Closed keenranger closed 4 years ago
You can control the rate at which the agents train using the sc2.yaml config file: pymarl/src/config/envs/sc2.yaml
adjust the step multiplier to 1x. step_mul: 8
watch agents train.
I prefer to set to evaluate mode in the default.yaml and watch through there after I have completed training but this is the quicker way to view agent actions.
@GJHall step_mul actually corresponds to the number of game steps per agent step. Changing it should have a big impact on the policy.
@keenranger what you can do is you can import time
and put a time.sleep(ms)
in the step()
function of the StarCraft2Env. This is probably the fastest way watching the agents' performance. Alternatively, you can safe a replay and then rewatch it.
Hope this helps.
I want to know my agent works well by watching game, but game speed is so fast that I don
t know what
s happening.I know that I can change the in-game speed by '+' and '-' key, but even if I slow down in-game speed, game is still too fast to watch.
Is there any command or something to slow down game speed??