Implement PatchGAN

[phdenzel]

For adversarial training losses (latent diffusion). To stabilize training, latent diffusion introduces an adaptive weight scheme:

• https://github.com/CompVis/latent-diffusion/blob/main/ldm/modules/losses/contperceptual.py#L32
• Eq. 7 in https://arxiv.org/pdf/2012.09841