Spaces:
No application file
Sauti Ya Kenya: Building an Authentic Kenyan Swahili TTS
Hello everyone,
I'm excited to introduce Sauti Ya Kenya, our latest project at Msingi AI! This initiative is dedicated to developing a truly Kenyan Swahili Text-to-Speech (TTS) system that captures the local accent, intonation, and unique speech patterns of native speakers.
Why Sauti Ya Kenya?
Currently, many TTS systems for Swahili sound foreign and fail to reflect the natural way Kenyan Swahili is spoken. With Sauti Ya Kenya, we aim to:
- Fine-tune existing models using high-quality, locally sourced speech data.
- Create a TTS system that resonates with native Kenyan speakers by accurately reproducing the local accent.
- Ensure that our voices are preserved and represented authentically in the digital space.
What We Need:
We are at the early stages and would love to collaborate with the community. Specifically, we’re looking for:
- Native Kenyan Swahili speakers: Help us record voice samples.
- Linguists & Language Experts: Assist in annotating and providing feedback on local dialects and nuances.
- Developers & TTS/NLP Specialists: Collaborate on fine-tuning and improving our model.
- Everyone: Share your ideas, suggestions, or questions to help us shape this project.
Your participation can make a huge impact on ensuring that Kenyan Swahili is accurately and authentically represented in AI. Let’s work together to build a system that not only speaks our language but does it right.
Feel free to comment below or reach out directly if you're interested in joining this effort!
Thank you,
Kiplangat Korir
Founder, Msingi AI