Tensorflow tf.constant
v = tf.Variable([0.0])with tf.GradientTape() as g: loss = tf.constant(v + v)g.gradient(loss, v).numpy()array([2.], dtype=float32)
Mysterious Monkey