Data#
A comprehensive data handling pipeline with modular components for anomaly detection tasks.
Core data structures that define how data is represented and validated throughout the pipeline. Features type-safe containers, dual backend support, and automatic validation.
Ready-to-use PyTorch Dataset implementations of standard benchmark datasets (MVTec, BTech) and support for custom datasets across multiple modalities (Image, Video, Depth).
Lightning implementations of these PyTorch datasets that provide automated data loading, train/val/test splitting, and distributed training support through the PyTorch Lightning DataModule interface.
Additional Resources#
Helper functions and utilities for data processing and augmentation.
Step-by-step guides on using the data components.