I indeed have stopped autoscalers that ran for a long time. Deleting them initially seems to work, but they immediately return as if never deleted.

I've already deleted many experiments and models, but this had absolutely no impact on the Metrics storage.

I have the email associated with the account. How should I share it privately?

Worse yet, I've tried both a conda-based environment and a virtualenv with pip for everything, and both fail in this way.

Boohoo! I was planning on using it heavily. 😞

@<1523701070390366208:profile|CostlyOstrich36> this is the workspace id: 0f88f2f4baff40e5be0f88acd0180447

I'm seeing an extremely short list, that is indeed missing scipy and scikit-learn, which are clearly installed and available.

The vanity name, your-name-here.clear.ml, fails, but app.clear.ml works. Thanks.

Thanks for the help, @<1523703097560403968:profile|CumbersomeCormorant74> . I've successfully deleted these hidden tasks (all but two that ran for a few minutes each). A day has passed, and the Metrics usage hasn't changed at all.

I have just created a clean python virtual env (which seems to work fine locally, but fails as above). Naturally, it is using clearml v1.9.1 .