This section provides practical guides for common tasks you will encounter when working with FAIRChem models. Whether you are running inference, training models, or integrating with simulation tools, these guides will help you get started quickly.
Use FAIRChem models with ASE for single-point calculations, relaxations, and molecular dynamics.
Efficiently run predictions on many structures using batched inference.
Run many independent ASE simulations with concurrent batched inference for improved GPU utilization.
Create custom datasets from ASE databases, CIF files, trajectories, and other formats.
Fine-tune pretrained UMA models on your own datasets for improved accuracy.
Train models from scratch using the FAIRChem training framework.
Evaluate pretrained models on standard benchmarks and metrics.
Run downstream property benchmarks like relaxations, elastic tensors, and phonons.
Use FAIRChem models with LAMMPS for large-scale molecular dynamics simulations.
Integrate FAIRChem with QuAcc for complex molecular simulation workflows.