Hi @<1547028031053238272:profile|MassiveGoldfish6> , you would want to use the output_uri parameter of Task.init

@<1523701070390366208:profile|CostlyOstrich36> , wondering if my clarification made sense and if what I am describing is possible?

Hi @<1547028031053238272:profile|MassiveGoldfish6> , I think this is what you're looking for - None

Sorry, I wasnt clear. I already have everything set up correctly in the clearml.conf to work with S3. I know this works because when I push datasets, the data is stored in S3. It's just that every time I run an ML experiment, clearml is using my machine as storage which is fine for experiment tracking and comparison. The only problem is that other members of my team want to use my models but they cant since they cant pull it. How do I tell clearml to use the S3 storage for an experiment instead of my local machine? Or simpler, just push my best model to S3 for anyone on my team to use