As for my reinforcement learning network, the dimension of “action” is 2.
I have to chose a distribution for it, then sample from the distribution to choose action
Can I just choose to generate two seperate distributions for it, or have to choose other distributions, such as GMM or multivariat normal?