This repository includes the program aggregating the results of reinforcement learning agents and visualizing them for my research.

pip install -r requirements.txt

Please write file patterns with the rules used by the Unix shell in fourrooms/config.json. The files is output to steps directory by the learning program of four-room domain. The path pattern example is out/dynamic_human_20221024_1650/steps/*.

Please write file patterns with the rules used by the Unix shell in pinball/config.json. The files is output to steps directory by the learning program of pinball domain. The path pattern example is out/dynamic_human_v3/steps/*.csv.

Befor running the analyzing program, you must export a csv file from tensorboard output files. Please modify file paths in transform.py.

cd pick_and_place
python transform.py

cd fourrooms
python analyze.py 

cd pinball
python analyze.py

cd pick_and_place
python analyze.py