ClearML FAQ | Hi, I'M Trying To Run My Code With A Pipelinecontroller Within A Docker Container Instead Of On My Local Computer. Currently Having Trouble Where The Code Doesn'T Run In The Repo As Expected; "No Repository Found, Storing Script Code Instead" Is The Warni

Answered

Hi, I'M Trying To Run My Code With A Pipelinecontroller Within A Docker Container Instead Of On My Local Computer. Currently Having Trouble Where The Code Doesn'T Run In The Repo As Expected; "No Repository Found, Storing Script Code Instead" Is The Warni

Hi, I'm trying to run my code with a PipelineController within a docker container instead of on my local computer. Currently having trouble where the code doesn't run in the repo as expected; "no repository found, storing script code instead" is the warning. Using set_repo to the github also presents a problem where I'm running a script in a folder in the repo, but the execution work directory is the root and not the intended folder. Any suggestions as to how to fix this? Is this a problem with the clearml.conf file? Or is there a way to change the execution work directory for the pipeline controller?

  				
Posted 
	10 months ago

					More
				  		
  Report
		
					BitingDeer35
				
					0
					 × 1

Votes Newest

Answers

Hi @<1633638724258500608:profile|BitingDeer35> ! Looks like the SDK doesn't currently allow to create steps/controllers with a designated cwd. You will need to call the set_script function on your step's tasks and on the controller for now.
For the controller: If you are using the PipelineDecorator, you can do something like: PipelineDecorator._singleton._task.set_script(working_dir="something") , before you are running the pipeline function. In the case of regular PipelineController s , you need to do something like controller._task.set_script("...") .
For steps: you need to iterate through each node and change the cwd before the nodes run:

for node in controller.nodes:
   node.job.task.set_script("...")

  				
Posted 
	10 months ago

					More
				  		
  Report
		
					SmugDolphin23
				
					0

Write your answer

927 Views

1 Answer

10 months ago

10 months ago

Tags

Similar posts

Have You Guys Forgot To Add Uv Caching? Im Trying On 1.9.3Rc0 And Uv Always Redownloads The Packages, Doesnt Use Cache. Or Is It My Skill Issue? All I Added: Package_Manager: {

Hi, I Am Logging Some Scalars From A Simple Ml Experiment, And Im Seing Different Curves Logged With Clearml Depending On When I Run The Code. At The Moment We Are Basically Drawing Some Random Numbers And Logging Their Values, But The Frequency Between E

Hi! Maybe Someone Will Be Able To Help Me. Im Trying To Build System: Step 0: Build Pipeline (Run Locally) Step 1: Executes The Taskscheduler, Step 2: Executes The Pipeline (Remotely) I Tried To Use Agent Services As On Image. But It Doesnt Work ( I’M Ab

Hello, What Are Mongodb Default Login Credentials For Self Hosted Clearml Server? We Are Moving Our Datasets To Another Place And So Far I Think Clearml Doesnt Support It. So I Would Like To Try Change The Database Entry (Im Guessing Its Stored Somewhere)

Hello Everone, I Have Hosted Clearml Server And Trained A Yolov8 Model To Test My Installations. The Model Was Trained Successfully And I Tried To Optimize The Hyderparameters By Using The Sample Code From Clearml But Im Getting Some Error In Doing So An

Hi, I Am Using Clearml By Building It As My Own Server. After The Message Below Was Displayed, The Operation Stopped Without Progress. In Clearml Server, It Is In “Running” State. “Clearml.Task - Info - No Repository Found, Storing Script Code Instead”

Hi All, Is There A Way To Make Agents That Run As Root, To Execute The Code Tasks And Of Pipeline Steps, As The User That Executed The Pipeline Or Task Code Instead Of As Root?

Hi, I Was Uploading An Image Artifact Using The Following But In The Preview I Only Get An Array Instead Of An Image. Am I Doing Something Wrong? ``` Im=Cv2.Imread('Pic.Jpg') Task.Upload_Artifact('Myimage',I'M) ```

Hello, I'M Really Enjoying The Clearml Experience And We'Re Using It Very Successfully At Work. I Have A Small Question Though: I'M Trying To Set The "File_History_Size" Sdk Parameter From Python Code Instead Of The Conf File. Is That Possible? Thanks!