AI can't speak my language.
8 million people speak Tiv. Google can't pronounce it. ChatGPT gets it wrong. Not a single AI system can serve them.
I'm building the solution.
SAVA Languages the first voice dataset for the Tiv language. Starting in Benue State.
We just published the SLANGUAGE Tiv Parallel Corpus on Hugging Face.
Tiv is spoken by 8 million people in Benue State, Nigeria. Yet it remains one of the most underrepresented languages in AI and NLP research globally.
We just published the SLANGUAGE Tiv Parallel Corpus on Hugging Face.
Tiv is spoken by 8 million people in Benue State, Nigeria. Yet it remains one of the most underrepresented languages in AI and NLP research globally.
Every sentence manually verified for tonal accuracy.
This is v1. A larger corpus of 7,000+ sentences is already complete, with audio recordings underway for TTS and ASR model training.
Tiv is tonal. It is complex. And it deserves to be represented in AI.