Nya-Həba Named Entity Recognition Corpus v1.0

The Nya-Həba Named Entity Recognition Corpus v1.0 is a manually annotated language resource for Nya-Həba (Kilba), a low-resource Chadic language spoken primarily in Adamawa State, Nigeria. The corpus contains 3,184 sentence instances and 44,308 tokens annotated using the BIO scheme for Person (PER), Location (LOC), Time (TIME), and Animal (ANIM) entities. The release includes a cleaned canonical CoNLL corpus, leakage-safe benchmark partitions, a strict unique-text benchmark, corpus statistics, annotation documentation, validation utilities, and an audit of duplicate-text annotation conflicts.