First of all, to handle SSL termination and secondly, to use subdomains instead of exposed external ports

Hi @<1554638179657584640:profile|GlamorousParrot83> , how did you set up your clearml.conf file? Can you share?

Just an update - it seems this is actually a caching issue that we're working to address... You're getting an outdated page ( https://allegro.ai/docs/examples/examples_hyperparam_opt/ ) while the new docs page is https://allegro.ai/docs/examples/optimization/hyper-parameter-optimization/examples_hyperparam_opt/

And does this always reproduce?

Hi TrickySheep9 ,
Not sure I understand what you mean...

I'll ask the UI team to take a look - any specific way to reproduce it?

This is something that needs to be addressed on the machine level. Deleting experiments tries modifying metrics/logs etc. but these are read-only at this stage...

Wait for it... 🙂

And here's for data stored in ES: https://clearml.slack.com/archives/CTK20V944/p1631037530225100?thread_ts=1631031994.223100&cid=CTK20V944

Hi RattyDolphin75 , which ClearML SDK version are you using?

You need only 5 - api, web and files host, key and secret

that's strange

Hi @<1549202366266347520:profile|GorgeousMonkey78> , we're investigating

Hi MortifiedCrow63 , I think custom SSO login is supported in the ClearML paid version

Hi @<1523701092473376768:profile|SuperiorPanda77> , while data in ES is deleted, compacting the indices is not something ES does immediately - you can look in the EE documentation in how to trigger that yourself (can't look for it now)

Hi @<1724235695645528064:profile|RattySeahorse41> , these messages basically means you're using the Self-service Free account in app.clear.ml, and your usage has reached the free usage limit for this month. One option is to upgrade to a Pro account.

When a task is taken from a queue, only a single agent will get it. If you use a pipeline, each agent can run a different step

Hi SkinnyPanda43 , what exactly do you mean by config values? The configuration objects contain configuration your code has uploaded during the run, but other config values (including hyper parameters etc.) are also stored in different places

There was a fix related to this feature on the latest agent version, you'll need to use 1.1.0

@<1784754456546512896:profile|ConfusedSealion46> can you share the logs you're getting on the task?

Hi FloppyDeer99 ,

after pulling it from a queue in agent

What do you mean exactly?

DilapidatedDucks58 when if you try to create a new project?

Hi DilapidatedDucks58 , that's very strange - can you attach a screenshot? I just tried that both in the demo server and our community server, and I can add description to new empty sub projects or edit the description for existing projects...

@<1710827340621156352:profile|HungryFrog27> (3) in this case means the agent detected that the task's state was changed from running to something else (the agent checks that 3 times to make sure, waiting 3 seconds between each check)

Seems good enough for me, as long as all that is well documented 🙂

The question is who is creating this queue 🙂 - just making sure again, the queue with the strange name is created after you enqueued a task to some existing queue using the UI

What server version are you using?

Hi GiddyTurkey39 , are we talking about two different issues? One is the Trains Server on GCP, the other is installing trains on your machine?

Where do you use tqdm?

Well, let me check 🙂