OCP10k

OCP 10 dataset https://github.com/Open-Catalyst-Project/ocp/blob/main/MODELS.md#is2re-models from jarvis.core.atoms import Atoms from jarvis.core.specie import atomic_numbers_to_symbols from ocpmodels.datasets import SinglePointLmdbDataset from jarvis.db.jsonutils import dumpjson from collections import defaultdict # dataset = SinglePointLmdbDataset({"src": "/wrk/knc6/oc/ocp/data/is2re/10k/train/data.lmdb"}) train = SinglePointLmdbDataset( {"src": "ocp/data/is2re/10k/train/data.lmdb"} ) val = SinglePointLmdbDataset( {"src": "ocp/data/is2re/all/val_id/data.lmdb"} ) test = SinglePointLmdbDataset( {"src": "ocp/data/is2re/all/test_id/data.lmdb"} ) print("n_train=", len(train)) print("n_val=", len(val)) print("n_test=", len(test))