Yanbei Chen

To get those images, it is suggested to contact the authors of FashionIQ.

Hi, This part can be done by pre-training the base network using the same loss, but the different texts (i.e. tagged captions in fasion200k). So the difference is only to...