The ungraded Data Labeling lab has an error or typo:
# Allow unigrams and bigrams
vectorizer = CountVectorizer(ngram_range=(1, 5))
If the comment is correct (i.e. “unigrams and bigrams” only), then the argument passed to CountVectorizer() should be ngram_range(1, 2), not ngram_range(1, 5).
Alternatively, if the argument is correct, then the comment should read:
# Allow unigrams, bigrams, trigrams, 4-grams and 5-grams