Skip to main content
1 of 2

why DDPG is an off-policy RL algorithm?

In DDPG, if there are no $\epsilon$-greedy and no action noise, is DDPG an on-policy algorithm?