Titan25 is an MLIP dataset constructed with GAIA.
It comprises 1.8M data points across 11 elements (C, H, N, O, Ag, Au, Cu, Pd, Pt, Rh, Ru),
which supports the development of MLIPs that robustly capture general chemical reactions, including interfacial processes.
To accelerate MLIP training, we recommend converting .extxyz files to .lmdb or similar formats.
GAIA paper includes
full details on the construction of the dataset
the evaluation results on the dataset
We also… See the full description on the dataset page:
https://huggingface.co/datasets/aixsim/Titan25.
