I want to use RL algorithm to train my project. But I cannot find that on flow directory, how many types of algorithms I could use, and there are some explain document of these algorithms, or If I want to create a custom algorithm, how to do that?
Asked
Active
Viewed 103 times
1 Answers
1
I'm assuming you've installed rllib to start training with Flow. If you go to the agents
directory located under Ray/ray/python/ray/rllib/agents/
you can find the supported algorithms such as PPO and DDPG.

Rayyan
- 210
- 2
- 14
-
yes, I find this directory, but it is not under flow directory, it is ok? – user11795736 Jul 25 '19 at 02:17
-
Yes that shouldn't be an issue, train away! – Rayyan Jul 25 '19 at 02:20