Does it support distributed agents?

[DazeHolic]

I wan to train a model with multy agents for speed up trainning. I know it support multi-agent in one scene. But my environment takes too much time to run and cost 60% CPU. So get more agents in one PC can't solve my problem. I want to train agents in several PCs.

How can I solve this problem?

Thanks!

[vincentpierre]

Hi, Distributed training is on our roadmap and we are currently working on it. It was announced in this blog post in the section ML-Agents Cloud. We do not yet have a release date...

[DazeHolic]

hey, vincentpierre. I'll always be waiting for this. Thank you for your efforts.

[Hsgngr]

We are so excited about Distributed RL, please inform us a soon as you start to work on it.

[ghost]

@vincentpierre I'm working on a distributed training and started to explore more into the ML agents python API; plan on this is to run AWS ECS cluster, and I wonder if you would be able to give us some directions on the best approach to handle this one? I appreciate the support. I am looking forward to getting some feedback.

[github-actions[bot]]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.