A Contrastive Learning Approach for Training Variational Autoencoder Priors

Open paper

Benchmark	Model	Rank	Results
image-generation-on-celeba-256x256	NCP-VAE	–	FID: 24.79
image-generation-on-celeba-64x64	NCP-VAE	–	FID: 5.25
image-generation-on-cifar-10	NCP-VAE	–	FID: 24.08