def spectral_normed_weight()

in utils.py [0:0]

• 34 lines of code
• 5 McCabe index (conditional complexity)

def spectral_normed_weight(w, name, lower_bound=False, iteration=1, fc=False):
    if fc:
        iteration = 2

    w_shape = w.shape.as_list()
    w = tf.reshape(w, [-1, w_shape[-1]])

    iteration = FLAGS.spec_iter
    sigma_new = FLAGS.spec_norm_val

    u = tf.get_variable(name + "_u",
                        [1,
                         w_shape[-1]],
                        initializer=tf.random_normal_initializer(),
                        trainable=False)

    u_hat = u
    v_hat = None
    for i in range(iteration):
        """
        power iteration
        Usually iteration = 1 will be enough
        """
        v_ = tf.matmul(u_hat, tf.transpose(w))
        v_hat = tf.nn.l2_normalize(v_)

        u_ = tf.matmul(v_hat, w)
        u_hat = tf.nn.l2_normalize(u_)

    u_hat = tf.stop_gradient(u_hat)
    v_hat = tf.stop_gradient(v_hat)

    sigma = tf.matmul(tf.matmul(v_hat, w), tf.transpose(u_hat))

    if FLAGS.spec_eval:
        dep = []
    else:
        dep = [u.assign(u_hat)]

    with tf.control_dependencies(dep):
        if lower_bound:
            sigma = sigma + 1e-6
            w_norm = w / sigma * tf.minimum(sigma, 1) * sigma_new
        else:
            w_norm = w / sigma * sigma_new

        w_norm = tf.reshape(w_norm, w_shape)

    return w_norm