sigmoid_tf.py

def sigmoid(z):
    x = tf.placeholder(tf.float32, name = 'x') # a placeholder tensor (variable)
    sigmoid = tf.sigmoid(x) # compting sigmoid function of "x"
    
    with tf.Session() as sess:
        result = sess.run(sigmoid, feed_dict = {x : z}) # runnig the computation graph (sigmoid) using a feed_dict
    return result

print(sigmoid(1))
print(sigmoid(10))
print(sigmoid(100))