# facebookresearch/fairchem

FAIR Chemistry's library of machine learning methods for chemistry

Repository: https://github.com/facebookresearch/fairchem
Canonical: https://ross.abutalabs.com/products/fairchem
Homepage: https://fair-chem.github.io/
Language: Python
License: NOASSERTION
License Family: other
Last push: 2026-08-26T21:23:13+00:00

## Health v2 (maintenance only)
Score: 99/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 98, longevity 100
- inputs: {"age_days": 2533, "days_push": 7, "days_rel": 14, "gap_med": 4, "n_releases_24m": 58}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2231, forks 499 (observed 2026-08-28T04:06:29.087981+00:00)

## What it is
FAIR Chemistry's centralized Python library of machine learning models, datasets, and applications for materials science and quantum chemistry. Its flagship UMA model is a universal machine-learning interatomic potential trained on 500M+ DFT calculations for molecules, materials, and catalysts.

## Use cases
- predict adsorption energies for heterogeneous catalysis
- run molecular dynamics with a universal interatomic potential
- compute phonons and elastic properties of bulk materials
- generate molecular conformers and predict electronic properties
- simulate molecular crystals and MOFs for CO2 capture
- fine-tune pretrained atomistic models on custom datasets
- run large-scale GPU-accelerated simulations via LAMMPS

## When to choose
- you need state-of-the-art ML potentials for atomistic simulations across molecules, materials, and catalysts
- you want to fine-tune or train graph neural network potentials on your own DFT data
- you need fast, energy-conserving inference as an alternative to expensive DFT calculations

## When to avoid
- you need exact DFT-level accuracy or Materials Project-compatible energetics without careful corrections
- you depend on FAIRChem v1 pretrained models, since v2 is a breaking change
- your work is unrelated to atomistic/chemical simulation

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, simulation, sdk
- domain: chemistry, machine-learning
- platform: python
- tags: interatomic-potentials, uma, molecular-dynamics, catalysis, materials-science, dft, atomistic-simulation, lammps, quantum-chemistry, computational-chemistry, linux, macos, gpu

## Member repositories
- facebookresearch/fairchem (main) score 99

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:29.087981+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:44:26.950189+00:00, confidence not recorded.
  - readme: https://github.com/facebookresearch/fairchem (fetched 2026-08-28T04:06:29.087981+00:00, sha 67ea748c92d0)
  - homepage: https://fair-chem.github.io/ (fetched 2026-08-29T10:25:15.565291+00:00, sha 26147758f407)
- Data as of 2026-08-30T08:39:29.467469+00:00.