What are the differences between casadi's dot product and numpy's dot product, and how can I calculate the dot product in Casadi?

Question

I am trying to write a conversion from a pytorch neural network to a casadi neural network.

I am getting an error where the dot product requires equal shapes, whereas numpy does not require that.

Example:

import numpy as np

A = np.array([[1, 2, 3]])
B = np.array([
    [1, 2, 3],
    [4, 5, 6],
    [7, 8, 9]
])
C = np.dot(A, B)  # array([[30, 36, 42]])

In casadi:

import casadi as ca

A = ca.MX(*A.shape)
B = ca.MX(*B.shape)
C = ca.dot(A, B)

Traceback (most recent call last):
  File "/home/tom/.local/lib/python3.10/site-packages/IPython/core/interactiveshell.py", line 3433, in run_code
    exec(code_obj, self.user_global_ns, self.user_ns)
  File "<ipython-input-72-07337ba0e18e>", line 1, in <module>
    C = ca.dot(A, B)
  File "/home/tom/miniconda3/envs/vpp3/lib/python3.10/site-packages/casadi/casadi.py", line 36361, in dot
    return _casadi.dot(*args)
RuntimeError: .../casadi/core/matrix_impl.hpp:2000: Assertion "x.size()==y.size()" failed:
dot: Dimension mismatch

How can I perform this dot product?

Numpy handles this mismatch by using [broadcasting](https://numpy.org/doc/stable/user/basics.broadcasting.html). I'd write an answer, but I don't know how to do the same thing in casadi. — Nick ODell, Aug 16 '23 at 16:03

What are the differences between casadi's dot product and numpy's dot product, and how can I calculate the dot product in Casadi?

0 Answers0