Barlow Graph Auto Encoder

Setup

Easiest way is to use docker and vscode.

1. Build the image from given file as:

docker build -t bgae:latest -f Dockerfile .

2. Use vscode remote-container to open the project inside docker container. Afterward, you can execute run_for_dataset.py for a single dataset, or run_hptune.py if you want to play around with hyperparameter tuning.

Config files

dataset_config.yaml is used for run_for_dataset.py and hptune_config.yaml is used for run_hptune.py. In both the files, there is a base configuration that can be patched by dataset-specific configurations as demonstrated by the examples inside these config files.