Asparagus is a modular machine learning framework designed for medical imaging applications.

Overview

Built on PyTorch, Lightning, and Hydra, Asparagus provides a complete pipeline for:

• Self-Supervised Pretraining — learn representations from unlabelled data via masked image modelling
• Supervised Training — train segmentation, classification, and regression models from scratch
• Finetuning — transfer pretrained weights to downstream tasks with minimal labelled data
• Evaluation — run large-scale multi-task evaluation via the EvalBox

Repositories