Hi @<1523707996645888000:profile|GrievingTurkey78> , did you configure the agent to install requirements from requirements.txt ? You can get more information using the --debug flag when running clearml-agent

I think if you provide an absolute path it should work 🙂

I see. I don't think it's supported but I think it would be a great idea for a feature. Maybe Open a Github feature request?

you're always running a single task at a time. The whole point is that everything is reported to the task (auto-magic bindings, console logs etc.), so there cannot be any ambiguity. You can close the current task ( task.close() ) and init a new one if you'd like, but you can't init several at the same time.

Hi @<1523701083040387072:profile|UnevenDolphin73> , not in the open source

Setting the upload destination correctly and doing the same steps again

That's a good question. In case you're not running in docker mode, the agent machine that runs the experiment needs to have Cuda/Cudnn installed. If you're running in docker mode you need to select a docker that already has those installed 🙂

And also look here 🙂 https://github.com/allegroai/trains-actions-train-model

Hi IrritableJellyfish76 , yes. It is available only for Scale & Enterprise versions

But... if the machine is shut down, how do you expect the agent to run?

Do you see any errors in the apiserver on startup or after?

Because that seems to be connected to data

No problem 🙂
Imagine that internally for an artifact that is saved in some address my-minio-host:9000/FILE
Internally ClearML keeps the link to the artifact as is. It doesn't matter where the ClearML backend is located/deployed since it will always be pointing to the same address. You could hack it around by doing changes on Mongo directly but I would strongly advise against it if you're not sure what you're doing

And the experiments ran on agents or locally (i.e pycharm/terminal/vscode/jupyter/...)

Hi @<1554638160548335616:profile|AverageSealion33> , if you must have some separation between dev and prod it would be a good idea. Can you elaborate on what you mean regarding clearml features?

Hi @<1523715429694967808:profile|ThickCrow29> , thank you for creating a github issue on this! Makes sense it would be moved to aborted 🙂

Hi RotundHedgehog76 , from API perspective I think you are correct

Looks like you're having issues connecting to the server through the SDK. Are you able to access the webUI? Is it a self hosted server?

Hi @<1768084624061239296:profile|QuaintWoodpecker78> , you have an error when you try to unzip? Are you downloading directly through the webUI? Where was the artifact stored?

Hi @<1635813046947418112:profile|FriendlyHedgehong10> , can you please elaborate on the exact steps you took? When you view the model in the UI - can you see the tags you added during the upload?

Hi @<1752139552044093440:profile|UptightPenguin12> , for that you would need to use the API and use the mark_completed call with the force flag on

Hi @<1523701601770934272:profile|GiganticMole91> , As long as experiments are deleted then their associated scalars are deleted as well.

I'd check the ES container for logs. Additionally, you can always beef up the machine with more RAM to give elastic more to work with.

Hi @<1574931891478335488:profile|DizzyButterfly4> , I think if you have a pandas object pd then the usage would be something like ds.set_metadata(metadata=pd, metadata_name="my pandas object")

I think you would be referencing the entire thing using the metadata_name parameter

Hi @<1719524641879363584:profile|ThankfulClams64> , what do you mean regarding ClearML GPU Compute? Do you mean the Genesis autoscaler?

My bad, if you set auto_connect_streams to false, you basically disable the console logging... Please see the documentation:

auto_connect_streams (Union[bool, Mapping[str, bool]]) – Control the automatic logging of stdout and stderr.

Hi @<1523704674534821888:profile|SourLion48> , can you try running it from the same machine that the server is running on?

Hi @<1717350332247314432:profile|WittySeal70> , where are the debug samples stored? Have you recently moved the server?

I was suspecting connectivity issues. Glad to hear it's working

The ubuntu is the client side or you changed OS on the server side?

Try this key pair on another machine, could be just invalid..