gradient_descent_init

torch_sim.optimizers.gradient_descent.gradient_descent_init(state, model, *, cell_filter=None, **filter_kwargs)[source]

Initialize a gradient descent optimization state.

Parameters:

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

  • state (SimState | dict[Literal['positions', 'masses', 'cell', 'pbc', 'atomic_numbers', 'system_idx'], ~torch.Tensor]) – SimState containing positions, masses, cell, etc.

  • cell_filter (CellFilter | CellFilterFuncs | None) – Filter for cell optimization (None for position-only optimization)

  • **filter_kwargs (Any) – Additional arguments passed to cell filter initialization

Returns:

Initialized OptimState with forces, energy, and optional cell state

Return type:

OptimState | CellOptimState

Notes

Use cell_filter=None for position-only optimization. Use cell_filter=UNIT_CELL_FILTER or FRECHET_CELL_FILTER for cell optimization.