So I'd have to make edits to the docker-compose file for clearml-serving; there would not be any issues arising due to that right?

How do I access the logs of the init process?

So I am able to access it via sending requests to the clearml fileserver but, any way to access it from the dashboard(the main app)?

We're initialising a task to ensure it appears on the experiments page;
Also not doing so gave us issues of ‘Missing parent pipeline task’ for a set of experiments we had done earlier

That aside, adding mock doesn't really change the error above though

SuccessfulKoala55 Any leads f or the above error i n running the script?

Exactly what I thought;
That's so odd

Oh

Tried all three solutions right now;
Oddly enough it doesn't work :(

Yes they have been set up as per the new credentials created on the server

What exact ly does the error mean

Yep, the pipeline finishes but the status is still at running . Do we need to close a logger that we use for scalers or anything?

Hey We figured a temporary solution - by importing the modules and reloading the contents of the artefact by pickle. It still gives us a warning, though training works now. Do send an update if you find a better solution

Hey, is it possible for me to upload a pdf as an artefact?

I replaced local host with the host name of the server in the credentials

Though as per your docs the add_requirements is for a requirements .txt

This issue was due to a wsl proxy problem; wsl’s host name couldn't be resolved by the server and that became a problem for running agents. It works fine on Linux machines so far, however.
So no worries :D

stuff is a package that has my local modules - I've added it to my path by sys.path.insert, though here it isn't able to unpickle

Not local .py files

I'm good, David; hope you're good too :D
Thanks for the help! It worked :D

Also, does PipelineDecorator.upload_model store anything on the fileserver ? I can't seem to understand the use of PipelineDecorator.upload_model() apart from making a model appear on the pipeline task

Also, if I were to move to g pu machines, how do I assign more cores to a worker or so?

However, I use this to create an instance of a dataloader(torch) this is fed into my next stage in the pipeline - though I import the local modules and add the folders to the path it is unable to unpickle the artifact

I'm asking this because my kwargs is observed as an empty dict if printed

Yep, that's exactly what's happening.
However here's what I want to do:
upload model to clearml’s fileserver and get the model url in the details for easy download

2022-07-12 13:41:39,309 - clearml.Task - ERROR - Action failed <400/12: tasks.create/v1.0 (Validation error (error for field 'name'. field is required!))> (name=custom pipeline logic, system_tags=['development'], type=controller, comment=Auto-generated at 2022-07-12 08:11:38 UTC by mbz1kor@BMH1125053, project=2ecfc7efcda448a6b6e7de61c8553ba1, input={'view': {}})

Traceback (most recent call last):

File "main.py", line 189, in <module>

executing_pipeline( 

File "/home/mbz1kor...

Nop e

Thanks for actively replying, David
Any update on the example for saving a model from within a pipeline( specifically in .pth or h5 formats?)

Can PipelineDecorator.upload_model be used to store models on the clearml fileserver?