Reproducing results on latest Starcraft version (5.0.9)

[ilyalasy]

Hi, thank you for your work!
I downloaded pretrained weights ("rl_22-02-07_16-26-48.pth") and then i ran:

from alphastarmini.core.rl import rl_vs_inner_bot_mp
rl_vs_inner_bot_mp.test(on_server=P.on_server, replay_path=P.replay_path)

Unfortunately, agent is only building probes, and definitely not acting according to anyhow optimal policy.
I am using latest SC2 version, is this the reason? Is it possible to provide model weights for a new version?

[liuruoze]

Yes, it is the reason.

Yes, we will provide a model for the new SC2 version in the future version.

You can wait for it or download a 3.16.1 version to experience our project.