Closed CHENGEZ closed 1 year ago
I'm also puzzled by this. I believe it depends on the type of group access that each user has on the server. For some people, the conda environment works well, for some other people, it does not.
A practical solution is to use a virtual environment instead of a conda environment. Please see this.
Be aware of disk space limitations, as installing more environments requires more space, you might need to uninstall some of the environments that you have already created, like the simpleEnv
.
HI, I have the same error message but my jobs already started to train on the cluster. I don't really get it why it works with this error message .....
I had the same issue but adding source ~/.bashrc
in a line before conda activate pylocoEnv
to the job scripts solved it
I had the same issue but adding
source ~/.bashrc
in a line beforeconda activate pylocoEnv
to the job scripts solved it
Thank you very much! It does solves it.
Hi,
I have installed conda on the euler server, and it works fine as I can create conda envs and activate them normally.
Below shows the created conda environments:
But when I follow the instructions in
HOWTO.md
to submit a job, I always get the following output in the.err
file:It seems the conda command is not found when called from the job script.
And of course since the environment is not successfully activated, "Module not found" issues also appear in the
.out
file. Could you please provide some hint on why conda works from the terminal but not from the job script?