OrthoIV example in Docs throws error

When I try to run the OrthoIV example from the docs (https://github.com/microsoft/EconML), I am able to fit the model and obtain results using the effect method. But when I try to compute effect confidence intervals using the effect_inference method, it throws and error saying Inference = None, even though the model has already been fit?

Am I missing something? How do I resolve this issue?

AttributeError Traceback (most recent call last)
in
----> 1 lb, ub = est.effect_inference(X_test.iloc[:, 1:-2]) # OLS confidence intervals

~/pycx_kernel/.deps/econml-0.7.0-py3-none-any.whl/econml/_ortho_learner.py in effect_inference(self, X, T0, T1)
603 def effect_inference(self, X=None, *, T0=0, T1=1):
604 self._check_fitted_dims(X)
--> 605 return super().effect_inference(X, T0=T0, T1=T1)
606 effect_inference.doc = LinearCateEstimator.effect_inference.doc
607

~/pycx_kernel/.deps/econml-0.7.0-py3-none-any.whl/econml/cate_estimator.py in call(self, *args, **kwargs)
170 return getattr(self._inference, name)(*args, **kwargs)
171 else:
--> 172 raise AttributeError("Can't call '%s' because 'inference' is None" % name)
173 return call
174

AttributeError: Can't call 'effect_inference' because 'inference' is None

Open the OrthoIV example in the linked EconML repository and reproduce the documented fit, effect, and effect_inference calls. Start by reading the shown effect_inference path and how the estimator is configured in the example. Done means the example either computes confidence intervals successfully or documents the required configuration and expected behavior.