Steps required to train a StyleGan-XL for a custom dataset

[MiguelCosta94]

What would be the steps required to train the network for a new dataset like the Visual Wake Words?

[xhlho]

Hello, have you completed the training on your own dataset? If it is implemented, please let me know what needs to be noted. Thank you in advance!

[MiguelCosta94]

Sorry, I did not finish the training. I ended up using another GAN architecture.

[xhlho]

Thank you for your answer. I have another small request. Could you please let me know which open source code it is? Thank you!

[MiguelCosta94]

I've used this framework to implement my GAN https://github.com/POSTECH-CVLab/PyTorch-StudioGAN/tree/master

[amanikiruga]

I followed the instructions laid out in the /.README.MD

specifically https://github.com/autonomousvision/stylegan-xl/tree/main?tab=readme-ov-file#data-preparation
and
https://github.com/autonomousvision/stylegan-xl/tree/main?tab=readme-ov-file#training

also be sure to follow the instructions in the dataset_tool.py in the javadocs written for the function convert_dataset

stylegan-xl/dataset_tool.py

Line 343 in 4241ff9

def convert_dataset(

Feel free to ask me any questions here as needed.