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
Add a tridiagonal eigh solver. #6622
Conversation
112ecc6
to
41b2bff
Compare
jax/_src/scipy/linalg.py
Outdated
beta_abs = jnp.abs(beta) | ||
beta_sq = beta * beta | ||
|
||
# Compute an interval containing the eigenvalue of T using the Gerschgorin |
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.
# Compute an interval containing the eigenvalue of T using the Gerschgorin | |
# Compute an interval containing the eigenvalue of T using the Gershgorin |
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.
https://en.wikipedia.org/wiki/Semyon_Aranovich_Gershgorin says this is an accepted transliteration!
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.
Okay you win, though the spelling is inconsistent with a comment a few lines down. (Maybe that's good, as all possible transliterations are represented.)
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.
(Since GitHub already marked me as a reviewer, given my drive-by comment...)
Beautiful code as always!
To be clear, Rasmus gets all the credit for this code. I just ported it, which happily at the level of NumPy-like ops is almost completely mechanical to do. (The autodiff stuff, when we have the pieces for it, will be a bit different.) |
Copying Rasmus code is a recipe for success in any endeavor! :D |
This code was written by Rasmus for TensorFlow; this PR ports it to JAX.