A Hearthstone Card visualizer in 3D with the help of Tensorboard using PCA or t-SNE
- First run the
cards.py
program, which creates the neccessary save file to run Tensorboard with. - Start Tensorboard with
tensorboard --logdir /tmp/hearthstone_cards/
- Click on the Embeddings tab
- Python 3
- Tensorflow
The card list is provided by http://hearthstoneapi.com. The program uses a local copy of cards.json
.
The card_list.py
program generates a (poorly) formatted webpage listing all the images.
It may take a long time and a require a lot of RAM to load!
If you don't want to run it, just open card_list.html
which is a local copy of the output.
- Run
card_list.py
- Open the url written in the console by the program
- Python 3
- Flask