Skip to content

MLatom 3.2.0

Choose a tag to compare

View all tags
@dralgroup dralgroup released this 14 Jul 08:11
· 124 commits to main since this release
v3.2.0
f285bd9

A major release with many new features, usability/performance improvements, and bug fixes:

  • Energy-weighted training of ANI machine learning potentials (downweights high-deformation-energy PES regions).
  • Diffusion Monte Carlo (DMC) for accurate frequencies/ZPVEs, via an interface to PyVibDMC.
  • AIQM1 on the cloud is much faster and more stable (numerical-derivative tweaks).
  • More robust frequency/thermochemistry with PySCF and a new freqprog keyword (PySCF is the default when Gaussian is absent).
  • Improved ORCA interface (DFT).
  • CCSD(T)*/CBS for charged and open-shell species.
  • Geometry optimization with constraints (via ASE).

Paper (energy-weighted training): Tell machine learning potentials what they are needed for, J. Phys. Chem. Lett. 2024, 15, 4451–4460.

Install: pip install mlatom==3.2.0 · full release notes · documentation

Assets 2
Loading