cost_tf.py

def cost(logits, labels):
    """
    Computes the cost using the sigmoid cross entropy
    
    Arguments:
    logits -- vector containing z, output of the last linear unit (before the final sigmoid activation) 
    labels -- vector of labels y (1 or 0) 
    """
    z = tf.placeholder(tf.float32, name = 'z')
    y = tf.placeholder(tf.float32, name = 'y')
    
    # the loss function -
    cost = tf.nn.sigmoid_cross_entropy_with_logits(logits = z, labels = y)
    
    # running the session -
    with tf.Session() as sess:
        cost = sess.run(cost, feed_dict = {z : logits, y : labels})
        
    sess.close()
    return cost

logits = np.array([0.29,0.34,0.75,0.88])
cost = cost(logits, np.array([0,0,1,1]))
print(cost)