Explore Library
Code QuizIntermediate

Handling Out-of-Vocabulary Tokens

Find why this encoder crashes on unknown words.

Codepython
vocab = {"<UNK>": 0, "hello": 1, "world": 2}

def encode(token, vocab):
    return vocab[token]

print(encode("hello", vocab))
print(encode("foobar", vocab))

What is the bug when encoding an out-of-vocabulary token?