DesCLIP: Robust Continual Adaptation via General Attribute Descriptions for Pretrained Vision-Language Models

Open in new window