Lens: Rethinking Training Efficiency for Foundational Text-to-Image Models Paper • 2605.21573 • Published May 20 • 111
