Skip to main content

Matplotlib

tip

If you are not already using ClearML, see Getting Started for setup instructions.

Matplotlib is a Python library for data visualization. ClearML automatically captures plots and images created with matplotlib.

All you have to do is add two lines of code to your script:

from clearml import Task

task = Task.init(task_name="<task_name>", project_name="<project_name>")

This will create a ClearML Task that captures:

  • Git details
  • Source code and uncommitted changes
  • Installed packages
  • Matplotlib visualizations
  • And more

View captured Matplotlib plots and images in the WebApp, in the experiment's Plots and Debug Samples tabs respectively.

Experiment plots

Automatic Logging Control​

By default, when ClearML is integrated into your script, it captures all of your matplotlib visualizations. But, you may want to have more control over what your experiment logs.

To control a task's framework logging, use the auto_connect_frameworks parameter of Task.init(). Completely disable all automatic logging by setting the parameter to False. For finer grained control of logged frameworks, input a dictionary, with framework-boolean pairs.

For example:

auto_connect_frameworks={
   'matplotlib': False, 'tensorflow': False, 'tensorboard': False, 'pytorch': True,
   'xgboost': False, 'scikit': True, 'fastai': True, 'lightgbm': False,
   'hydra': True, 'detect_repository': True, 'tfdefines': True, 'joblib': True,
   'megengine': True, 'catboost': True
}

Manual Logging​

To augment its automatic logging, ClearML also provides an explicit logging interface.

Use Logger.report_matplotlib_figure() to explicitly log a matplotlib figure, and specify its title and series names, and iteration number:

logger = task.get_logger()

area = (40 * np.random.rand(N))**2
plt.scatter(x, y, s=area, c=colors, alpha=0.5)
logger.report_matplotlib_figure(title="My Plot Title", series="My Plot Series", iteration=10, figure=plt)
plt.show()

The logged figure is displayed in the experiment's Plots tab.

Experiment Matplotlib plots

Matplotlib figures can be logged as images by passing report_image=True to Logger.report_matplotlib_figure(). View the images in the experiment's DEBUG SAMPLES tab.

Experiment debug sample

See Manual Matplotlib Reporting example.