
Two coverage tiers and multiple efficiency options: zero-dependency logistic-regression classifiers, fastText, and neural models built on XLM-R, AfriBERTa, and our own ModernBERT Nigerian encoder.


LIDNeural language identification across Nigerian languages (YorĂčbĂĄ, Igbo, Hausa, Pidgin, English).

LIDZero-dependency language identification across Nigerian languages (YorĂčbĂĄ, Igbo, Hausa, Pidgin, English).

LIDNeural language identification across Nigerian languages (YorĂčbĂĄ, Igbo, Hausa, Pidgin, English).

LIDfastText language identification across 25 world languages.

LIDNeural language identification across 25 world languages.

LIDNeural language identification across 25 world languages.
More of the stack, organised the same way.

Frontier-merged LLMs: from MIST-Mini-8B to MIST-1-140B, including reasoning-tuned Thinking variants.

Diacritic & tone restoration for YorĂčbĂĄ, Igbo, Hausa and beyond.

Vision models: super-resolution upscaling, denoising, and image steganography.