Masked Autoencoder for Distribution Estimation(MADE) for MNIST

PyTorch implementation of MADE for MNIST

Description

This is implementation of Masked Autoencoder for Distribution Estimation(MADE). To facilitate sampling numbers, I used one-hot encoding of numbers as additional input. The shape of input is batch size * 794 (not 784). This implementation includes direct connection between input and output and connectivity-agnostic training. But i didn't implement order-agonostic training.

Results

The following results were obtained with the default setting. (command: python made.py)

Reconstruction	Sampling

References

The implementation is based on:
[1] https://arxiv.org/abs/1502.03509