daniel_deng

期待这个功能，切换分类比较方便

> I've done some experiments with optimizing dlatent vs qlatent. I've observed that when optimizing qlatent against a real image (I tried a few images of celebrities), the result does...

No pro, thanks for your work!

I encounter the same issue, and finally found that I ignore to use ONLY the gradient of each discriminator part or generator part. If you don't add params like `var_list=generator_vars`,...

I remove all tensorflow pkg and do `pip3 install tensorflow-gpu==1.14 --user` this problem still exists. I add `allow_soft_placement` to gpu config and didn't solve this problem.

I tried this by adjust different learning rate to G and D, but the results is even worse.

I guess the reason of memory overflow largely due to preprocessing using pretrained-model. To overcome this, it's reasonable if I preprocess data without pretrained-model, but train model with pretrained-model checkpoint...

It is on the first stage of initializing feature extractors, in generating `FeaturisedData` before `xxx.jsonl.gz` generated. I think there are some approaches to solve the question: 1. Skip pretrain model...

SinGAN could be added

I have the same issue, I suspect this is related to the version of torch-geometry. This is my torch version: torch 2.0.1 torch-geometric 2.3.1 torch-scatter 2.1.1+pt20cu117 torchvision 0.15.2