Professor/kinyarwanda-tts-dataset-kin

Kinyarwanda TTS Dataset (Split Version)

This dataset is a reformatted version of the mbazaNLP/kinyarwanda-tts-dataset.

Modifications

• Original data was provided as a single set of 3,992 clips.
• This version has been split into Train (80%), Validation (10%), and Test (10%) sets.
• Audio files have been processed into the Hugging Face datasets format for easier loading.

Credits & Acknowledgements

Original data created and provided by Mbaza NLP. All credit for the data collection and linguistic verification goes to the original authors.