TernaryCLIP: Efficiently Compressing Vision-Language Models with Ternary Weights and Distilled Knowledge

Open in new window