This is a enhanced implementation of Variational Autoencoder. Both fully connected and convolutional encoder/decoder are built in this model. Please star if you like this implementation.
$python vae_train_amine.py # for training
$python sample.py # for sampling
- Removed standard derivation learning on Gaussian observation decoder.
- Set the standard derivation of observation to hyper-parameter.
- Add deconvolution CNN support for the Anime dataset.
- Remove Anime dataset itself to avoid legal issues.
There are two pretrained models
- Anime
- MNIST
The weights of pretrained models are locaded in weights folder