AfriBERT Kenya Masked Language Modeling
Compare base AfriBERT against the Kenya-adapted model on Swahili, Sheng, Kenyan institutional text, M-PESA language, and English-Swahili code-switching.
1 10
How to use
Add <mask> where you want the model to predict a token.
[MASK] is also accepted and converted automatically.
For private models, set HF_TOKEN before launching the app.
The same base AfriBERT tokenizer is used for both models.
Side-by-side comparison
Detailed predictions
Examples