/binarygan

Code for "Training Generative Adversarial Networks with Binary Neurons by End-to-end Backpropagation"

Primary LanguagePython

BinaryGAN

Prepare Training Data

  • Download MNIST database by running the script:

    ./training_data/download_mnist.sh

  • or download it manually:

    1. Download MNIST database here
    2. Decompress all the .gz files
    3. Move the decompressed files to ./training_data/mnist

  • Store the data to shared memory (optional)

    Make sure the SharedArray package has been installed.

    python ./training_data/load_mnist_to_sa.py ./training_data/mnist/ \
--merge --binary

Configuration

Modify config.py for configuration.

  • Quick setup

    Change the values in the dictionary SETUP for a quick setup. Documentation is provided right after each key.

  • More configuration options

    Four dictionaries EXP_CONFIG, DATA_CONFIG, MODEL_CONFIG and TRAIN_CONFIG define experiment-, data-, model- and training-related configuration variables, respectively.

    The automatically-determined experiment name is based only on the values defined in the dictionary SETUP, so remember to provide the experiment name manually when you modify any other configuration variables so that you won't overwrite a trained model.

Train the model

python train.py