Tensor-Vector Inner Product using Einsum

This is an example of how to properly use the einsum function to compute a tensor-vector inner product.

from csdl_om import Simulatorimport numpy as npfrom csdl import Modelimport csdl

class ExampleInnerTensorVector(Model):
    def define(self):        a = np.arange(4)        vec = self.declare_variable('a', val=a)
        # Shape of Tensor        shape3 = (2, 4, 3)        c = np.arange(24).reshape(shape3)
        # Declaring tensor        tens = self.declare_variable('c', val=c)
        # Inner Product of a tensor and a vector        self.register_output(            'einsum_inner2',            csdl.einsum_new_api(                tens,                vec,                operation=[('rows', 0, 1), (0, ), ('rows', 1)],            ))

sim = Simulator(ExampleInnerTensorVector())sim.run()
print('a', sim['a'].shape)print(sim['a'])print('c', sim['c'].shape)print(sim['c'])print('einsum_inner2', sim['einsum_inner2'].shape)print(sim['einsum_inner2'])

[0. 1. 2. 3.]c (2, 4, 3)[[[ 0.  1.  2.]  [ 3.  4.  5.]  [ 6.  7.  8.]  [ 9. 10. 11.]]
 [[12. 13. 14.]  [15. 16. 17.]  [18. 19. 20.]  [21. 22. 23.]]]einsum_inner2 (2, 3)[[ 42.  48.  54.] [114. 120. 126.]]