Jake Henning

@AL0494241 what version of pip do you have installed? Can you attach the full log?

It does! based on the cuda version inside the container The fix in the latest agent solves the issue of re-downloading torch if the version do not match exactly e.g....

Hi @Guydada , are you sure your clearml.conf is configured correctly? Can you share it (without the credentials, of course)

Yeah, I think this is actually quite a neat solution that also uses the HOCON capabilities (good thing we use it 🙂)

Hi @terbed , the original pip you attached was created during the local run of the task, before running it remotely?