What is the game configuration in the online evaluation?

We have not yet decided on this, but we will release all details before the evaluation system opens. Also for your reference, one of the most important goals of this challenge is to encourage research on learning generalized agent behavior in the open-world environment. So we expect to have some environmental diversity in the evaluation. Participants are also encouraged to train their agents in many different game configurations (such as different maps, spawning locations, supply distributions, etc.).

Contact cog2022@inspirai.com if you have further questions about the challenge