L2RPN-Baseline

The curriculumagent.baseline.CurriculumAgent is a simplified interface to run either a part or the full Teacher-Tutor-Junior-Senior pipeline. This CurriculumAgent interface is also part of the L2RPN-Baseline and can be found in the l2rpn-baselines repository.

To run the Curriculum agent first import the agent with: from curriculumagent.baseline import CurriculumAgent. Afterwards, initialize the agent with your preferred Grid2Op Environment agent = CurriculumAgent(...). Now you either have two options:

1. You can load a previously trained model and actions with agent.load(...) and then use the agent with agent.act(...). If you want to retrain the model on the same environment you can call agent.train(...), e.g., when you have new chronics.

2. Alternatively, if you don’t have no model and/or actions you can train the full pipeline with agent.train_full_pipeline(...). This will generate both actions and after a while a new model. Note that this might take some time.

After the training, you can create the output agent similar to the MyAgent method by calling agent.save(...). Further, if you want to create a submission folder, you can call agent.create_submission(...). Note that in the baseline folder, we created some the default training results on the IEEE14 grid, called “l2rpn_case14_sandbox”.