NVTVRescaleState

class torch_sim.integrators.nvt.NVTVRescaleState(positions, masses, cell, pbc, atomic_numbers, charge=None, spin=None, system_idx=None, *, momenta, energy, forces)

Bases: MDState

State information for an NVT system with a V-Rescale thermostat.

This class represents the complete state of a molecular system being integrated in the NVT (constant particle number, volume, temperature) ensemble using a Velocity Rescaling thermostat. The thermostat maintains constant temperature through stochastic velocity rescaling.

Variables:

• positions (Tensor) – Particle positions with shape [n_particles, n_dimensions]

• masses (Tensor) – Particle masses with shape [n_particles]

• cell (Tensor) – Simulation cell matrix with shape [n_dimensions, n_dimensions]

• pbc (Tensor | list[bool] | bool) – Whether to use periodic boundary conditions

• momenta (Tensor) – Particle momenta with shape [n_particles, n_dimensions]

• energy (Tensor) – Energy of the system

• forces (Tensor) – Forces on particles with shape [n_particles, n_dimensions]

Parameters:

• positions (Tensor)

• masses (Tensor)

• cell (Tensor)

• pbc (Tensor | list[bool] | bool)

• atomic_numbers (Tensor)

• charge (Tensor | None)

• spin (Tensor | None)

• system_idx (Tensor | None)

• momenta (Tensor)

• energy (Tensor)

• forces (Tensor)

Notes

• The V-Rescale thermostat provides proper canonical sampling

• Stochastic velocity rescaling ensures correct temperature distribution

• Time-reversible when integrated with appropriate algorithms

get_number_of_degrees_of_freedom()

Calculate the degrees of freedom per system.

Return type:

Tensor