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.


LIDLABSNeural language identification across Nigerian languages (YorΓΉbΓ‘, Igbo, Hausa, Pidgin, English).

LIDLABSZero-dependency language identification across Nigerian languages (YorΓΉbΓ‘, Igbo, Hausa, Pidgin, English).

LIDLABSNeural language identification across Nigerian languages (YorΓΉbΓ‘, Igbo, Hausa, Pidgin, English).

LIDLABSfastText language identification across 25 world languages.

LIDSTABLENeural language identification across 25 world languages, tuned for long-form text: documents, articles, and passages.

LIDSTABLENeural language identification across 25 world languages, tuned for short text: search queries, chat messages, and user input.
More of the stack, organised the same way.