Based on the cognitive-agent example, I did some simple changes to the code using q-learning for the training. When running training the code gets stuck at env.step(action) and the ZMQ socket error occurs. What can I do to fix this issue?
Seems like this issue is related to the Gameover condition.
Hi Piotr, Hope you are doing fine.
Based on the cognitive-agent example, I did some simple changes to the code using q-learning for the training. When running training the code gets stuck at env.step(action) and the ZMQ socket error occurs. What can I do to fix this issue? Seems like this issue is related to the Gameover condition.
Is it ok to change the Gameover condition?