DeepFashion is a large-scale fashion image dataset that contains over 800,000 diverse fashion images ranging from well-posed shop images to unconstrained consumer photos.
We use part of DeepFashion to implement our editing system. All the codes are written in Pytorch
DeepFashion can be used to train and evaluate algorithms of fashion retrieval, fashion synthesis, and GANs for fashion generation and editing.
- Demo of interactive fashion image editing
- Google Drive: downloading link
- Google Drive: downloading link
Extract the dataset and checkpoints with the same directory of demo.py
The use of this software is RESTRICTED to non-commercial research and educational purposes.