search

NVlabs / stylegan3

Official PyTorch implementation of StyleGAN3
Other
6.38k stars 1.12k forks source link

Input Channels > 3 #201

Open sorooshpoorgholi opened 1 year ago

sorooshpoorgholi commented 1 year ago

How can I train the model on a dataset where the input channels are more than 3 (e.g 4)?

PDillis commented 1 year ago

You'll have to modify the dataset_tool.py and architecture of the Generator and Discriminator, so that the number of channels can be different from 1 or 3. I've done this already in my repo, so try it and let me know if it works (I tested it with some dummy RGBA data and it worked, but note that you can't use all types of augmentations with RGBA data).