From my understanding the agent basically runs a pip install command (usually pointing to some autogenerated list of packages - depending on configs).

You can treat the 'installed packages' section as if it were a 'requirements.txt' file.

Hi @<1717350332247314432:profile|WittySeal70> , for pipelines you should use pipe.add_parameter . Under the hood the controller is a special type of Task. ClearML always automatically detects argparser, however for pipelines this methodology does not apply and requires pipe.add_paramter

Hi @<1726047624538099712:profile|WorriedSwan6> , ClearML uses ElasticSearch & MongoDB as databases for all of that information. I suggest checking online for backup procedures of these databases in K8s

Oh LOL 😛

Hi MagnificentMosquito84 , is this a self hosted server? What version is it? Do you have visibility into the logs?

RoughTiger69 , do you have a rough estimate on the size that breaks it?

RotundSquirrel78 , try going into localhost:8080/login

Hi @<1547028031053238272:profile|MassiveGoldfish6> , are you self hosted or on the community server? What project is this, a pipelines/dataset project or just some regular project?

Can you add the full error please?

Hi!
Can you say what's the size of your clearml folders?

Hi EnviousPanda91 , what version of ClearML are you using? Are you running on a self hosted server?

If you're running on a remote interpreter you can install clearml plugin so it will detect repo properly

Hello MotionlessCoral18 ,

Can you please add a log with the failure?

Is it possible that you don't have permissions for deletion on that Azure account with your credentials?

Try with delete_files=False parameter

Also with delete_external_files=False

Try running the following script

from clearml import Task 
import time

task = Task.init(output_uri="

")

print("start sleep")
time.sleep(20)
print("end sleep")

Please add the logs

With what host config were you trying the last attempts?

Hi @<1547752799075307520:profile|ZippyCamel28> , to address your points

• What do you mean by 'reload'?
• You need to go into the project and archive the experiments in order to delete the project + experiments in the archive
• There are some configurations you can play with to report 'less' metrics. For example sdk.metrics.plot_max_num_digits You should read here - None . To get an idea of the size of an experiment think of an...

What is the combination of --storage and configuration that worked in the end?

try None :port/bucket

Non-AWS S3-like services (e.g. MinIO):

:port/bucket

I think it is one of the parameters of the task. Fetch a Task and see what properties the artifact has 🙂

Hi @<1665891247245496320:profile|TimelyOtter30> , not sure I follow. It looks like a misconfiguration. I think you need to see the correct settings here: None , also note the direct reference to minio 🙂

Can you give an example of how you're saving this artifact?

Hi @<1523702932069945344:profile|CheerfulGorilla72> , making sure I understand - You basically want to select an input model via the UI?

Hi JitteryCoyote63 , you can get around it using the auto_connect_frameworks parameter in Task.init()

I see, thanks for the input!

Or are you trying to change something in the docker compose?

Hi @<1534496192929468416:profile|EagerGiraffe33> , what if you try to put a specific version of pytorch you've tested on your remote environment in the requirements section of the cloned task?