Data Generation:
 - follow the readme install instructions in generator/
 - generate state-value pairs for your problems (see generator/README.md)

Training Setup:
 - install the python requirements in strips-hgn (pip install -r requirements.txt)
 - In the following we assume that my_domain is the domain your training data is
 based on. Create a folder strips-hgn/experiments/ijcai21/json_data/my_domain/json_data
 and copy the *-worker-i.json files generated during data generation in this folder.
 - Create a symlink in strips-hgn/experiments/ijcai21/json_data/my_domain to the
 directory on which your training data is based on
  - e.g. ln -s ~/benchmarks/my-domain/ problems
  - copy the script strips-hgn/experiments/ijcai21/json_blocksworld.py and adapt
  it to your needs (change base_directory to json_data/my_domain)
 - running the adapted scripts trains strips HGN networks for your training data
 - models are saved in strips-hgn/results/


