{"success":true,"data":{"author":"Microsoft Research AI for Science","bucketAccessType":"public","bucketName":"microsoft/msr-acc-tae25","bucketPrefix":"","category":"Chemistry & Materials","cloudProvider":"huggingface","createdAt":"2026-11-14T14:15:22.772Z","description":"MSR-ACC/TAE25 is part of the Microsoft Research Accurate Chemistry Collection: 73,040 total atomization energies at the CCSD(T)/CBS level obtained with the W1-F12 thermochemical protocol. It exhaustively covers closed-shell, charge-neutral, covalently bound equilibrium molecules with up to five non-hydrogen atoms drawn from elements up to argon and without significant multireference character. The publishers built it to train and evaluate machine-learned exchange-correlation functionals, density functional and semi-empirical methods, and graph neural networks beyond organic chemistry. Molecules are stored in QCSchema-compatible columns (symbols, geometry, charge, multiplicity) with reference TAE labels (tae@w1-f12), component energies, singlet-triplet gaps and molecular graphs. The repository holds three Parquet files (all, train, valid), about 0.13 GB (measured 2026-11-14); open them with pandas, pyarrow or datasets. Released by Microsoft under CDLA-Permissive-2.0 (Zenodo 10.5281/zenodo.15387280).","displayTitle":"MSR-ACC/TAE25: 73k CCSD(T)/CBS Total Atomization Energies","doi":"10.5281/zenodo.15387280","id":"568dcef8-4695-55d9-8e26-a869459ccd2f","license":"CDLA-Permissive-2.0","licenseEvidenceUrl":"https://zenodo.org/records/15387280","objectCountEstimate":5,"probeStatus":"ok","referenceUrl":"https://arxiv.org/abs/2506.14492","region":"us","schemaCard":{"accessHints":[{"language":"python-datasets","snippet":"from datasets import load_dataset\nds = load_dataset(\"microsoft/msr-acc-tae25\")"},{"language":"python-hub","snippet":"from huggingface_hub import snapshot_download\nsnapshot_download(repo_id=\"microsoft/msr-acc-tae25\", repo_type=\"dataset\")"}],"caveats":["HuggingFace metadata: author=microsoft, license:cdla-permissive-2.0, size_categories:10K<n<100K, format:parquet, downloads=193"],"format":"mixed","generatedAt":"2026-11-14T14:15:24.631Z","layout":{"example":"msr-acc-tae25-all.parquet","pattern":"{repoId}/{filename}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"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","encoding":"hex","key":"msr-acc-tae25-all.parquet","truncatedAt":512}],"sampleObjects":["msr-acc-tae25-all.parquet","msr-acc-tae25-train.parquet","msr-acc-tae25-valid.parquet"],"sidecars":[{"bytes":"---\nlicense: cdla-permissive-2.0\ntags:\n- total-atomization-energies\n- quantum-chemistry\n- computational-chemistry\n- chemistry\n- coupled-cluster\npretty_name: MSR-ACC/TAE25\n---\n\n# Microsoft Research - Accurate Chemistry Collection: Total Atomization Energies\n\n## Description\n\nThe Microsoft Research Accurate Chemistry Collection (MSR-ACC) provides a collection of accurate coupled cluster labels for training machine learning functionals.\nMSR-ACC/TAE25 comprising 73,040 total atomization energies at the CCSD(T)/CBS level obtained with the W1-F12 thermochemical protocol.\nThe dataset is constructed to exhaustively cover the chemical space of closed-shell, charge-neutral, covalently bound equilibrium molecular structures containing up to 5 non-hydrogen atoms drawn from elements up to argon and lacking significant multireference character.\nThe dataset generation is described in [Accurate Chemistry Collection: Coupled cluster atomization energies for broad chemical space\n](https://arxiv.org/abs/2506.14492).\n\nMSR-ACC/TAE25 opens the way for developing and evaluating generally applicable machine-learning, density functional theory, and semi-empirical methods, and has already been used for training the first exchange correlation functional to reach chemical accuracy for atomization energies in [Accurate and scalable exchange-correlation with deep learning](https://arxiv.org/abs/2506.14665).\nIn contrast to smaller or more specialized TAE databases, MSR-ACC/TAE25 provides a large and chemically diverse test set for identifying systematic errors and validating approximate electronic structure methods, for example, in describing bond energies across a wide range of challenging s- and p-block compounds.\nThe size, diversity, and accuracy of the dataset make it useful not only for developing deep-learning DFT methods, but also for training and validating models such as graph neural networks, critically testing their ability to generalize beyond the typical organic chemistry focus of datasets like GDB-9.\nFinally, the dataset can be filtered to create highly specific benchmarks to answer practical questions, such as selecting the most accurate functional for a particular type of chemical system (e.g., phosphorus sulfide compounds).\n\n## Schema\n\nThe molecules in this dataset are stored in QCSchema format, the columns *symbols*, *geometry*, *molecular_charge*, and *molecular_multiplicity* are sufficient to create the full QCSchema document.\nAdditionally, the *name* field provides the sum formula and the *atomic_numbers* column provides the nuclear charges of each atom.\nThe SHA1 hash of each QCSchema document is provided in the *sha1* column.\nFinally, the following extra columns are provided:\n\n- *atom-count:all* : Number of all atoms\n- *atom-count:non-h* : Number of non-hydrogen atoms\n- *graph:all* : Molecular graph including all atoms\n- *graph:non-h* : Molecular graph including non-hydrogen atoms\n- *singlet-triplet-gap-s0-t1@\\** : Energy difference between s0 and t1 state at specified level of theory\n- *tae:frac[(T)]@\\** : Fraction of perturbative triples contribution to the total atomization energy\n- *tae@\\** : Total atomization energy at specified level of theory\n- *tae[\\*]@\\** : Component of the total atomization energy at specified level of theory\n\nThe *tae@w1-f12* column provides the reference label with the highest accuracy compared to experiment.\n\n## Usage\n\nTo evaluate on MSR-ACC/TAE25 the user needs to compute the total atomization energy of the molecule.\nThe total atomization energy for a molecule *M* with *n* atoms of type *A* is defined as\n\n$$\n\\text{TAE} = n\\times E(A) - E(M)\n$$\n\nThe atoms *A* are expected to be in their ground state multiplicity.\n\nFor completeness we provide the ground state multiplicities of all atomic species in MSR-ACC/TAE25 here for reference:\n\n- Hydrogen (H): doublet\n- Lithium (Li): doublet\n- Beryllium (Be): singlet\n- Bor (B): doublet\n- Carbon (C): triplet\n- Nitrogen (N): quartet\n- Oxygen (O): triplet\n- Fluorine (F): doublet\n- Sodium (Na): doublet\n- Magnesium (Mg): singlet\n- Aluminium (Al): doublet\n- Silicon (Si)","encoding":"utf8","key":"README.md","name":"README.md","size":7598,"truncated":true}],"sourceObjectKey":"msr-acc-tae25-all.parquet","starterCell":"from datasets import load_dataset\nds = load_dataset(\"microsoft/msr-acc-tae25\")\nprint(ds)","starterSamplePath":"msr-acc-tae25-all.parquet","totalObjects":5},"summary":"MSR-ACC/TAE25 is part of the Microsoft Research Accurate Chemistry Collection: 73,040 total atomization energies at the CCSD(T)/CBS level obtained with the W1-F12 thermochemical…","title":"MSR-ACC/TAE25: 73k CCSD(T)/CBS Total Atomization Energies","updatedAt":"2026-11-14T17:43:12.318Z","visibility":"public","websiteUrl":"https://huggingface.co/datasets/microsoft/msr-acc-tae25"}}