- Place the
scenario2_cyborg--
file in thescenario
directory of Cyborg. - Replace the
linux_user_host_image1
file in theimage
folder of Cyborg with the new image file. - Use the modified_train.py for training the agent (this updates the action space to the reduced one)
- Update your training script to point to the new scenario file
scenario2_cyborg--
. - The action space for the blue team has been reduced from 145 to 144 due to the reduced action space. Update the action space in your training script accordingly. Use the provided enumerated action list for scenario2 (
blue_enum_action.txt
) or for the modified scenario (blue_enum_action_cyborg--.txt
) to select the appropriate action space. - For the Cardiff agent, decrease all actions by 1 in the training script to align with the updated action space.
- Update your evaluation script (evaluation.py)to point to the new scenario file
scenario2_cyborg--
. - If you are using cardiff_uni structure, since the action space for the blue team has been reduced from 145 to 144 due to the reduced action space, we need to update the action space in your `MainAgent.py' script accordingly. Modified evalaution script and MainAgent script can be downloaded.