hs366399 / Image-Super-Resolution-Using-VAE-GAN-with-PyTorch

The model uses the AE-GAN (Autoencoder Generative Adversarial Network) architecture for generating upsampled images. The model is trained on Celeb-A image (1024 x 1024) dataset where input image is of 128x128 and generated image is of shape 480x480.
14Updated 4 years ago

Related projects: