Skip to content

Add specification for computing the matrix transpose (matrix_transpose) #248

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Sep 13, 2021

Conversation

kgryte
Copy link
Contributor

@kgryte kgryte commented Aug 30, 2021

This PR:

  • adds a specification for computing the transpose of a matrix (or a stack of matrices).

Background

The addition of a matrix_transpose API is meant to replace the use of the transpose API. The transpose API was found to be problematic due to its API supporting more general axes permutation (see gh-228).

Similar to other linear algebra APIs, the proposed matrix_transpose API operates on stack of matrices and thus more succinctly and explicitly allows expressing matrix operations in accordance with mathematical intuition and obviates the need for doing transpose(x, axes=[-2,-1]) whenever wanting to perform a batched matrix transpose. Batching is the default behavior.

The matrix_transpose API serves as a functional complement to the proposed .mT attribute (see gh-246).

@kgryte
Copy link
Contributor Author

kgryte commented Sep 13, 2021

Adding the matrix_transpose was discussed in the consortium meeting held 09-02-2021. No objections were raised, as this PR provides the desired behavior of batched matrix transpose.

Merging...

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant