Skip to content

Latest commit

 

History

History
15 lines (12 loc) · 948 Bytes

README.md

File metadata and controls

15 lines (12 loc) · 948 Bytes

My Masters Thesis

conditional VAE (CVAE), VQ-VAE + PixelCNN, VQ-VAE2 + Gated PixelCNN, cVAE-cGAN in tensorflow 2.x
I have not implemented the VQ-VAE2 with PixelSnail due to my poor knowledge about it.
The dataset is https://www.kaggle.com/datasets/andyczhao/covidx-cxr2

CVAE is based on: https://keras.io/examples/generative/vae/
VQ-VAEs and PixelCNN are based on: https://keras.io/examples/generative/vq_vae/ and https://github.com/rosinality/vq-vae-2-pytorch
Gated PixelCNN is based on: https://github.com/rogertrullo/Gated-PixelCNN-Pytorch
cVAE-cGAN is based on https://www.kaggle.com/code/thomasdubail/vae-gan

Reconstructions:

Reconstructions

Generations:

Generations