Function tokens_to_vector

unsupervised_class/books.py:68–73

(tokens)

# now let's create our input matrices - just indicator variables for this example - works better than proportions
def tokens_to_vector(tokens):
    x = np.zeros(len(word_index_map))
    for t in tokens:
        i = word_index_map[t]
        x[i] += 1
    return x

N = len(all_tokens)
D = len(word_index_map)