Welcome to FineTuneMLP_SA_Design! It provides data generation,training and evaluation code/data for ration design of single atom catalyst. This projects uses GemNet-OC20 + OC22 pretrained model delveloped in Open Catalyst Project Rational Design of SA Workflow: We explore the promise of fine-tuning strategies that leverage small data to improve global optimization. We provide two successful fine-tuning approaches for Pt/CeO2, which can serve as benchmarks for the research community.

• This codebase was mostly built based on Open Catalyst Project

Please Tuned on for the upcoming publications:

@article{ft_sa_screening,
    author = {Seokhyun Choung†, Wongyu Park† and Jeong Woo Han*},
    title = {Rapid mapping of Stable and Active  Single Atom Active Site Motifs Accelerated by MLP based Automated Workflow},
    journal = {?},
    year = {2024},
    doi = {?},
}