ClearML/clearml-docs
1
0
Fork 0
mirror of https://github.com/clearml/clearml-docs synced 2025-03-19 19:49:03 +00:00
Code Issues Actions 1 Packages Projects Releases Wiki Activity

Merge branch 'main' of https://github.com/allegroai/clearml-docs

Browse Source
This commit is contained in:
revital 2025-01-19 07:29:17 +02:00
commit ff90deaa23
6 changed files with 82 additions and 73 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
docs/fundamentals/task.md
View File

@ -53,7 +53,7 @@ The captured [code execution information](../webapp/webapp_exp_track_visual.md#e
* Python environment
* [Execution configuration](#execution-configuration) and hyperparameters
The captured [execution output](../webapp/webapp_exp_track_visual.md#experiment-results) includes:
The captured [execution output](../webapp/webapp_exp_track_visual.md#task-results) includes:
* [Console output](../webapp/webapp_exp_track_visual.md#console)
* [Scalars](../webapp/webapp_exp_track_visual.md#scalars)
* [Plots](../webapp/webapp_exp_track_visual.md#plots)

BIN
docs/img/webapp_tracking_33.png
View File

Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 141 KiB

After

Width:  |  Height:  |  Size: 49 KiB

BIN
docs/img/webapp_tracking_34.png
View File

Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 141 KiB

After

Width:  |  Height:  |  Size: 68 KiB

BIN
docs/img/webapp_tracking_34b.png Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 69 KiB

2
docs/webapp/webapp_exp_table.md
View File

@ -233,7 +233,7 @@ to open the context menu
| Abort All Children | Manually terminate all *Running* tasks which have this task as a parent | *Running* or *Aborted* | None for parent task, *Aborted* for child tasks |
| Retry | Enqueue a failed task in order to rerun it. Make sure you have resolved the external problem which previously prevented the task’s completion. | *Failed* | *Pending* |
| Publish | Publish a task to prevent changes to its tracking data, inputs, and outputs. Published tasks and their models are read-only. *Published* tasks cannot be enqueued, but they can be cloned, and their clones can be edited, tuned, and enqueued. | *Completed*, *Aborted*, or *Failed*. | *Published* |
| Add Tag | Tag tasks with color-coded labels to assist you in organizing your work. See [tagging tasks](webapp_exp_track_visual.md#tagging-experiments). | Any state | None |
| Add Tag | Tag tasks with color-coded labels to assist you in organizing your work. See [tagging tasks](webapp_exp_track_visual.md#tagging-tasks). | Any state | None |
| Clone | Make an exact, editable copy of a task (for example, to reproduce a task, but keep the original). | *Draft* | Newly cloned task is *Draft* |
| Move to Project | Move a task to another project. | Any state | None |
| Compare | Compare selected tasks (see [Comparing Tasks](webapp_exp_comparing.md)) | Any state | None |

151
docs/webapp/webapp_exp_track_visual.md
View File

@ -1,14 +1,14 @@
---
title: Tracking Experiments and Visualizing Results
title: Tracking Tasks and Visualizing Results
---
While an experiment is running, and any time after it finishes, track it and visualize the results in the ClearML Web UI,
While a task is running, and any time after it finishes, track it and visualize the results in the ClearML Web UI,
including:
* [Execution details](#execution) - Code, the base Docker image used for [ClearML Agent](../clearml_agent.md), output destination for artifacts, and the logging level.
* [Execution details](#execution) - Code, the container image used for [ClearML Agent](../clearml_agent.md), output destination for artifacts, and the logging level.
* [Configuration](#configuration) - Hyperparameters, user properties, and configuration objects.
* [Artifacts](#artifacts) - Input model, output model, model snapshot locations, other artifacts.
* [Info](#info) - Extended experiment information, such as the experiment start, create, and last update times and dates, user creating the experiment, and its description.
* [Info](#info) - Extended task information, such as the start, create, and last update times and dates, user creating the task, and its description.
* [Console](#console) - stdout, stderr, output to the console from libraries, and ClearML explicit reporting.
* [Scalars](#scalars) - Metric plots.
* [Plots](#plots) - Other plots and data, for example: Matplotlib, Plotly, and ClearML explicit reporting.
@ -16,39 +16,38 @@ including:
## Viewing Modes
The ClearML Web UI provides two viewing modes for experiment details:
The ClearML Web UI provides two viewing modes for task details:
* The info panel
* [Info panel](#info-panel)
* [Full screen details mode](#full-screen-details-view)
* Full screen details mode.
Both modes contain all experiment details. When either view is open, switch to the other mode by clicking <img src="/docs/latest/icons/ico-info-min.svg" alt="Table/Full screen view" className="icon size-md space-sm" />
(**View in experiments table / full screen**), or clicking <img src="/docs/latest/icons/ico-bars-menu.svg" alt="Bars menu" className="icon size-md space-sm" /> (**menu**) > **View in experiments
Both modes contain all task details. When either view is open, switch to the other mode by clicking <img src="/docs/latest/icons/ico-info-min.svg" alt="Table/Full screen view" className="icon size-md space-sm" />
(**View in task table / full screen**), or clicking <img src="/docs/latest/icons/ico-bars-menu.svg" alt="Bars menu" className="icon size-md space-sm" /> (**menu**) > **View in tasks
table / full screen**.
### Info Panel
The info panel keeps the experiment table in view so that [experiment actions](webapp_exp_table.md#task-actions)
The info panel keeps the task table in view so that [task actions](webapp_exp_table.md#task-actions)
can be performed from the table (as well as the menu in the info panel).
![Info panel](../img/webapp_tracking_40.png)
Click <img src="/docs/latest/icons/ico-compact-view.svg" alt="Compressed view" className="icon size-md space-sm" /> to
hide details in the experiment table, so only the experiment names and statuses are displayed
hide details in the task table, so only the task names and statuses are displayed
![Compressed info panel](../img/webapp_tracking_41.png)
### Full Screen Details View
The full screen details view allows for easier viewing and working with experiment tracking and results. The experiments
table is not visible when the full screen details view is open. Perform experiment actions from the menu.
The full screen details view allows for easier viewing and working with task tracking and results. The task
table is not visible when the full screen details view is open. Perform task actions from the menu.
![Full screen view](../img/webapp_tracking_33.png)
## Execution
An experiment's **EXECUTION** tab of lists the following:
A task's **EXECUTION** tab of lists the following:
* Source code
* Uncommitted changes
* Installed Python packages
@ -59,8 +58,8 @@ In full-screen mode, the source code and output details are grouped in the **DET
### Source Code
The Source Code section of an experiment's **EXECUTION** tab includes:
* The experiment's repository
The Source Code section of a task's **EXECUTION** tab includes:
* The task's repository
* Commit ID
* Script path
* Working directory
@ -70,7 +69,7 @@ The Source Code section of an experiment's **EXECUTION** tab includes:
### Uncommitted Changes
ClearML displays the git diff of the experiment in the Uncommitted Changes section.
ClearML displays the git diff of the task in the Uncommitted Changes section.
![Uncommitted changes section](../img/webapp_exp_uncommitted_changes.png)
@ -88,12 +87,12 @@ using to set up an environment (`pip` or `conda`) are available. Select which re
### Container
The Container section list the following information:
* Image - a pre-configured Docker that ClearML Agent will use to remotely execute this experiment (see [Building Docker containers](../clearml_agent/clearml_agent_docker.md))
* Arguments - add Docker arguments
* Setup shell script - a bash script to be executed inside the Docker before setting up the experiment's environment
* Image - a pre-configured container that ClearML Agent will use to remotely execute this task (see [Building Docker containers](../clearml_agent/clearml_agent_docker.md))
* Arguments - add container arguments
* Setup shell script - a bash script to be executed inside the container before setting up the task's environment
:::important
To [rerun](webapp_exp_tuning.md) an experiment through the UI in the listed container, the ClearML Agent executing the experiment must be running in
To [rerun](webapp_exp_tuning.md) a task through the UI in the listed container, the ClearML Agent executing the task must be running in
Docker mode:
```bash
@ -119,7 +118,7 @@ All parameters and configuration objects appear in the **CONFIGURATION** tab.
### Hyperparameters
Hyperparameters are grouped by their type and appear in **CONFIGURATION** **>** **HYPERPARAMETERS**. Once an experiment
Hyperparameters are grouped by their type and appear in **CONFIGURATION** **>** **HYPERPARAMETERS**. Once a task
is run and stored in ClearML Server, any of these hyperparameters can be [modified](webapp_exp_tuning.md#modifying-experiments).
#### Command Line Arguments
@ -157,14 +156,14 @@ The **TF_DEFINE** parameter group shows automatic TensorFlow logging.
### User Properties
User properties allow to store any descriptive information in a key-value pair format. They are editable in any experiment,
except experiments whose status is *Published* (read-only).
User properties allow to store any descriptive information in a key-value pair format. They are editable in any task,
except *Published* ones (read-only).
![User properties section](../img/webapp_tracking_21.png)
### Configuration Objects
ClearML tracks experiment (Task) model configuration objects, which appear in **Configuration Objects** **>** **General**.
ClearML tracks a task's model configuration objects, which appear in **Configuration Objects** **>** **General**.
These objects include those that are automatically tracked, and those connected to a Task in code (see [`Task.connect_configuration`](../references/sdk/task.md#connect_configuration)).
![Configuration objects](../img/webapp_tracking_24.png)
@ -200,7 +199,7 @@ The task's input and output models appear in the **ARTIFACTS** tab. Each model e
* ID
* Configuration.
Input models also display their creating experiment, which on-click navigates you to the experiment's page.
Input models also display their creating task, which on-click navigates you to the task's page.
![Models in Artifacts tab](../img/webapp_exp_artifacts_01.png)
@ -210,10 +209,10 @@ to navigate to its page in the **MODELS** tab (see [Model Details](webapp_model_
## Info
The **INFO** tab shows extended experiment information:
* [Latest experiment events log](#latest-events-log)
* [Experiment description](#description)
* [Experiment details](#experiment-details)
The **INFO** tab shows extended task information:
* [Latest task events log](#latest-events-log)
* [Task description](#description)
* [Task details](#task-details)
### Latest Events Log
@ -237,19 +236,19 @@ ClearML maintains a system-wide, large but strict limit for task history items.
:::
### Description
Add descriptive text to the experiment in the **Description** section. To modify the description, hover over the
Add descriptive text to the task in the **Description** section. To modify the description, hover over the
description box and click **Edit**.
### Experiment Details
The **Experiment Details** section lists information describing the experiment:
### Task Details
The **Task Details** section lists information describing the task:
* The parent experiment
* The parent task
* Project name
* Creation, start, and last update dates and times
* User who created the experiment
* Experiment state (status)
* Whether the experiment is archived
* Runtime properties - Information about the machine running the experiment, including:
* User who created the task
* Task state (status)
* Whether the task is archived
* Runtime properties - Information about the machine running the task:
* Operating system
* CUDA driver version
* Number of CPU cores
@ -259,21 +258,21 @@ The **Experiment Details** section lists information describing the experiment:
* Host name
* Processor
* Python version
* Experiment Progress
* Task Progress
![Info tab](../img/webapp_tracking_31.png)
## Experiment Results
## Task Results
:::tip Embedding ClearML Visualization
You can embed experiment plots and debug samples into ClearML [Reports](webapp_reports.md). These visualizations are
updated live as the experiment(s) updates. The Enterprise Plan and Hosted Service support embedding resources in external
You can embed task plots and debug samples into ClearML [Reports](webapp_reports.md). These visualizations are
updated live as the task(s) updates. The Enterprise Plan and Hosted Service support embedding resources in external
tools (e.g. Notion). See [Plot Controls](#plot-controls).
:::
### Console
The complete experiment log containing everything printed to stdout and stderr appears in the **CONSOLE** tab. The full log
The complete task log containing everything printed to stdout and stderr appears in the **CONSOLE** tab. The full log
is downloadable. To view the end of the log, click **Jump to end**.
![Console tab](../img/webapp_tracking_32.png)
@ -289,7 +288,10 @@ Scalar series can be displayed in [graph view](#graph-view) (default) or in [met
#### Graph View
Scalar graph view (<img src="/docs/latest/icons/ico-charts-view.svg" alt="Graph view" className="icon size-md space-sm" />)
shows scalar series plotted as a time series line chart. The series are sub-sampled for
shows scalar series plotted as a time series line chart. By default, a single plot is shown for each scalar metric,
with all variants overlaid within.
The series are sub-sampled for
display efficiency. For high resolution, view a series in full screen mode by hovering over the graph and clicking <img src="/docs/latest/icons/ico-maximize.svg" alt="Maximize plot icon" className="icon size-md space-sm" />.
:::info Full Screen Refresh
@ -304,26 +306,24 @@ a `Summary` table.
Use the scalar tools to improve analysis of scalar metrics. In the info panel, click <img src="/docs/latest/icons/ico-settings.svg" alt="Settings gear" className="icon size-md space-sm" /> to use the tools. In the full screen details view, the tools
are on the left side of the window. The tools include:
* **Group by** - Select one of the following:
* **Metric** - All variants for a metric on the same plot
<Collapsible type="screenshot" title="View a screenshot">
<a id="group_by"/>
* **Group by** - Select one of the following:
* **Metric** - Displays all variants for a metric on the same plot. For example, if you have a "Test" metric with
"loss" and "accuracy" variants, both variants will appear on the same plot that is titled "Test".
![Plots grouped by metric](../img/webapp_tracking_33.png)
</Collapsible>
* **None** - Group by metric and variant (individual metric-variant plots).
<Collapsible type="screenshot" title="View a screenshot">
* **None** - Displays individual plots for each metric-variant combination, grouped into sections by metric. For
example, a "Test" metric with "loss" and "accuracy" variants will have a separate plot for each variant under the
"Test" group.
![Plots groups my metric and variant](../img/webapp_tracking_34.png)
</Collapsible>
* Horizontal axis - Select the x-axis units:
* Iterations
* Time from start - Time since experiment began
* Time from start - Time since task began
* Wall time - Local clock time
* Curve smoothing - Choose which smoothing algorithm to use from the dropdown menu: Exponential moving average, Gaussian,
or Running Average. Use the slider to configure the smoothing factor or specify a value manually.
@ -331,8 +331,14 @@ are on the left side of the window. The tools include:
plots to display. For example, to display specific plots, click **HIDE ALL**, and then click <img src="/docs/latest/icons/ico-show.svg" alt="Eye Show" className="icon size-md space-sm" />
on each plot you want to view.
To embed scalar plots in your [Reports](webapp_reports.md), hover over a plot and click <img src="/docs/latest/icons/ico-plotly-embed-code.svg" alt="Embed code" className="icon size-md space-sm" />,
which will copy to clipboard the embed code to put in your Reports. In contrast to static screenshots, embedded resources
To embed scalar plots in your [Reports](webapp_reports.md), hover over a plot and click Embed <img src="/docs/latest/icons/ico-plotly-embed-code.svg" alt="Embed code" className="icon size-md space-sm" />,
which will copy to clipboard the embed code to put in your Reports. To quickly get the embed codes for all plots of a
specific metric, click Embed <img src="/docs/latest/icons/ico-plotly-embed-code.svg" alt="Embed code" className="icon size-md space-sm" />
on the group section header (available when plots are [grouped by](#group_by) `None`).
![Embed metric group](../img/webapp_tracking_34b.png)
In contrast to static screenshots, embedded resources
are retrieved when the report is displayed allowing your reports to show the latest up-to-date data.
See additional [plot controls](#plot-controls) below.
@ -367,7 +373,10 @@ Plotly plots. Individual plots can be shown / hidden or filtered by title.
![Plots tab](../img/webapp_tracking_35.png)
For each metric, the latest reported plot is displayed.
Plots are grouped into sections by metric. To quickly get the embed codes for all plots of a specific metric, click Embed <img src="/docs/latest/icons/ico-plotly-embed-code.svg" alt="Embed code" className="icon size-md space-sm" />
on the group section header.
For each metric/variant combination, the latest reported plot is displayed.
When viewing a plot in full screen (<img src="/docs/latest/icons/ico-maximize.svg" alt="Maximize plot icon" className="icon size-md space-sm" />),
older iterations are available through the iteration slider (or using the up/down arrow keyboard shortcut). Go to the
@ -416,7 +425,7 @@ These controls allow you to better analyze the results. Hover over a plot, and t
### Debug Samples
Experiment outputs such as images, audio, and videos appear in **DEBUG SAMPLES**. These include data generated by
Task outputs such as images, audio, and videos appear in **DEBUG SAMPLES**. These include data generated by
libraries and visualization tools, and explicitly reported using the [ClearML Logger](../fundamentals/logger.md).
You can view debug samples by metric in the reported iterations. Filter the samples by metric by selecting a metric from the
@ -435,7 +444,7 @@ buttons (or using the left/right arrow keyboard shortcut).
![Debug Samples image viewer](../img/webapp_tracking_44.png)
## Tagging Experiments
## Tagging Tasks
<div class="vid" >
<iframe style={{position: 'absolute', top: '0', left: '0', bottom: '0', right: '0', width: '100%', height: '100%'}}
@ -449,18 +458,18 @@ buttons (or using the left/right arrow keyboard shortcut).
<br/>
Tags are user-defined, color-coded labels that can be added to experiments (and pipelines, datasets, and models),
allowing to easily identify and group experiments. Tags can help in organizing, querying, and automating experiments.
For example, tag experiments by the machine type used to execute them, label versions, team names, or any other
Tags are user-defined, color-coded labels that can be added to tasks (and pipelines, datasets, and models),
allowing to easily identify and group tasks. Tags can help in organizing, querying, and automating tasks.
For example, tag tasks by the machine type used to execute them, label versions, team names, or any other
category.
You can use tags to filter your experiments in your experiment table (see [Filtering Columns](webapp_exp_table.md#filtering-columns))
or when querying experiments in your code (see [Tag Filters](../clearml_sdk/task_sdk.md#tag-filters)). You can trigger
experiment execution according to their tags (see [TriggerScheduler](../references/sdk/trigger.md)) or automatically
You can use tags to filter the tasks in your task table (see [Filtering Columns](webapp_exp_table.md#filtering-columns))
or when querying tasks in your code (see [Tag Filters](../clearml_sdk/task_sdk.md#tag-filters)). You can trigger
task execution according to their tags (see [TriggerScheduler](../references/sdk/trigger.md)) or automatically
deploy models according to their tags (see [ClearML Serving](../clearml_serving/clearml_serving_tutorial.md#automatic-model-deployment)).
**To add tags:**
1. Click the experiment **>** Hover over the tag area **>** **+ADD TAG** or <img src="/docs/latest/icons/ico-bars-menu.svg" alt="Bars menu" className="icon size-md space-sm" /> (menu)
1. Click the task **>** Hover over the tag area **>** **+ADD TAG** or <img src="/docs/latest/icons/ico-bars-menu.svg" alt="Bars menu" className="icon size-md space-sm" /> (menu)
1. Do one of the following:
* Add a new tag - Type the new tag name **>** **(Create New)**.
* Add an existing tag - Click a tag.
@ -471,6 +480,6 @@ deploy models according to their tags (see [ClearML Serving](../clearml_serving/
## Locating the Experiment (Task) ID
## Locating the Task ID
The task ID appears in the experiment page's header.
The task ID appears in the task page's header.