I think these are the configurations you are looking for:
sdk.development.worker.report_event_flush_threshold
sdk.development.worker.console_cr_flush_period
inside clearml.conf
I would suggest the website ๐
Hi ReassuredOwl55 , can you please elaborate on your use case or exactly what you're trying to achieve?
Hi SarcasticSquirrel56 ,
You can use output_uri=<S3_BUCKET>
in Task.init()
to upload artifacts to s3. Is that what you're looking for?
UnevenDolphin73 , I've encountered a similar issue with s3. I believe it's going to be fixed in the next release ๐
Hi WhoppingMole85 , can you please provide a small code snippet to play with? How are you saving the tables?
Hi @<1607909176359522304:profile|UnevenCow76> , I suggest you review the following video on serving - None
This also explains how to visualize different metrics in Grafana
I think AnxiousSeal95 updates us when there is a new version or release ๐
Looping in @<1523703436166565888:profile|DeterminedCrab71> & @<1523701435869433856:profile|SmugDolphin23> for visibility
Hi @<1523701132025663488:profile|SlimyElephant79> , to answer your questions:
-
Does ClearML store any dataset outside of this S3 storage (and local storage) for preview or compression purposes?Some preview data might be stored inside mongodb (if it's a table for example). This of course can be disabled so no data/previews are exposed to the ClearML backend.
-
Are any data packets related to model or data versioning routed through the API server? (We are pretty confident it shouldn't, but j...
@<1691620877822595072:profile|FlutteringMouse14> , can you try upgrading to the latest available? 1.8.1rc2
Hi AttractiveShrimp45 , can you please elaborate on what you mean by KPIs artifact?
Hi @<1570583227918192640:profile|FloppySwallow46> , please don't @ the entire channel for help ๐
If a task is in pending it means that no agent picked it up yet. Maybe the agent is unavailable or the process crashed. Check in that direction
You can, but if it's a large organization having a dedicated account for this would probably be best
What are the Elasticsearch, Mongo and apiserver versions in the docker compose are? Backup will only work in this scenario when they are exactly the same between 2 systems.
Is it the services docker that comes with the docker compose or did you run your own agent?
Hi @<1655381985293504512:profile|LividGorilla28> , this is because your browser cannot show local files as it is a sandbox environment. Opening a new tab and going directly to a local file by pasting the path is the only way to view images that are saved locally.
The correct way is to store it on some storage solution such as s3/gcp/minio which can serve the data to the browser
Hi @<1594863230964994048:profile|DangerousBee35> , this is pretty much it. I think the default one suggested is a good one
Hi @<1695969549783928832:profile|ObedientTurkey46> , do you have a code snippet that reproduces this behaviour?
My guess other agents are sitting on different machines, did you verify that the credentials are the same between the different clearml.conf
files? Maybe @<1523701087100473344:profile|SuccessfulKoala55> might have an idea
YummyLion54 , Hi
What versions of clearml/clearml-agent are you using?
When running it without the agent, do the hydra configurations show up properly in the UI?
You can provide it in the extra configurations sections