How do I create a StyleGAN2 model to generate new images?

I want to upload thousands of training images and generate a StyleGAN2 model and have access to the hyperparameters to tell it to generate new images.

I also would like to upload my existing model PKL file and train it some more.

Is this possible on your platform?

Unfortunately, we do not currently support training any GAN models. Custom models you can train include:

  • Pose estimation
  • Image Labeling
  • Image Segmentation
  • Object Detection

If you have a TensorFlow Lite or Core ML model, you can manage it in Fritz, but those are the only mobile formats supported.