Hello everyone
I’m new to ClearML and I’ve faced one issue with models auto logging. Tried to google an answer online, but failed..
Generally I use the most simple https://github.com/allegroai/clearml/blob/master/examples/frameworks/pytorch/pytorch_tensorboard.py example. But when I save the model - the only thing that I can get - it’s local path, not saved weights on the server. Meanwhile there is no any errors from the ClearML logs.
On the other hand if I store some regular artifact(image, for example) - everything works as expected.
Do you know, is this feature supported only in case of remote cloud storage connected to a Clear ML Server?
Tech specs:
We’re using self hosted ClearML via Kuber WebApp: 1.6.0-213 • Server: 1.6.0-213 • API: 2.20 pypi clearml==1.6.0 (the same issue on 1.6.2 as well)
Update(solution):
If you meet the same issue - just add additional argument output_uri=True during Task.init() call (see details in the thread)
Hi GreasyRaven35
You should set the output_uri, in Task init, it will auto upload the model, and register the remote location URLtask = Task.init(..., output_uri=True)You can also specify a target bucket, if you configured credentials (e.g. output_uri=" s3://bucket ")
Thanks a lot for your help, everything works like a charm!
