Karol-G / GaNDLF

A generalizable application framework for segmentation, regression, and classification using PyTorch

GaNDLF

A Generally Nuanced Deep Learning Framework for segmentation, regression and classification.

Why use this?

• Supports multiple
  • Deep Learning model architectures
  • Data dimensions (2D/3D)
  • Channels/images/sequences
  • Prediction classes
  • Domain modalities (i.e., Radiology Scans and Digitized Histopathology Tissue Sections)
• Robust data augmentation, courtesy of TorchIO
• Built-in nested cross validation (and related combined statistics), with support for parallel HPC-based computing
• Handles imbalanced classes (e.g., very small tumor in large organ)
• Multi-GPU (on the same machine - distributed) training
• Leverages robust open source software
• No need to write any code to generate robust models
• Automatic mixed precision support

Citation

Please cite the following article for GaNDLF (full PDF):

@misc{gandlf2021,
      title={GaNDLF: A Generally Nuanced Deep Learning Framework for Scalable End-to-End Clinical Workflows in Medical Imaging}, 
      author={Sarthak Pati and Siddhesh P. Thakur and Megh Bhalerao and Ujjwal Baid and Caleb Grenko and Brandon Edwards and Micah Sheller and Jose Agraz and Bhakti Baheti and Vishnu Bashyam and Parth Sharma and Babak Haghighi and Aimilia Gastounioti and Mark Bergman and Bjoern Menze and Despina Kontos and Christos Davatzikos and Spyridon Bakas},
      year={2021},
      eprint={2103.01006},
      archivePrefix={arXiv},
      primaryClass={cs.LG}
}

Documentation

Start at https://cbica.github.io/GaNDLF/. Includes the following:

• Installation
• Usage
• Extension

Disclaimer

• The software has been designed for research purposes only and has neither been reviewed nor approved for clinical use by the Food and Drug Administration (FDA) or by any other federal/state agency.
• This code (excluding dependent libraries) is governed by the license provided in https://www.med.upenn.edu/cbica/software-agreement.html unless otherwise specified.

Contact

For more information or any support, please post on the Discussions section or contact CBICA Software.