**calorine: A Python Package for Constructing and Sampling Neuroevolution Potential Models** **Authors:** Eric Lindgren, Magnus Rahm, Erik Fransson, Fredrik Eriksson, Nicklas Österbacka, Zheyong Fan, and Paul Erhart **Journal:** Journal of Open Source Software **DOI:** 10.21105/joss.06264 **Summary:** Molecular dynamics (MD) simulations are crucial in computational chemistry, physics, and materials science for understanding microscopic processes and developing new materials. Traditional methods often use empirical interatomic potentials or force fields, which are fast but inaccurate, or ab-initio methods, which are accurate but computationally expensive. Machine-learned interatomic potentials (MLIPs) have emerged as a promising alternative, combining the speed of heuristic force fields with the accuracy of ab-initio techniques. Neuroevolution potentials (NEPs), implemented in the GPUMD package, are a highly accurate and efficient class of MLIPs. NEP models have been used to study various properties in materials, including radiation damage in tungsten, phase transitions, and thermal transport in two-dimensional materials. **calorine** is a Python package that simplifies the construction, analysis, and use of NEP models via GPUMD. It provides a Python interface to access the functionality of GPUMD, making it easy to integrate into Python-based workflows. **calorine** also exposes two ASE Calculator objects, one using the CPU and one using the GPU, to enhance the transferability of NEP models to other codes and machines without discrete GPUs. Examples of recent work supported by **calorine** include studies on through-plane lattice thermal conductivity in van-der-Waals structures and dynamic modes in halide perovskites under continuous-order phase transitions. **Related Software:** - **PyNEP**: Focuses on NEP construction within GPUMD. - **GPYUMD**: Focuses on MD simulations within GPUMD. **Acknowledgements:** The authors acknowledge contributions from Petter Rosander and funding from the Swedish Research Council, the Swedish Foundation for Strategic Research, and computational resources provided by the National Academic Infrastructure for Supercomputing in Sweden (NAISS) and the Swedish National Infrastructure for Computing (SNIC). **References:** The article includes references to recent publications and software packages related to NEPs and MD simulations.**calorine: A Python Package for Constructing and Sampling Neuroevolution Potential Models** **Authors:** Eric Lindgren, Magnus Rahm, Erik Fransson, Fredrik Eriksson, Nicklas Österbacka, Zheyong Fan, and Paul Erhart **Journal:** Journal of Open Source Software **DOI:** 10.21105/joss.06264 **Summary:** Molecular dynamics (MD) simulations are crucial in computational chemistry, physics, and materials science for understanding microscopic processes and developing new materials. Traditional methods often use empirical interatomic potentials or force fields, which are fast but inaccurate, or ab-initio methods, which are accurate but computationally expensive. Machine-learned interatomic potentials (MLIPs) have emerged as a promising alternative, combining the speed of heuristic force fields with the accuracy of ab-initio techniques. Neuroevolution potentials (NEPs), implemented in the GPUMD package, are a highly accurate and efficient class of MLIPs. NEP models have been used to study various properties in materials, including radiation damage in tungsten, phase transitions, and thermal transport in two-dimensional materials. **calorine** is a Python package that simplifies the construction, analysis, and use of NEP models via GPUMD. It provides a Python interface to access the functionality of GPUMD, making it easy to integrate into Python-based workflows. **calorine** also exposes two ASE Calculator objects, one using the CPU and one using the GPU, to enhance the transferability of NEP models to other codes and machines without discrete GPUs. Examples of recent work supported by **calorine** include studies on through-plane lattice thermal conductivity in van-der-Waals structures and dynamic modes in halide perovskites under continuous-order phase transitions. **Related Software:** - **PyNEP**: Focuses on NEP construction within GPUMD. - **GPYUMD**: Focuses on MD simulations within GPUMD. **Acknowledgements:** The authors acknowledge contributions from Petter Rosander and funding from the Swedish Research Council, the Swedish Foundation for Strategic Research, and computational resources provided by the National Academic Infrastructure for Supercomputing in Sweden (NAISS) and the Swedish National Infrastructure for Computing (SNIC). **References:** The article includes references to recent publications and software packages related to NEPs and MD simulations.