Fixed the negative gradient modulus error.

It was a 1 extra conjugation (both transposition and conjugation cancel out)

Author: mesonepigreco

Modules/SchaMinimizer.py

@@ -416,9 +416,9 @@ def minimization_step(self, custom_function_gradient = None):
 
 
         # Store the gradient in the minimization
-        self.__gc__.append(np.real(np.einsum("abc,acb", dyn_grad, np.conj(dyn_grad))))
-        self.__gc_err__.append(np.real(np.einsum("abc, acb", err, np.conj(err))))
-            
+        self.__gc__.append( np.sqrt(np.sum( np.abs(dyn_grad)**2)))
+        self.__gc_err__.append( np.sqrt(np.sum( np.abs(err)**2)))
+
         # Perform the step for the dynamical matrix respecting the root representation
         new_dyn = PerformRootStep(np.array(self.dyn.dynmats, order = "C"), dyn_grad,
                                   self.min_step_dyn, root_representation = self.root_representation,