Closed grossmanlev closed 6 years ago
The ways to speedup the code: 1) Make sure you're building with "cmake -DCMAKE_BUILD_TYPE=RelwithDebInfo" or "Release". 2) Disable sync mode (add flag --no-sync) 3) Disable Soccerwindow (add flag --headless)
It should run at several hundred FPS, depending on the number of players you have activated.
On Tue, Nov 28, 2017 at 8:04 PM, Lev Grossman notifications@github.com wrote:
Hello,
I am looking to do some RL training on an agent. However, the simulator is currently pretty slow, especially to do something like Q-Learning. Is there a way to speed up the simulator/make the graphical agent now show (if that's the slow-down)?
Thanks!
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/LARG/HFO/issues/60, or mute the thread https://github.com/notifications/unsubscribe-auth/AABNOQ87psuVo657TTIPq0XqEjg3ZKPYks5s7NfDgaJpZM4QuZBd .
Well, I have built with "RelwithDebInfo," and I am running the following command:
./bin/HFO --defense-npcs=2 --offense-npcs=2 --no-sync --fullstate --headless
However, it takes about 4 seconds per every 100 frames or so.
Does this command run for you at several hundred FPS? If so, any ideas what I could have done wrong in the setup?
Thanks!
Sorry, you'll want to remove the --no-sync flag and it should go much faster.
On Wed, Nov 29, 2017 at 12:08 PM, Lev Grossman notifications@github.com wrote:
Well, I have build with "RelwithDebInfo," and I am running the following command:
./bin/HFO --defense-npcs=2 --offense-npcs=2 --no-sync --fullstate --headless
However, it takes about 4 seconds per every 100 frames or so.
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/LARG/HFO/issues/60#issuecomment-347980872, or mute the thread https://github.com/notifications/unsubscribe-auth/AABNOaHI5dtISzW18Rsa7TwM2TSM2kuzks5s7bnOgaJpZM4QuZBd .
Oh great! Thank you so much!
Hello,
I am looking to do some RL training on an agent. However, the simulator is currently pretty slow, especially to do something like Q-Learning. Is there a way to speed up the simulator/make the graphical agent now show (if that's the slow-down)?
EDIT: Headless doesn't seem to make it go any faster. Also, for reference, I am using attacker npcs -- is this the cause of the slowness? As npcs on their own are also as slow as my learning agent.
Thanks!