This is an implementation of a variational autoencoder using convolutional layers. I will show an example of using it with a sketch dataset.
Project convnet2
python train.py -mode train -config configs/sbir_vae.config -name SBIR
python train.py -mode predict -config configs/sbir_vae.config -name SBIR