Max Entropy Result for value = 7 yields 5.5 in calc_mean_score

[lbundalian]

in the code:
def get_max_entropy_distribution(mean):
SAMPLESPACE = np.arange(10)
features = get_features()

model = MinDivergenceModel(features, samplespace=SAMPLESPACE, algorithm='CG')

# set the desired feature expectations and fit the model
X = np.array([[mean]])
model.fit(X)

return model.probdist()

the resulting entropy is an array consisting of 0.9999998 or something, when i tried to revert it using calc_mean_score, it yields 5.5