ElegantRL
ElegantRL copied to clipboard
A confusion about helloworld_DQN_single_file.py
Why is warm up for ReplayBuffer used in helloworld_DQN_single_file.py but not in elegantrl?
Warming up for replaybuffer is not necessary for off-policy DRL algorithm. I plan to cancel warning up.
Warming up for replaybuffer is not necessary for off-policy DRL algorithm. I plan to cancel warning up. Thanks for your answer, I also found that Warming up for replaybuffe is not a necessary step for DQN, and there are many example codes that do not use this.