The idea is to genererate Magic The Gathering cards with Deep Convolutional Generative Adversarial Network (and Tensorflow). For now the cards generated look like this:
- Docker 18.03 and up
- nvidia-docker 2.0 and up
Building docker image
git clone https://github.com/Laurans/Magic_Cards_generation.git cd Magic_Cards_generation docker build -t workspace_magic_gan .
cd Magic_Cards_generation nvidia-docker run -it -p 8888:8888 -v `pwd`:/workspace/ --name magic_gan workspace_magic_gan
It will run a jupyter server.
If you want to extract data you can find the scripts in
cd src/extract_data python get_images_links.py python extract_images.py python check_images.py
Beware to check your dataset for consistency
Training or using model
There is a jupyter notebook to train and use the model :