462.07 hours along with text data. The dataset covers six low-resource languages of Tanzania: Sukuma, Hehe, Zinza, Runyambo, Subi, and Luguru. The resource was created specifically to develop speech and language technologies for rare languages.
What you can do with it
- Training and fine-tuning automatic speech recognition (ASR) models for Tanzanian languages
- Adapting multilingual text generation models for rare African languages
- Evaluating speech synthesis and recognition quality in low-resource languages
- Tasks
- Automatic speech recognition, Text generation
- Size
- 10K–100K rows
- Downloads this month
- 54.4K
- Likes
- 0
How to load it
from datasets import load_dataset
ds = load_dataset("Ethnic-Voices-Tz/ethnic-voices-tanzania")