Skip to content

Project Structure Guide

AUDIT supports flexible project organization, but we recommend following the default structure for clarity, reproducibility, and ease of use. This structure is modular and designed to help you manage datasets, configurations, outputs, and logs efficiently.

your_project/
├── datasets/
├── config/
├── outputs/
├── logs/

Datasets Directory (datasets/)

Central location for all datasets used in training and testing. Organize each dataset in its own subdirectory, with images, ground truth segmentations, and model predictions clearly separated.

Info

The reserved word __pred_ is used in AUDIT for model predictions.

Example: 

datasets/
├── dataset_1/
│   ├── dataset_1_images/
│   ├── dataset_1_seg/
│   │   ├── model_1/
│   │   ├── model_2/
├── dataset_2/

Configuration Directory (config/)

Store all configuration files needed for feature extraction, metric evaluation, and app settings.

Example: 

config/
├── feature_extraction.yaml
├── metric_extraction.yaml
├── app.yaml

Outputs Directory (outputs/)

All results generated by AUDIT are saved here, including extracted features and evaluation metrics.

Example: 

outputs/
├── features/
├── metrics/

Logs Directory (logs/)

Contains logs for debugging, monitoring, and record-keeping.

Example: 

logs/
├── features/
├── metric/

🏁 Getting Started

  • The recommended structure is automatically created when you run AUDIT for the first time.
  • Example datasets and configuration files are available in the repository.
  • For details on configuration and running AUDIT, see the other sections in the documentation.

Tip

Using the default structure ensures seamless integration and reproducibility.