adviser_reisekosten

1. Install requirements into your python environment: pip install -r requirements.txt
2. Data is in train_graph.json and test_graph.json
3. If training multiple models, you can start and enable a redis cache in train_chatbot.py and then execute sh start_cache.sh
4. To train a policy, change the configuration as needed in train_chatbot.py and execute python train_chatbot.py (requires GPU - make sure to deactivate caching options in configuration if run without redis)