Nayana: A Foundation for Document-Centric Vision-Language Models via Multi-Task, Multimodal, and Multilingual Data Synthesis.

Adithya S. Kolavi, Samarth P, Vyoman Jain

Browse the full ICCV paper archive.