The Hmar Heritage Foundation has released a Hugging Face Space for HmarBERT and made its associated datasets publicly available. The initiative aims to provide digital resources for the Hmar language, which previously lacked a significant online presence.
The project includes a model inference space hosted on Hugging Face by user azinamotoe. The organization also maintains a dataset repository under the hmar-heritage-org account. The team invites feedback on dataset structure and future development steps.
This release provides a foundational resource for generating text in Hmar, addressing the language's historical absence from digital platforms.