How can I acclerate my reinforcement learning porcess?

Typically: "How do I... ", "How can I... " questions
Post Reply
yifanzhou
Posts: 13
Joined: 03 Jun 2020, 06:17

How can I acclerate my reinforcement learning porcess?

Post by yifanzhou »

Hello,
I am using reinforcement learning with remote API (python) and synchronous operation mode. But I find that it is very slow. I wonder if it can be accelerated.
(1) The render may take some time, so if it is possible to close the render? Or start the simulation in the terminal?
(2) I tried to change the remote API function operation mode but it seems only the "simx_opmode_blocking" works.
Thanks for any advice.

coppelia
Site Admin
Posts: 10339
Joined: 14 Dec 2012, 00:25

Re: How can I acclerate my reinforcement learning porcess?

Post by coppelia »

Hello,

you can start CoppeliaSim in headless mode, via the -h command line argument. Additionally, since you are using Python, I highly recommend to have a look at the ZeroMQ-based remote API, since it is much simple to use, and can possibly speed-up the behaviour of your client.

Cheers

Post Reply