ClearML/clearml-docs
1
0
Fork 0
mirror of https://github.com/clearml/clearml-docs synced 2025-06-26 10:06:56 +00:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

pipeline remote execution

Browse Source
This commit is contained in:
revital 2023-08-21 13:25:24 +03:00
parent f90acd9db1
commit 789f07eb29
2 changed files with 18 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

BIN
docs/img/pipelines_edit.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 133 KiB

18
docs/pipelines/pipelines_sdk_tasks.md
View File

@ -250,3 +250,21 @@ Run the pipeline by using one of the following methods:
respective queues or in the default execution queue. respective queues or in the default execution queue.
* [`PipelineController.start_locally`](../references/sdk/automation_controller_pipelinecontroller.md#start_locally) - launches * [`PipelineController.start_locally`](../references/sdk/automation_controller_pipelinecontroller.md#start_locally) - launches
the pipeline controller locally. To run the pipeline steps locally as well, pass `run_pipeline_steps_locally=True`. the pipeline controller locally. To run the pipeline steps locally as well, pass `run_pipeline_steps_locally=True`.
### Remote Execution
When you run your pipeline, ClearML collects all the information required to reproduce it (DAG, configuration,
installed packages, uncommitted changes etc.). By default, when you rerun the pipeline through the ClearML WebApp, the
pipeline is constructed from the original codebase.
To change this behavior, pass `always_create_from_code=False` when instantiating a `PipelineController`. When rerun,
the pipeline is generated from the pipeline configuration stored in the pipeline task. This allows you to modify the
pipeline configuration via the UI, without changing the original codebase.
To change the pipeline configuration via the UI:
1. Clone the pipeline controller task
1. Go to the cloned task’s **Configuration > Configuration Objects > Pipeline**
1. Hover over the configuration object > click **Edit**
When the pipeline is rerun, it will use the new configuration values, overriding the original values set in the code.
![Edit pipeline via UI](../img/pipelines_edit.png)