This is a pytorch implemenation of CVAE on MNIST dataset. It also provides reconstruction of images during test time and also new image generation.
To run the program
python cvae_implementaion.py
you can also control several paramenters of this program as well
python cvae_implementaion.py --lr=0.001 --batch_size=64 --itr=40 --latent_dim=40
Below is an example of image reconstruction after 10 and 40 epochs. The top section represents the original image and the bottom part is of the reconstructed images.
Reconstruction after 10 Epochs | Reconstruction after 40 Epochs |
Using CVAE we can also generate images.