Running a distributed program cannot find the correct library

Questions
#1

Hello, I am running the program under /dgl/example/pytorch/graphsage/dist. According to the instructions of the readme, I have run to the last step, but during the running process, the error is showed that the numpy cannot be found, but it is obvious I downloaded them and I suspect it’s an environment setup issue, but I’m not sure what went wrong.
My environment is:
Ubuntu-18.04
miniconda3
pytorch1.12
dgl0.9.1

Here is my error message:

%Z$W}}QJ9UMC)`L_PZMCD
%Z$W}}QJ9UMC)`L_PZMCD1920×724 234 KB

#2

is numpy installed on all the machines’ environment you’re training on? not only the machine you triggered the train.

#3

Yes, both machines are the same configuration

#4

To add, both of my machines can run simple GCN programs individually, so I don’t think there might be a problem with the running environment.

#5

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.