VAE-GAN in MXNet

• Implementation of Autoencoding beyond pixels using a learned similarity metric, based on the Tensorflow implementation: https://github.com/JeremyCCHsu/tf-vaegan

• Please refer to their official Github for details: Autoencoding Beyond Pixels

• As the name indicates, VAE-GAN replaces GAN's generator with a variational auto-encoder, resulting in a model with both inference and generation components.

Experiements

• Dataset: caltech 101 silhouettes dataset from https://people.cs.umass.edu/~marlin/data.shtml

Prerequisites

• Opencv
• Python packages required: scipy, scikit-learn and Pillow, opencv python package

Environment Tested On

Deep Learning AMI (Ubuntu) - 2.0, p2.8xlarge

Usage

If you want to train and test with the default options do the following:

1. Download the default dataset and convert from matlab file format to png file format

python convert_data.py

2. Train on the downloaded dataset and store the encoder model and generator model params.

python vaegan_mxnet.py --train

3. Test on the downloaded dataset

python vaegan_mxnet.py --test --testing_data_path /home/ubuntu/datasets/caltech101/test_data

• Using existing models:

python vaegan_mxnet.py --test --testing_data_path [your dataset image path] --pretrained_encoder_path [pretrained encoder model path] --pretrained_generator_path [pretrained generator model path] [options]

• Train a new model:

python vaegan_mxnet.py --train --training_data_path [your dataset image path] [options]

• Training on the CPU:

python vaegan_mxnet.py --train --use_cpu --training_data_path [your dataset image path] [options]