Multi tool use

How to multiply k 2x2 Matrices by k 2x2 matrices in tensorflow?

I have 2 3D tensors in tensorflow, where the two tensors have the shape Kx2x2. The tensors represent a set of 2x2 matrices. Is there a way to multiply the 2x2 matrices in the first tensor with the corresponding matrix in the second so that I get a Kx2x2 tensor in the end?

3 Answers
3

You can use tf.matmul.

tf.matmul

c = tf.matmul(a, b)


You can do tf.matmul(a,b).

According to the tensorflow documentation,

tf.matmul returns:
A Tensor of the same type as a and b where each inner-most matrix is the product of the corresponding matrices in a and b.

output[..., i, j] = sum_k (a[..., i, k] * b[..., k, j]), for all indices i, j.

Note: This is matrix product, not element-wise product.

https://www.tensorflow.org/api_docs/python/tf/matmul

tf.matmul is the correct operator for matrix multiplication. If the dimension of the tensor is over 2, the inner 2 specify the shape of the matrix. Hence the shape of two tensors must be [a_1, a₂, ..., a_n, x, y] and [a_1, a₂, ..., a_n, y, z], respectively ([K, 2, 2] in the OP's case).

tf.matmul

Sample Code

# Suppose X and Y are two tensors of the shape [K, 2, 2]
result = tf.matmul(X, Y)


By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.