Fitting model in hmmlearn

Question

I'm not sure I get how hmmlearn expects the input data. Below is a modified example from the tutorial.

import numpy as np
from hmmlearn import hmm

states = ['up', 'down']
start_probs = np.array([0.6, 0.4])

vocabulary = [0, 1, 2, 3]

emission_probs = np.array(
    [
        [0.25, 0.1, 0.4, 0.25],
        [0.2, 0.5, 0.1, 0.2],
    ]
)

trans_mat = np.array(
    [
        [0.8, 0.2], 
        [0.2, 0.8],
    ]
)

observations = np.random.choice(vocabulary, (10, 7))
lengths = [len(item) for item in observations]

# Set up model:
model = hmm.MultinomialHMM(
    n_components=len(states),
    n_trials=len(observations[0]),
    n_iter=50,
    init_params='',
)

model.n_features = len(vocabulary)
model.startprob_ = start_probs
model.transmat_ = trans_mat
model.emissionprob_ = emission_probs

model.fit(observations, lengths)
logprob, received = model.decode(observations)

At this point I'm getting this error:

 in _AbstractHMM._check_and_set_n_features(self, X)
    525 if hasattr(self, "n_features"):
    526     if self.n_features != n_features:
--> 527         raise ValueError(
    528             f"Unexpected number of dimensions, got {n_features} but "
    529             f"expected {self.n_features}")
    530 else:
    531     self.n_features = n_features
    ValueError: Unexpected number of dimensions, got 7 but expected 4

I'm not sure why it is expecting 4, which is just the size of the vocabulary. Isn't lengths supposed to be the size of the sequences?