Xingang Pan

Hi, is there any chance of merging this pull request? :-)

@Prannoy-Noel Hi, thanks for your interest. This repo might be helpful: https://github.com/XingangPan/OCDA-Driving-Example

@seyeon956 Thanks for your interest in our work. Here are my answers: 1. We use vgg16_bn provided by torchvision. 2. Yes. The overall process is as what you said. 3....

@Nadavc220 For question2, we follow the architecture of https://github.com/wasidennis/AdaptSegNet/blob/master/model/deeplab_vgg.py And added batchnorm to the architecture. The feature before the classifier is used for class memory bank. As described in the...

@zhangxiangyu19 It means the class is unknown and is chosen from a set of random samples.

@wty-ustc Yes, the result is consistent with what I tried. Optimizing just for latent code would produce very inaccurate reconstruction for BigGAN, even if you enlarge the learning rate. That's...

For church, you should use checkpoint at checkpoints/stylegan2/stylegan2-church-config-f.pt (--size 256 --channel_multiplier 2). You can find these settings at https://github.com/XingangPan/GAN2Shape/blob/main/configs/church.yml

@kekqiu You may use this repo to transfer .pkl to pytorch weights: https://github.com/rosinality/stylegan2-pytorch

Hi, you need to first perform GAN inversion to get the latent code of the images. To do this you may use code from https://github.com/rosinality/stylegan2-pytorch/blob/master/projector.py or use our code here...

@DarKayserLeo Hi, please run the script under https://github.com/rosinality/stylegan2-pytorch or https://github.com/XingangPan/GAN2Shape/tree/main/gan2shape/stylegan2/stylegan2-pytorch