I saw several comparison between SGD, RMSPROP and ADAM but what I am looking for is their comparsion in DQN algorithm?
What is best to select as optimizer SGD or Adam in DQN? Why?
Please check the accuracy of my thinking:
-What is their main differences?
Adam is more stable than SGD in training of DQN?
Speed of convergence of ADAM is more in training of DQN? why? Because of considering 2 coefficient?
In Testing case SGD with momentum is more stable than ADAM?why?
-the speed of convergence of Adam and adaptive approach is more higher but in long time SGD has a better performance? if correct why?
-What we are able to say in comparison between ADAM and RMSPROP?