Loss score does not update in custom loss

Question

I have the below custom loss which during training does not udpate

def cluster_loss(z, centers):

 partial = tf.math.pow(tf.squeeze(tf.norm(z - centers, ord=\'euc