What is the maximum test set accuracy achieved by the neural network model through threshold optimization?

What is the maximum test set accuracy achieved by the neural network model through threshold optimization?: a task in data-agent-harbor-train (Harbor dataset). Files (in /home/user/input, no subfolders): - sonar.all-data.csv