BUG: a failing modeling test on windows #18333

__________________ test_models_fitting[LMLSQFitter-model33] ___________________

model = {'bounding_box': False, 'class': <class 'astropy.modeling.functional_models.Sersic2D'>
Name: Sersic2D
N_inputs: 2
N_ou...y / sr>)], 'parameters': {'amplitude': <Quantity 3. MJy / sr>, 'ellip': 0, 'n': 4, 'r_eff': <Quantity 2. arcsec>, ...}}
fitter = <astropy.modeling.fitting.LMLSQFitter object at 0x000001A9E092FBC0>

    @pytest.mark.skipif(not HAS_SCIPY, reason="requires scipy")
    @pytest.mark.filterwarnings(r"ignore:.*:RuntimeWarning")
    @pytest.mark.filterwarnings(r"ignore:Model is linear in parameters.*")
    @pytest.mark.filterwarnings(r"ignore:The fit may be unsuccessful.*")
    @pytest.mark.filterwarnings(r"ignore:humlicek2 has been deprecated since .*")
    @pytest.mark.parametrize("model", MODELS)
    @pytest.mark.parametrize("fitter", fitters)
    def test_models_fitting(model, fitter):
        fitter = fitter()
    
        bad_voigt = model["class"] == Voigt1D and ("method" not in model["parameters"])
        if (
            (
                isinstance(fitter, LevMarLSQFitter)
                and model["class"] in NON_FINITE_LevMar_MODELS
            )
            or (
                isinstance(fitter, TRFLSQFitter)
                and (model["class"] in NON_FINITE_TRF_MODELS or bad_voigt)
            )
            or (
                isinstance(fitter, LMLSQFitter)
                and (model["class"] in NON_FINITE_LM_MODELS or bad_voigt)
            )
            or (
                isinstance(fitter, DogBoxLSQFitter)
                and model["class"] in NON_FINITE_DogBox_MODELS
            )
        ):
            return
    
        m = model["class"](**model["parameters"])
    
        if m.has_bounds and isinstance(fitter, LMLSQFitter):
            pytest.skip("The LMLSQFitter fitter does not support models with bounds")
    
        if len(model["evaluation"][0]) == 2:
            x = np.linspace(1, 3, 100) * model["evaluation"][0][0].unit
            y = np.exp(-(x.value**2)) * model["evaluation"][0][1].unit
            args = [x, y]
        else:
            x = np.linspace(1, 3, 100) * model["evaluation"][0][0].unit
            y = np.linspace(1, 3, 100) * model["evaluation"][0][1].unit
            z = np.exp(-(x.value**2 + y.value**2)) * model["evaluation"][0][2].unit
            args = [x, y, z]
    
        # Test that the model fits even if it has units on parameters
>       m_new = fitter(m, *args)
                ^^^^^^^^^^^^^^^^

..\..\.tox\py312-test-alldeps\Lib\site-packages\astropy\modeling\tests\test_models_quantities.py:837: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
..\..\.tox\py312-test-alldeps\Lib\site-packages\astropy\modeling\fitting.py:275: in wrapper
    model_new = func(self, model, xdata, ydata, zdata, **kwargs)
                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
..\..\.tox\py312-test-alldeps\Lib\site-packages\astropy\modeling\fitting.py:1730: in __call__
    return super().__call__(
..\..\.tox\py312-test-alldeps\Lib\site-packages\astropy\modeling\fitting.py:288: in wrapper
    return func(self, model, x, y, z=z, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
..\..\.tox\py312-test-alldeps\Lib\site-packages\astropy\modeling\fitting.py:1429: in __call__
    init_values, fitparams, cov_x = self._run_fitter(
..\..\.tox\py312-test-alldeps\Lib\site-packages\astropy\modeling\fitting.py:1603: in _run_fitter
    self.fit_info = optimize.least_squares(
..\..\.tox\py312-test-alldeps\Lib\site-packages\scipy\_lib\_util.py:689: in inner
    return func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^
..\..\.tox\py312-test-alldeps\Lib\site-packages\scipy\optimize\_lsq\least_squares.py:1015: in least_squares
    result = call_minpack(vector_fun.fun, x0, vector_fun.jac, ftol, xtol, gtol,
..\..\.tox\py312-test-alldeps\Lib\site-packages\scipy\optimize\_lsq\least_squares.py:75: in call_minpack
    x, info, status = _minpack._lmder(
..\..\.tox\py312-test-alldeps\Lib\site-packages\scipy\optimize\_differentiable_functions.py:734: in fun
    self._update_fun()
..\..\.tox\py312-test-alldeps\Lib\site-packages\scipy\optimize\_differentiable_functions.py:698: in _update_fun
    self.f = self.fun_wrapped(xp_copy(self.x))
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
..\..\.tox\py312-test-alldeps\Lib\site-packages\scipy\optimize\_differentiable_functions.py:415: in __call__
    return np.atleast_1d(self.fun(x))
                         ^^^^^^^^^^^
..\..\.tox\py312-test-alldeps\Lib\site-packages\scipy\optimize\_lsq\least_squares.py:263: in __call__
    return self.f(x, *self.args, **self.kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

self = <astropy.modeling.fitting.LMLSQFitter object at 0x000001A9E092FBC0>
fps = array([   21.14561837,   -19.94026772,  -188.2192972 ,   -45.48335025,
          11.87787074,    21.69193093, -8425.06283925])
fit_param_indices = {0, 1, 2, 3, 4, 5, ...}
args = (<Sersic2D(amplitude=3., r_eff=2., n=4., x_0=1., y_0=2., ellip=0., theta=0.)>, None, array([1.        , 1.02020202, 1.....33342676e-08, 5.01481171e-08, 3.96425452e-08,
       3.12866777e-08, 2.46517859e-08, 1.93922565e-08, 1.52299797e-08]))
model = <Sersic2D(amplitude=3., r_eff=2., n=4., x_0=1., y_0=2., ellip=0., theta=0.)>
weights = None

    def objective_function(self, fps, *args, fit_param_indices=None):
        """
        Function to minimize.
    
        Parameters
        ----------
        fps : list
            parameters returned by the fitter
        args : list
            [model, [weights], [input coordinates]]
        fit_param_indices : list, optional
            The ``fit_param_indices`` as returned by ``model_to_fit_params``.
            This is a list of the parameter indices being fit, so excluding any
            tied or fixed parameters.  This can be passed in to the objective
            function to prevent it having to be computed on every call.
            This must be optional as not all fitters support passing kwargs to
            the objective function.
        """
        model = args[0]
        weights = args[1]
        inputs = args[2:-1]
        meas = args[-1]
    
        fps = fitter_to_model_params_array(
            model,
            fps,
            self._use_min_max_bounds,
            fit_param_indices=fit_param_indices,
        )
    
        if weights is None:
            value = np.ravel(model.evaluate(*inputs, *fps) - meas)
        else:
            value = np.ravel(weights * (model.evaluate(*inputs, *fps) - meas))
    
        if not np.all(np.isfinite(value)):
>           raise NonFiniteValueError(
                "Objective function has encountered a non-finite value, "
                "this will cause the fit to fail!\n"
                "This can be caused by non-finite values in the input data "
                "or weights, which can be removed with fit(..., "
                "filter_non_finite=True), or by diverging model parameters "
                "that yield non-finite model values."
            )
E           astropy.modeling.fitting.NonFiniteValueError: Objective function has encountered a non-finite value, this will cause the fit to fail!
E           This can be caused by non-finite values in the input data or weights, which can be removed with fit(..., filter_non_finite=True), or by diverging model parameters that yield non-finite model values.

..\..\.tox\py312-test-alldeps\Lib\site-packages\astropy\modeling\fitting.py:1177: NonFiniteValueError