PyLandscape

Introduction

pylandscape is a Pytorch library for loss landscape analysis of neural networks. The library enables computing the following metrics:

• CKA similarity
• Hessian metrics
• Mode connectivity
• Loss surface

NOTE: All the functionalities relative to the computation of the Hessian metrics have been embedded via PyHessian. If your interested in learning more about how these metrics are computed have a look to their Repository.

Usage

Install from Pip

You can install the library from pip:

pip install pylandscape

Research

If you used Pylandscape consider to cite:

@misc{baldi2025losslandscapeanalysisreliable,
      title={Loss Landscape Analysis for Reliable Quantized ML Models for Scientific Sensing}, 
      author={Tommaso Baldi and Javier Campos and Olivia Weng and Caleb Geniesse and Nhan Tran and Ryan Kastner and Alessandro Biondi},
      year={2025},
      eprint={2502.08355},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/2502.08355}, 
}