stylegan-t

stylegan-t copied to clipboard

What is the expected date of code release?

9

piotr-komorowski

Licensing

2

Is this project gonna be forbidden to be used for commercial purposes for free?

andzejsp

There seem to be aliasing issues with the resulting images. Which version was used for stylegan-t?

TortoiseFanboy

The mechanism for modulation in figure 3b is quite strange. Based on the below figure, I guess that the modulation style s multiply to the feature output from previous block...

thuanz123

Why divide by 'batch_size' when calculating GAN loss? > loss_gen = (F.relu(torch.ones_like(gen_logits) + gen_logits)).mean() / batch_size > loss_real = (F.relu(torch.ones_like(real_logits) - real_logits)).mean() / batch_size > loss_gen = (-gen_logits).mean() / batch_size...

higun2

Dear authors, Thanks for the amazing work. I'm reaching out to ask you an opinion about a problem I am currently facing. I am working with an architecture that is...

manigalati

Hey. Is there any chance you have this dataset cached locally and can send it to me? https://huggingface.co/datasets/ChristophSchuhmann/improved_aesthetics_6plus I'm going to train miniSDXL (like lambdalabs/miniSD-diffusers, but for SDXL) and need...

kopyl

abt the PATH_TO_NETWORK_PKL

2

Can you provide a link to the pre-trained model?. Thanks

BoxFishLab

Hyperparameters not matched between code and paper

2

Thanks for your great work. In Table 4 from the paper EMA is 0.9978, but in the code [generator.py line504](https://github.com/autonomousvision/stylegan-t/blob/36ab80ce76237fefe03e65e9b3161c040ae888e3/networks/generator.py#LL504) it's 0.995. Which is the actual value? Also the batch...

keyu-tian

run in ubuntu

2

when I run the code in ubuntu, which has more than one gpu, it appeared the below error: (pytorch) m11113013@htsc-Server3:~/ProjectCode/MasterProject2$ python -m torch.distributed.run --standalone --nproc_per_node 1 train.py --outdir ./training-runs/ --cfg...

lcxsnow