Overview:

Code repository for machine-learning research position with Professor Mingda Li. Includes a multi-class classification model for predicting the magnetic order of magnetic structures and a binary classification model for predicting the propogation vector of magnetic structures. The input data are from MAGNDATA and Materials Project.

Instructions:

The folder magnetic_ordering includes source code and pretrained models for magnetic order classification. To retrain the model, first download the data from materials project using preload_data/download_structures.ipynb, then run the python script magnetic_ordering.py. To load the pretrained models in folder trained_model, run python script loadmodel.py.

The folder prop_vector includes source code and pretrained models of propagation vector classification. To retrain the model, run python script prop_vector.py.

Questions?

Reach out to Harry Heiberger (harryh@mit.edu), Helena Merker (hmerker@mit.edu), or Linh Nguyen (linhnk@mit.edu).