In DDPG, if there are no $\epsilon$-greedy and no action noise, is DDPG an on-policy algorithm?
why DDPG is an off-policy RL algorithm?
GoingMyWay
- 150
- 8