These are examples of GAN/VAE architectures, some of which are from the book: "Hands-On Image Generation with Tensorflow" This includes: Image -> Image Translation techniques like:
- Pix2Pix (Supervised)
- GauGAN (Supervised, but MultiModel, For Segmentation Masks)
- CycleGAN (Unsupervised)
- BicycleGAN (Supervised, MultiModel, Any paired dataset)
- StarGAN (Unsupervised, MultiDomain)
Stylization techniques like:
- Nueral Style Transfer
- Arbitrary Style Transfer
High Fidelity Techniques:
- PgGAN (High Res, very slow and lower quality)
- StyleGAN (High Res, very high quality, quite slow)
Wide Range Techniques:
- SelfAttention GAN (Low Res)
- BigGAN (High Res)