with tf.Session() as s:
    lin_space = s.run(tf.linspace(-3.,9.,500))
    print('The shape of lin_space: ',lin_space.shape)
    print('The type of lin_space: ', type(lin_space))
    # this returns an array of shape (500,) and we specified that xx is (1,500) dimensional, so we need to reshape 
    lin_space = lin_space.reshape(1,500)
    cost_fcn = s.run(yy, feed_dict = {xx_placeholder : lin_space})

The shape of lin_space:  (500,)
The type of lin_space:  <class 'numpy.ndarray'>