Examples

This section contains comprehensive examples demonstrating how to use SACRO-ML for various machine learning scenarios and frameworks.

• Notebook Examples
• Scikit-learn Examples
• PyTorch Examples
• Risk Assessment Examples
• SafeModel Examples
• Unsupported Model Examples
• User Stories

Example summaries

Below are short descriptions of each examples page. Click a link to open the corresponding examples document.

• Notebook examples — Interactive Jupyter notebooks demonstrating SACRO-ML usage across different models and workflows; good for quick-start tutorials and step-by-step demonstrations.

• Scikit-learn examples — Examples focused on the scikit-learn ecosystem: dataset preparation, model training, evaluation, and disclosure risk analyses.

• PyTorch examples — PyTorch-based examples (including CIFAR and simple model training) illustrating integration with SACRO-ML for deep learning use cases.

• Risk examples — Scripts and notebooks centered on risk assessment methods, metrics, and reporting approaches.

• SafeModel examples — Demonstrations of safe model construction and usage of the SACRO-ML SafeModel API and reporting utilities.

• Unsupported examples — Experimental or unsupported example code showing edge-case usage and advanced techniques for experienced users.

• User stories — Real-world user story notebooks and scripts illustrating typical workflows and how SACRO-ML can be applied to practical scenarios.