search

RenYurui / Global-Flow-Local-Attention

The source code for paper "Deep Image Spatial Transformation for Person Image Generation"
https://renyurui.github.io/GFLA-web
Other
566 stars 84 forks source link

About the training data resolution #59

Closed ChongjianGE closed 4 years ago

ChongjianGE commented 4 years ago

Hi @RenYurui , Thanks for the great work. I wonder what's the resolution of your training data?

ChongjianGE commented 4 years ago

Meanwhile, could you please help share the detailed file structure of the training data? Thank you.

RenYurui commented 4 years ago

I think you want to ask the resolution of the training images. We use 256 * 256 images for training. You can download the demo images to check the structure of the dataset by using ./download.sh