Join SEA-VL: Build Culturally Relevant Vision-Language Models for Southeast Asia

Contribute to SEA-VL, a initiative creating high-quality vision-language datasets for Southeast Asian languages. Help build inclusive AI resources and co-author publications. Join our community on GitHub and Discord.

We invite you to contribute to SEA-VL, an open-source initiative to create high-quality vision-language datasets for Southeast Asian languages. In Phase 1, we’re collecting self-taken, culturally relevant images and captions until February 15, 2025. By joining us, you can help build publicly available vision-based resources for SEA languages, fostering inclusivity in AI research. Active contributors will be considered co-authors for the publication targeted for ACL 2025 or EMNLP 2025.

To participate, share this initiative, contribute culturally-relevant images with captions, or offer feedback to strengthen the dataset creation process. Visit our SEA-VL announcement page to learn more and join our community on GitHub and Discord.

Tags: SEA-VL, vision-language models, Southeast Asia, AI research, dataset creation, ACL 2025, EMNLP 2025, GitHub, Discord