KPCA fails in SUOD model

I am trying to ensamble multiple KPCA models under the SUOD framework. Fitting a single model works as expected:

import numpy as np
from pyod.models.kpca import KPCA
from pyod.models.suod import SUOD

data = np.load('/14_glass.npz')

X, y = data['X'], data['y']

clf = KPCA()
clf.fit(X)

However, wrapping two or more models in SUOD results in an error:

model_list = [KPCA(gamma=.1), KPCA(gamma = .5)]

clf2 = SUOD(base_estimators=model_list, combination='average')
clf2.fit(X)

ValueError: Unknown kernel None