Distributed training using DeepMD.

[jiangzihanict]

I found that there is content related to distributed training in the source code. It looks like the PS architecture is used.

So, can I use DeepMD for distributed training now? Are there any related tutorials and examples?

Thanks!