{"success":true,"data":{"author":"Meta FAIR Chemistry, Lawrence Berkeley National Laboratory and collaborators (Levine et al.)","bucketAccessType":"public","bucketListable":true,"bucketName":"dl.fbaipublicfiles.com","bucketPrefix":"","category":"Chemistry & Materials","cloudProvider":"api","createdAt":"2026-11-19T00:05:36.113Z","description":"The original files of Open Molecules 2025 (OMol25), the molecular DFT dataset from Meta FAIR Chemistry, Lawrence Berkeley National Laboratory and collaborators: more than 100 million single-point calculations at wB97M-V/def2-TZVPD in ORCA 6 over biomolecules, metal complexes, electrolytes and recomputed community datasets, covering 83 elements and systems of up to 350 atoms. Every structure carries total energy (eV) and atomic forces (eV/A), with total charge and spin multiplicity in atoms.info next to per-structure extras such as HOMO-LUMO gap and Mulliken, Lowdin and NBO charges. This row mounts the six tar.gz archives Meta serves from its public file server, exactly as the fairchem documentation and training configs expect them: train_4M (20 GB, the 3,986,754 structure subset), val (21 GB, 2,762,021 structures with labels), test (8 GB, 2,805,046 structures for leaderboard submissions), neutral_train (108 GB), neutral_val (125 MB, 27,697 structures) and the full train split (489 GB). Each archive unpacks to a folder of ASE DB LMDB shards (data0000.aselmdb and so on), which fairchem opens with AseDBDataset. Use these, not a Parquet conversion, to train or fine-tune UMA, eSEN and other machine learning interatomic potentials with fairchem, or to reproduce the published splits. The archives are gzip streams, so copy one to notebook storage and unpack it there; train_4M unpacks to about 22 GB. The data are CC-BY-4.0. The model checkpoints in the gated facebook/OMol25 Hugging Face repository are under the separate FAIR Chemistry License and are not part of this row. Cite Levine et al., arXiv:2505.08762 (2025).","doi":"10.48550/arXiv.2505.08762","fileManifestCount":6,"fileManifestTotal":6,"fuseMountBrowsable":true,"githubUrl":"https://github.com/facebookresearch/fairchem","id":"bd68f492-0395-5be1-b4cd-9df7088a6e01","license":"CC-BY-4.0","licenseEvidenceUrl":"https://raw.githubusercontent.com/facebookresearch/fairchem/main/docs/molecules/datasets/omol25.md","referenceUrl":"https://doi.org/10.48550/arXiv.2505.08762","region":"us-east-1","schemaCard":{"accessHints":[{"language":"python","snippet":"# pip install fairchem-core   (brings ase and the aselmdb backend)\nimport os, tarfile\nfrom fairchem.core.datasets import AseDBDataset\nbase = os.environ.get(\"SCIGANTIC_MOUNT_PATH\", \"/mnt/http-archive/data\")\n# Start with the smallest split: 125 MB, 27,697 neutral validation structures\nwith tarfile.open(f\"{base}/neutral_val.tar.gz\", \"r|gz\") as t:\n    t.extractall(\"/tmp/omol25\", filter=\"data\")\nds = AseDBDataset({\"src\": \"/tmp/omol25/neutral_val\"})\nprint(len(ds), \"structures\")\natoms = ds.get_atoms(0)\nprint(atoms.get_chemical_formula(), \"charge\", atoms.info[\"charge\"], \"spin\", atoms.info[\"spin\"])\nprint(\"energy (eV):\", atoms.get_potential_energy(), \"forces:\", atoms.get_forces().shape)"}],"author":"hand-authored (seed-omol25-original-files.ts)","caveats":["The files are gzip-compressed tars, so they cannot be opened in place: stream one with tarfile mode \"r|gz\" into local storage first. train_4M needs about 22 GB unpacked, val about 23 GB; the full train split is 489 GB compressed and needs a large volume.","test.tar.gz is the leaderboard split; score predictions through the fairchem leaderboard rather than locally.","Reading *.aselmdb needs fairchem-core (it registers the aselmdb backend for ase.db).","Model checkpoints (UMA, eSEN) are not here. They sit in the gated facebook/OMol25 Hugging Face repo under the FAIR Chemistry License."],"columns":[{"description":"Geometry in angstrom; no periodic cell","name":"positions, numbers","type":"ase.Atoms"},{"description":"Total DFT energy in eV, atoms.get_potential_energy()","name":"energy","type":"float"},{"description":"Atomic forces in eV/A, atoms.get_forces()","name":"forces","type":"float[n_atoms,3]"},{"description":"Total charge of the system","name":"info.charge","type":"int"},{"description":"Spin multiplicity (2S+1)","name":"info.spin","type":"int"},{"description":"Which OMol25 subset the structure came from","name":"info.source","type":"str"},{"description":"Frontier orbital values from the same calculation","name":"info.homo_lumo_gap, homo_energy","type":"float"},{"description":"Per-atom partial charges","name":"info.mulliken_charges, lowdin_charges, nbo_charges","type":"float[n_atoms]"}],"format":"tar.gz archives of ASE DB LMDB shards (*.aselmdb)","generatedAt":"2026-11-18T00:00:00Z","layout":{"example":"neutral_val.tar.gz -> neutral_val/data0000.aselmdb","pattern":"<split>.tar.gz -> <split>/data<NNNN>.aselmdb (+ .aselmdb-lock)"},"mountPath":"/mnt/http-archive/data","sampleObjects":["neutral_val.tar.gz","test.tar.gz","train_4M.tar.gz","val.tar.gz","neutral_train.tar.gz","train.tar.gz"],"starterCell":"# pip install fairchem-core   (brings ase and the aselmdb backend)\nimport os, tarfile\nfrom fairchem.core.datasets import AseDBDataset\nbase = os.environ.get(\"SCIGANTIC_MOUNT_PATH\", \"/mnt/http-archive/data\")\n# Start with the smallest split: 125 MB, 27,697 neutral validation structures\nwith tarfile.open(f\"{base}/neutral_val.tar.gz\", \"r|gz\") as t:\n    t.extractall(\"/tmp/omol25\", filter=\"data\")\nds = AseDBDataset({\"src\": \"/tmp/omol25/neutral_val\"})\nprint(len(ds), \"structures\")\natoms = ds.get_atoms(0)\nprint(atoms.get_chemical_formula(), \"charge\", atoms.info[\"charge\"], \"spin\", atoms.info[\"spin\"])\nprint(\"energy (eV):\", atoms.get_potential_energy(), \"forces:\", atoms.get_forces().shape)"},"sourceType":"http","title":"Open Molecules 2025 (OMol25), Original ASE-LMDB Release from Meta FAIR Chemistry","updatedAt":"2026-11-19T00:06:22.209Z","visibility":"public","websiteUrl":"https://fair-chem.github.io/molecules/datasets/omol25.html"}}