do_rss#
- autoplex.data.rss.jobs.do_rss(mlip_type=None, iteration_index=None, mlip_path=None, structure=None, scalar_pressure_method='exp', scalar_exp_pressure=100, scalar_pressure_exponential_width=0.2, scalar_pressure_low=0, scalar_pressure_high=50, max_steps=1000, force_tol=0.01, stress_tol=0.01, Hookean_repul=False, hookean_paras=None, write_traj=True, num_processes_rss=1, device='cpu', isol_es=None)[source]#
Perform sandom structure searching (RSS) using a MLIP.
- Parameters:
mlip_type (str, mandatory) – Choose one specific MLIP type: ‘GAP’ | ‘ACE’ | ‘NequIP’ | ‘M3GNet’ | ‘MACE’.
iteration_index (str, mandatory) – Index for the current iteration.
mlip_path (str, mandatory) – Path to the MLIP model.
structure (list of Structure, mandatory) – List of structures to be relaxed.
scalar_pressure_method (str, optional) – Method for scalar pressure. Default is ‘exp’.
scalar_exp_pressure (float, optional) – Scalar exponential pressure. Default is 100.
scalar_pressure_exponential_width (float, optional) – Width for scalar pressure exponential. Default is 0.2.
scalar_pressure_low (float, optional) – Low limit for scalar pressure. Default is 0.
scalar_pressure_high (float, optional) – High limit for scalar pressure. Default is 50.
max_steps (int, optional) – Maximum number of steps for relaxation. Default is 1000.
force_tol (float, optional) – Force tolerance for relaxation. Default is 0.01.
stress_tol (float, optional) – Stress tolerance for relaxation. Default is 0.01.
Hookean_repul (bool, optional) – Whether to apply Hookean repulsion. Default is False.
hookean_paras (dict, optional) – Parameters for Hookean repulsion as a dictionary of tuples. Default is None.
write_traj (bool, optional) – Whether to write trajectory. Default is True.
num_processes_rss (int, optional) – Number of processes used for running RSS.
device (str, optional) – specify device to use cuda or cpu.
isol_es (dict[int, float] | None) –
- Returns:
Output dictionary containing the results of the RSS relaxation.
- Return type:
dict