ā Extensions to image generation and embedding models for SEA contexts
ā Evaluation across translated and region-specific multimodal benchmarks
This work is made possible by the amazing SEACrowd community, bringing together contributors across languages, cultures, and domains.