unrecall

why does the target action a' in Q(s', a') for training critic net directly come from the target actor net

1

Hi, jimkon In the original paper, the action for training critic net comes from the full policy. But, in your master, the action is just given by the target actor...

when i change discount_factor from 0.95 to 1.0, the AC algorithm can't converge to a optimal policy. Besides sometimes Continuous MountainCar Actor Critic Solution works at first 30 or 40...