I try to do a reinforcement learning enviroment with tf_agents in Tensorflow. Is it possible to dynamically change the size of the observation array?
For example I wa