mirror of
https://github.com/clearml/clearml-docs
synced 2025-01-31 06:27:22 +00:00
55 lines
2.2 KiB
Markdown
55 lines
2.2 KiB
Markdown
---
|
|
title: Reproducing Experiments
|
|
---
|
|
|
|
Reproduce experiments on local or remote machines in one of the following ways:
|
|
* Cloning any experiment - Make an exact copy, while maintaining the original experiment
|
|
* Resetting an experiment whose status is not *Published* - Delete the previous run's logs and output
|
|
|
|
After cloning or resetting, enqueue the reset or newly cloned experiment for execution by a worker.
|
|
|
|
Experiments can also be modified and then executed remotely, see [Tuning Experiments](webapp_exp_tuning.md).
|
|
|
|
## Cloning
|
|
To clone an experiment:
|
|
1. In the experiments table, right-click the experiment to reproduce and click **Clone**.
|
|
1. In the `Clone Experiment` modal, set the following:
|
|
* Project - The project where the experiment will be saved
|
|
* Description (optional)
|
|
* Set `<cloned_experiment>` as parent - Select to set this experiment as the new experiments' parent task
|
|
1. Click **Clone**
|
|
|
|
:::note
|
|
By default, the new experiment's parent task is set to the original task's parent, unless the original task does not
|
|
have a parent, in which case the original task is set as the parent. Select `Set <cloned_experiment> as parent` to force
|
|
the original experiment to become the clone's parent.
|
|
:::
|
|
|
|
|
|
|
|
|
|
## Resetting
|
|
|
|
To reset an experiment:
|
|
1. In the experiments table, right-click the relevant experiment and click **Reset**.
|
|
1. In the `Reset Experiment` modal, if you want the experiment's artifacts and debug samples to be deleted from the
|
|
ClearML file server, click the checkbox
|
|
1. Click **Reset**
|
|
|
|
|
|
|
|
## Final Steps
|
|
|
|
At the end of the process you are left with an experiment with draft status, meaning that it is editable.
|
|
|
|
Re-execute the new experiment:
|
|
1. If desired, modify the experiment's configuration (see [Tuning Experiments](webapp_exp_tuning.md)).
|
|
1. Enqueue the experiment for execution. Right-click the experiment > Enqueue > Select a queue > **ENQUEUE**.
|
|
|
|
:::note
|
|
Make sure that a [ClearML Agent](../clearml_agent.md) has been assigned to the selected queue
|
|
:::
|
|
|
|
A ClearML Agent will fetch the experiment from the queue and execute it. The experiment can now be tracked and its
|
|
results visualized.
|