The code was implemented for the Garment Transfer obejctive, where top cloth of the source image was rendered on the target image.
The improvement of the code with the hyper-parameter tunings are under process. The inferenece code with better results will be updated soon.
The LIP dataset(Image-based Multi-pose Virtual Try On) was used to train the model.