@<1523701070390366208:profile|CostlyOstrich36> i just edited my question, can you refer to it pls.

great, thank you!

@<1822805241150574592:profile|ShinySparrow39> thanks, but i don't understand your workaround. how does it enables overrides from clearm UI? specifically for hyperparameters app which doesn't seems to work with logged configurations files.

we are using the community server (pro account). full configuration is attached.

update:
--docker-args "-v some_dir:other_dir -v some_dir:other_dir" is the correct format

I'm currently overcoming it by just adding a "_ " ("data_ something" -> "_data_something")

setting ignore_remote_overrides = True help solve the issue, but obviously we can't use it as a solution. what reasons might be that it would take so much time when trying to find params override in the backend? is it a network issue? maybe needs to change the machine network configuration?

Thanks!

It worked, thanks! i spent a few hours trying to figure it out 😅

thanks!
by application you mean the web UI? is it possible to estimate the amount of calls it's performing when idle? while viewing an experiment? or some other typical usage?

after playing with it a bit more, i figured out that you need to set _ allow omegaconf_edit_ to False and the parameters naming should be done with dot notation. for example Hydra/some_key. other_key.my _key . there is no documentation regarding hydra with hyperparameters tuning, it is best to add a section regarding it in the link you attached.

hi, thanks for the reply, i can access the web UI, I am using the Pro plan (clearm's host)

well there is no default one.. and in the docs there is nothing about it. would be nice to add a minimal requirements for the AMI in the docs, instead of just writing "The AWS AMI to launch".
thanks for the answer :)

Hi, sure, there is nothing special there, even some redundancy.

def initialize_clearml_task(
        project_name: str = None,
        task_name: str = None,
        task_type: str = None,
        tags: list[str] = None,
) -> tuple[Task, Logger]:
    """
    Initialize and configure a ClearML task.

    Parameters
    ----------
    project_name : str
        Name of the ClearML project.
    task_name : str
        Name of the ClearML task.
    task_type : str
        Type of the ClearM...

It seems that when the working directory is set to '.' (root of the cloned repo) I am able to import my package as expected.
I thought about your solution but then it requires me to push every time I change my package which is inconvenient.

conf file:

# ClearML SDK configuration file
api { 
    # Tomer Roditi's workspace
    web_server: 

    api_server: 

    files_server: 

    # corractions server
    credentials {"access_key": "****", "secret_key": "****"}
}
sdk {
    # ClearML - default SDK configuration

    storage {
        cache {
            # Defaults to system temp folder / cache
            default_base_dir: "~/.clearml/cache"
            # default_cache_manager_size: 100
  ...

@<1523701070390366208:profile|CostlyOstrich36> thanks for the reply!

yes i'm using app.clear.ml
the vm is initialized via the clearml autoscalers, in the aws autoscaler i didn't have to do some network configurations there, thus i assume that it should be the same in the gcp VMs.

can you direct me to tests that should reveal lagging issues?

yes @<1523701087100473344:profile|SuccessfulKoala55> , i'm using the pro plan, "Hyperparameter Optimization" app

from the consule logs i can see that:
entry_point = preprocessing.py
working_dir = Scripts/Volvo/volvo_data_alignment

installed packages before the task is running:

# Python 3.10.10 | packaged by Anaconda, Inc. | (main, Mar 21 2023, 18:39:17) [MSC v.1916 64 bit (AMD64)]

GitPython == 3.1.31
bokeh == 2.4.3
boto3 == 1.26.158
botocore == 1.29.158
clearml == 1.11.1
h5py == 3.8.0
holoviews == 1.16.0
joblib == 1.2.0
lightgbm == 3.3.5
mat73 == 0.60
matplotlib == 3.7.1
numpy == 1.23.5
pandas == 1.5.3
pytest == 7.3.1
scikit_learn == 1.3.0
scipy == 1.10.1
shapely == 2.0.1
sktime == 0.21.0
statsmodels == 0.13.5
t...

so how can i make clearml to install my req file? i thought that it is automatically detects the req file in the cloned repository..

it seems like it does finds the requirments file otherwise why would it install tsfel tsfresh and some other "non-basic" packages?

my repo structure is:
|----my_package
|----scripts
|----some_folder
|----some_folder
|----task_script.py
|----requirements.txt

@<1523701070390366208:profile|CostlyOstrich36> Hi, i would expect a feature that looks something like this:
clearml Task CLI option "--mount-files" (or other informative name) which would be used to add local files to the sent task in the following format:

clearml-task --project examples --name remote_test --script my_script.py --mount-files "local_file_1:target_path_1, local_file_2:target_path_2"

of course there would be some size limit to the mounted files (same as you do with l...

@<1822805241150574592:profile|ShinySparrow39> thanks for replying but i think you got me wrong. in your suggestion you can connect only files that are committed and pushed to git when running remotely, which is exactly what i'm trying to find a workaround to.

any idea what it might be? or how can i test it with them?

After running your suggestion (which worked), i'll be more accurate, the issue is with datasets and not tasks. try running this:

from clearml import Task, Dataset
import pandas as pd

task = Task.init(project_name='examples', task_name='Artifacts with data_')
ds = Dataset.create(dataset_project="examples", dataset_name="test", use_current_task=True)

df = pd.DataFrame(
    {
        'num_legs': [2, 4, 8, 0],
        'num_wings': [2, 0, 0, 0],
        'num_specimen_seen': [10, 2, 1, 8]
 ...

since it is specifying the pandas==1.5.3 i thought it might be pulling the req file from my main branch so i updated the req file there as well (since i am working on a different branch) and it didn't changedanything (as you might expect)

we would love to get some clarifications as well, how does strings should be provided? what is the step size param in the uniform type? is it normal that all unused hyperparams in the HPO are displayed as str typein the UI?
thanks!

yes the agent has already cloned the same repo in the first task (from the same account with the same user and token).
do you mean the full log of the machine itself? the full log of the failed task is attached already