Defining the model

Next, we define the model as (x_tensor × w) + b:

model = tf.matmul(x_tensor, w) + b