Autoencoders

PyTorch implementation of various autoencoder architectures using the MNIST dataset.
Results are in the jupyter notebooks

Architectures implemented :

1. Vanilla autoencoder (using FCN)
2. Denoising autoencoder (using FCN)(using convolutions)

TO-DO :

• Sparse autoencoder
• Variational autoencoder
• Beta-VAE

Resources

1. From Autoencoder to beta-VAE (including research papers cited in the blog)