Received AttributeError: 'numpy.ndarray' object has no attribute 'lower' with permutation_importance

Question

I want to use permutation_importance to calculate feature importance. From the doc, I understood that X_train needs to be array, and y_train needs to be array-like. However,