CLIP as RNN: Segment Countless Visual Concepts without Training Endeavor Paper • 2312.07661 • Published Dec 12, 2023 • 16