Maya: Multimodal Aya

Curating a new multilingual image-text dataset and training Aya/SigLIP models with that data.

We develop an open-source multi-modal multi-lingual LLM that possesses both text and image capabilities in multiple languages (now eight).