Processed Fe-C-O Training Dataset and MACE Model for Bulk-Mediated Carbon-Oxygen Chemistry in Iron Catalysts

Zenodo release package for the Fe-C-O MACE manuscript.

Files to upload:- dataset.tgz: processed train/validation/test EXTXYZ files.- fe-c-o-mace.model: trained MACE model used for MD.- mace-train.sh: SLURM training wrapper.- mace-config.yaml: MACE training configuration.- mace-preprocess.sh: preprocessing script for generating HDF5 inputs.- VASP-AIMD.tgz: VASP AIMD input/provenance package containing POSCAR,  CONTCAR, INCAR, KPOINTS, and source-folder metadata. Raw OUTCAR files are not  included here because the source OUTCAR set is approximately 22 GB.- MANIFEST.md: file sizes and MD5 checksums.