=================================================================================

With Eager execution in TensorFlow 2.0, we don’t need a tf.Session to run our code. However, with graph mode in TensorFlow 1.0, evaluation happens only after we’ve wrapped our code with tf.Session by following the steps:
          i) Create the tf.Graph object and set it to the default graph for the current scope.
          ii) Declare the computation part in TensorFlow: c=tf.matmul(m,n).
          iii) Define the variable sharing and scope, as required.
          iv) Create and configure the tf.Session to build the graph and connect to the tf.Session.
          v) Initialize all the variables in advance.
          vi) Use the tf.Session.run method to start the computation.
          vii) tf.Session.run triggers a procedure to compute the final output.

============================================