fire_step

torch_sim.optimizers.fire.fire_step(state, model, *, dt_max=1.0, n_min=5, f_inc=1.1, f_dec=0.5, alpha_start=0.1, f_alpha=0.99, max_step=0.2, fire_flavor='ase_fire')[source]

Perform one FIRE optimization step.

Parameters:
  • model (ModelInterface) – Model that computes energies, forces, and optionally stress

  • state (FireState | CellFireState) – Current FIRE optimization state

  • dt_max (float) – Maximum allowed timestep

  • n_min (int) – Minimum steps before timestep increase

  • f_inc (float) – Factor for timestep increase when power is positive

  • f_dec (float) – Factor for timestep decrease when power is negative

  • alpha_start (float) – Initial velocity mixing parameter

  • f_alpha (float) – Factor for mixing parameter decrease

  • max_step (float) – Maximum distance an atom can move per iteration

  • fire_flavor (FireFlavor) – Optimization flavor (“vv_fire” or “ase_fire”)

Returns:

Updated FireState after one optimization step

Return type:

FireState | CellFireState