Explore Library
Code QuizIntermediate

One-Hot Vector Wrong Length

Identify the off-by-one sizing bug in one-hot encoding.

Codepython
def one_hot(token, vocab):
    vec = [0] * (len(vocab) - 1)
    idx = vocab.index(token)
    vec[idx] = 1
    return vec

vocab = ['cat', 'dog', 'fish']
print(one_hot('fish', vocab))

What is the bug in this one-hot encoding function?