Oh, I think that is for a very small data. I don't think it works for me.

I know. And the very fast help 🙏 😀

Can I see visualization of for example categorical columns as bar graphs?

Thank you so much.

Pefect, thanks. I will take a look at that.

It is an autoscaler for gcp. I think, there are unnecessary configs that were used in aws.

@<1523701070390366208:profile|CostlyOstrich36>

clearml==1.14.1
That is the version.

@<1523701070390366208:profile|CostlyOstrich36> I have been exploring. The problem seems to be when the docker container is using the cached dir.

Using cached repository in "/root/.clearml/vcs-cache/****.git.0081a6bc4d7afe6adde369e6aeab9406/****.git"

When inside that directory and tries to fetch, it asks for credentials. when it clones, it doesn't.

cloning: git@github.com:****/****.git
Using user/pass credentials - replacing ssh url 'git@github.com:****/****.git' with https ...

configurations:
  extra_clearml_conf: 'sdk.aws.s3.region="us-west-2"
                        agent.extra_docker_arguments=["--shm-size=90g"]
                        agent.extra_docker_shell_script=["git config --global credential.helper cache --timeout=604800",]'
  extra_trains_conf: ''
  extra_vm_bash_script: ''
  queues:
    gcp-v100:
    - - gcp-v100
      - 4
    gcp-l4:
      - - gcp-l4
        - 4
    gcp-cpu:
    - - gcp-cpu
      - 4
  resource_configurations:
    gcp-v100:
    ...

they are different tasks. I start a new task but it can be same commit sometimes.

Okay. thanks. But, about the overriding, I tried committing and when commited it works. So, I think that means the configuration is not overriden else where.

Another very related question is, does uncommitted changes work for sub modules as well? I mean, when there is a directory from a different repository cloned as a submodule.

both actually.

Sorry, It just worked now. I think It was slow internet connection issue. It just went away today.

CostlyOstrich36

It doesn't work when I insert the credentilas individually either. I am using EC2 as clearml server.

Thanks, it did

deployed it myself. It worked fine before I changed to ubuntu24.04 yesterday. We have been using clearml for years this way

I tried it on another laptop and I also tried it on the server machine where the clearml server is deployed. I get same error. @<1523701070390366208:profile|CostlyOstrich36>

It works on other machines. Can I clearml-init on a virtual environment? I installed clearml in a virtual environment.

I will do that

I see. So, is it same thing when network is slow and there is a mistake in url?

import yaml
from clearml.automation.auto_scaler import AutoScaler, ScalerConfig
from gcp_driver import GCPDriver

with open('gcp_autoscaler.yaml') as f:
    conf = yaml.load(f, Loader=yaml.SafeLoader)

    driver = GCPDriver.from_config(conf)

    conf = ScalerConfig.from_config(conf)
    autoscaler = AutoScaler(conf, driver)
    autoscaler.start()

That is the python code.

Thanks CostlyOstrich36
But, I am not using report_media() function. The debug samples (confusion matrices) are saved from tensorboard.

They are actually from tracked files. Actually, I get the uncommitted changes under Execution tab.

@<1523705004920147968:profile|CloudySwallow27>

But, it uses the committed changes instead of these values.

CloudySwallow27

The devops changed the url and I had to go through some steps to find out what the problem was.