lambda_m
This keyword sets the weight of Cartesian magnetic-force RMSE in the loss.
Syntax
lambda_m <weight>
Alternatively, use both of the following keywords:
lambda_m_parallel <weight_parallel>
lambda_m_perp <weight_perp>
Parameters
The weight must be non-negative and defaults to 1. It multiplies the RMSE of the Cartesian magnetic force
whose unit is eV/muB. Only frames containing mforce:R:3 and
atoms enabled by spin_type enter this RMSE. A
frame without magnetic-force labels can still contribute energy, force, or
virial data.
Set lambda_m 0 to remove the Cartesian magnetic-force RMSE from the
optimization objective.
To fit magnetic-moment magnitude and direction separately, set both lambda_m_parallel and lambda_m_perp instead of lambda_m.
Each coefficient must be finite and non-negative. A zero coefficient disables that component.
These keywords cannot be combined with lambda_m.
When neither is specified, lambda_m controls the Cartesian objective.
Output
In split mode, loss.out adds four columns after rmse_tau_test and before
any energy-difference columns: rmse_m_parallel_train, rmse_m_perp_train,
rmse_m_parallel_test, and rmse_m_perp_test. Read the # columns header;
the Cartesian and torque diagnostics are retained. Training metrics describe
the reported batch. Test metrics are zero
when no test set is supplied.
Examples
Fit the Cartesian magnetic force:
lambda_m 1.0
Alternatively, fit parallel and perpendicular components with separate weights:
lambda_m_parallel 1.0
lambda_m_perp 1.0
Notes
Independent parallel and perpendicular losses
For residual \(\delta\boldsymbol M_i=\boldsymbol M_i-\boldsymbol M_i^{ref}\) and the input spin direction \(\hat{\boldsymbol s}_i\), define
Both reductions use the same labeled, active atoms as lambda_m; the
normalization counts one parallel and two perpendicular degrees of freedom.
At exactly zero spin, squared residuals are assigned isotropically: one third
to the parallel channel and two thirds to the perpendicular channel. Every
nonzero spin defines a direction, without a small-moment cutoff. During
optimization, squared configuration weights multiply each squared
residual; reported RMSEs are unweighted.
SNES minimizes a weighted sum of RMSEs. In general no pair of
constant coefficients makes the sum of the parallel and perpendicular RMSEs
equal to Cartesian RMSE. Use lambda_m to supervise Cartesian magnetic force.
The raw metrics obey \(3 R_M^2=R_\parallel^2+2R_\perp^2\).
lambda_tau controls an independent torque term. Torque weights the
perpendicular residual by spin magnitude and is not the same as R_perp.
These loss settings do not change the model format or predicted magnetic force.
On restart, the objective is taken from
nep.in; the model/restart files do not store these loss coefficients.