Parameters for deep q-network using tensorflow-agents and other general optimisation

Question

I have a couple of questions about tensorflow-agents and deep Q-learning. It contains 2 parts, it\'s ok to answer any of them.

In cartpole example of t