Fetching the paper…

Unsupervised Vision-and-Language Pre-training Without Parallel Images and Captions · Around