We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
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
Follow up to #2
we should probably store C_S = cholesky(W.S)
C_S = cholesky(W.S)
C_B for
C_B
B = C_S.U' \ (W.A * cholesky(W.D).U') C_B = cholesky(Symmetric(I + B'B))
as well as the original matrix. As these tend to be needed for many operations, and don't need to be recomputed as long as the matrix isn't changed.
One question is making sure this works OK with AD.
The text was updated successfully, but these errors were encountered:
No branches or pull requests
Follow up to #2
we should probably store
C_S = cholesky(W.S)
C_B
foras well as the original matrix.
As these tend to be needed for many operations, and don't need to be recomputed as long as the matrix isn't changed.
One question is making sure this works OK with AD.
The text was updated successfully, but these errors were encountered: