how to fit linear mixed model with a fixed variance component

[JJ-Zhang-DS]

Describe the bug

I want to build a linear mixed model with fixed var components.
variance components = 5 and 10 (experimental error and sampling error, respectively)

But looks like the var components can not be passed to the model. How to fix it? thank you.

UserWarning: Argument params not used by MixedLM.fit
warnings.warn("Argument %s not used by MixedLM.fit" % x)

Code Sample, a copy-pastable example if possible

formula = 'y ~ trt'

# Fit the model
model = smf.mixedlm(formula, data, groups=data['rep'])

# Fix the variance components
vcov = {'rep: Intercept': 10, 'Residual': 5}

# Create a MixedLMParams object with fixed variance components
params = sm.regression.mixed_linear_model.MixedLMParams.from_components(vcomp=vcov)

# Fit the model with fixed variance components
model_fix = model.fit(params=params, method='lbfgs')

# Print a summary of the model results
print(model_fix.summary())

[kshedden]

This might be possible using the free argument to the fit method.