An inquiry when using multiple compute nodes

[qhtjrmin]

Hi, I am interested in training using your BNS-GCN in a distributed environment. However, I have encountered some difficulties running the setup with just the reddit_multi_node.sh script.
Could I get some guidance on how to properly configure the environment for distributed experiments?
Thanks for your help.

[ch-wan]

Hi, thank you for your interest to BNS-GCN. There are many issues that can cause the failure of launching multi-node training. Could you please share me with the error message? Simply launching reddit_multi_node.sh should not work as you need to specify some arguments for distributed training by following this.

If you are asking how to setup a distributed training environment, you have find the official PyTorch tutorial helpful.

Thanks.

[ch-wan]

@qhtjrmin I'm going to close this issue due to its inactivity. Please feel free to reopen it if the above response is not sufficient. Thanks.