CoMP: Continual Multimodal Pre-training for Vision Foundation Models Paper • 2503.18931 • Published Mar 24 • 30
