TensorBoard PR Curve

The tensorboard_pr_curve.py example demonstrates the integration of ClearML into code that uses TensorFlow and TensorBoard.

The example script does the following:

• Creates a task named tensorboard pr_curve in the examples project.
• Creates three classes, R, G, and B, and generates colors within the RGB space from normal distributions. The true label of each random color is associated with the normal distribution that generated it.
• Computes the probability that each color belongs to the class, using three other normal distributions.
• Generate PR curves using those probabilities.
• Creates a summary per class using tensorboard.plugins.pr_curve.summary.
• ClearML automatically captures TensorBoard output, TensorFlow Definitions, and output to the console.

Plots

In the ClearML Web UI, the PR Curve summaries appear in the task's page under PLOTS.

• Blue PR curves

  

• Green PR curves

  

• Red PR curves

  

Hyperparameters

ClearML automatically logs TensorFlow Definitions. They appear in CONFIGURATION > HYPERPARAMETERS > TF_DEFINE.

Console

All console output appears in CONSOLE tab.