PURE-melo / Spectral_Normalization-Tensorflow

Simple Tensorflow Implementation of "Spectral Normalization for Generative Adversarial Networks" (ICLR 2018)

Spectral_Normalization-Tensorflow

Simple Tensorflow Implementation of Spectral Normalization for Generative Adversarial Networks (ICLR 2018)

Usage

> python main.py --dataset mnist --sn True

Summary

Simple Code

def spectral_norm(w, iteration=1):
   w_shape = w.shape.as_list()
   w = tf.reshape(w, [-1, w_shape[-1]])

   u = tf.get_variable("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))

   with tf.control_dependencies([u.assign(u_hat)]):
       w_norm = w / sigma
       w_norm = tf.reshape(w_norm, w_shape)


   return w_norm

How to use

   w = tf.get_variable("kernel", shape=[kernel, kernel, x.get_shape()[-1], channels])
   b = tf.get_variable("bias", [channels], initializer=tf.constant_initializer(0.0))

   x = tf.nn.conv2d(input=x, filter=spectral_norm(w), strides=[1, stride, stride, 1]) + b

Related works

• Group Normalization-Tensorflow

Author

Junho Kim