search

linfengWen98 / CAP-VSTNet

[CVPR 2023] CAP-VSTNet: Content Affinity Preserved Versatile Style Transfer
MIT License
138 stars 10 forks source link

About the dataset #6

Closed zuiaimiaomiao closed 1 year ago

zuiaimiaomiao commented 1 year ago

I would like to ask a question, is the dataset something that needs to be constructed on its own, the training process in the md file seems to use the same images for content and style.

linfengWen98 commented 1 year ago

Just give a folder directory containing all images. During training, two different batches of images will be selected.