Function main

nlp_class2/recursive_theano.py:251–294

(is_binary=True)

def main(is_binary=True):
    train, test, word2idx = get_ptb_data()

    for t in train:
        add_idx_to_tree(t, 0)
    train = [tree2list(t, -1, is_binary) for t in train]
    if is_binary:
        train = [t for t in train if t[3][-1] >= 0] # for filtering binary labels

    # sanity check
    # check that last node has no parent
    # for t in train:
    #     assert(t[1][-1] == -1 and t[2][-1] == -1)

    for t in test:
        add_idx_to_tree(t, 0)
    test = [tree2list(t, -1, is_binary) for t in test]
    if is_binary:
        test = [t for t in test if t[3][-1] >= 0] # for filtering binary labels

    train = shuffle(train)
    # train = train[:2000]
    n_pos = sum(t[3][-1] for t in train)
    # print("num pos train:", n_pos)
    # idx2word = {v:k for k, v in word2idx.items()}
    # for i in range(4):
    #     words, _, _, labels = train[i]
    #     print_sentence(words, idx2word)
    #     print("label:", labels[-1])
    test = shuffle(test)
    test = test[:1000]

    V = len(word2idx)
    print("vocab size:", V)
    D = 10
    K = 2 if is_binary else 5

    model = RecursiveNN(V, D, K)
    model.fit(train, learning_rate=1e-2, reg=1e-2, mu=0, epochs=20, activation=T.tanh, train_inner_nodes=False)
    print("train accuracy:", model.score(train))
    print("test accuracy:", model.score(test))

    # make sure program doesn't end until we close the plot
    plt.show()


if __name__ == '__main__':