-
Notifications
You must be signed in to change notification settings - Fork 25k
Modify symmetric eigendecomposition derivative #23018
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
Conversation
The derivative of the symmetric eigendecomposition was previously a triangular matrix. Changelog: - Modify the derivative of symeig from a triangular matrix to a symmetric matrix. Test Plan: - Existing gradcheck and gradgradchecks are ported to test_autograd to verify that the change is correct. Input to symeig is symmetrized before passing
cc: @calincru |
I get what was done, @vishwakftw , but I think some folks might benefit from even a single sentence about why it was done. Could you clarify? |
@jeffreyksmithjr I have added a comment in Functions.cpp about this change. Hope it clarifies the reason for the change. |
Thanks for solving this, @vishwakftw! |
@pytorchbot rebase this please |
@pytorchbot rebase this please |
@pytorchbot rebase this please |
@pytorchbot rebase this please |
@fmassa this is ready for review. |
@pytorchbot rebase this please |
@pytorchbot rebase this please |
@fmassa is on vacation so I'm unsubscribing him from this issue and kicking this back into triage |
Thanks! |
@pytorchbot rebase this please |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@ezyang has imported this pull request. If you are a Facebook employee, you can view this diff on Phabricator.
The derivative of the symmetric eigendecomposition was previously a triangular matrix.
Changelog:
Test Plan: