Sess.run() And ".eval()" In Tensorflow Programming

In Tensorflow programming, can someone please tell what is the difference between '.eval()' and 'sess.run()'. What do each of them do and when to use them?

Solution 1:

A session object encapsulates the environment in which Tensor objects are evaluated.

If x is a tf.Tensor object, tf.Tensor.eval is shorthand for tf.Session.run, where sess is the current tf.get_default_session.

You can make session the default as below

x = tf.constant(5.0)
y = tf.constant(6.0)
z = x * y

with tf.Session() as sess:
  print(sess.run(z))   # 30.0print(z.eval())      # 30.0

The most important difference is you can use sess.run to fetch the values of many tensors in the same step as below

print(sess.run([x,y])) # [5.0, 6.0]print(sess.run(z))     # 30.0

Where as eval fetch single tensor value at a time as below

print(x.eval()) # 5.0print(z.eval()) # 3.0

TensorFlow computations define a computation graph that has no numerical value until evaluated as below

print(x) # Tensor("Const_1:0", shape=(), dtype=float32)

In Tensorflow 2.x (>= 2.0), You can use tf.compat.v1.Session() instead of tf.session()