A Benchmark for Quantum Chemistry Relaxations via Machine Learning Interatomic Potentials

TL;DR

This paper introduces PubChemQCR, a large-scale dataset of DFT-based molecular relaxation trajectories for small organic molecules, enabling the development and benchmarking of machine learning interatomic potentials for quantum chemistry applications.

Contribution

The paper provides the largest publicly available dataset of DFT relaxation trajectories, including energy and force labels, and benchmarks multiple MLIP models for quantum chemistry simulations.

Findings

PubChemQCR contains approximately 3.5 million trajectories and 300 million conformations.

Benchmark results highlight the performance of nine MLIP models on the dataset.

The dataset facilitates the development of transferable MLIPs for molecular relaxation tasks.

Abstract

Computational quantum chemistry plays a critical role in drug discovery, chemical synthesis, and materials science. While first-principles methods, such as density functional theory (DFT), provide high accuracy in modeling electronic structures and predicting molecular properties, they are computationally expensive. Machine learning interatomic potentials (MLIPs) have emerged as promising surrogate models that aim to achieve DFT-level accuracy while enabling efficient large-scale atomistic simulations. The development of accurate and transferable MLIPs requires large-scale, high-quality datasets with both energy and force labels. Critically, MLIPs must generalize not only to stable geometries but also to intermediate, non-equilibrium conformations encountered during atomistic simulations. In this work, we introduce PubChemQCR, a large-scale dataset of molecular relaxation trajectories curated from the raw geometry optimization outputs of the PubChemQC project. PubChemQCR is the largest publicly available dataset of DFT-based relaxation trajectories for small organic molecules, comprising approximately 3.5 million trajectories and over 300 million molecular conformations computed at various levels of theory. Each conformation is labeled with both total energy and atomic forces, making the dataset suitable for training and evaluating MLIPs. To provide baselines for future developments, we benchmark nine representative MLIP models on the dataset. Our resources are publicly available at https://huggingface.co/divelab