Machine Learning-Based Prediction of the Martensite Start Temperature

This repository presents the PyTorch code for predicting the Martensite start temperature for steel alloys. Pull the repository and use the src/main.py to test your own data or use our web application.

Free Web Application

A ready to use prediction tool can be found here: https://eah-jena-ms-predictor.streamlit.app/ You can host the application yourself locally using the streamlit_app.py.

Citation

@article{Wentzien2024,
author = {Wentzien, Marcel and Koch, Marcel and Friedrich, Thomas and Ingber, Jerome and Kempka, Henning and Schmalzried, Dirk and Kunert, Maik},
title = {Machine Learning-Based Prediction of the Martensite Start Temperature},
journal = {steel research international},
volume = {n/a},
number = {n/a},
pages = {2400210},
doi = {https://doi.org/10.1002/srin.202400210},
url = {https://onlinelibrary.wiley.com/doi/abs/10.1002/srin.202400210},
}

Database

Complete Database available in this repository as csv: https://github.com/EAH-Materials/MartensiteStart_DeepLearning/blob/main/data/MsDatabase_2022_complete.csv