Skip to content

Fix JAX eigh #9

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.

Sign up for GitHub

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 2 commits into from
Jan 18, 2025
Merged

Fix JAX eigh #9

Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
edbf1d9
Fix JAX eigh
Jan 16, 2025
7df72d0
Making compatible with older JAX versions
Jan 18, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 7 additions & 2 deletions tensorcircuit/backends/jax_ops.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -151,11 +151,16 @@

@jax.custom_vjp
def adaware_eigh(A: Array) -> Array:
return jnp.linalg.eigh(A)
result = jnp.linalg.eigh(A)
e = result.eigenvalues
Copy link
Member

@refraction-ray refraction-ray Jan 18, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

One problem of the change proposed here is that it is not compatible with older versions of Jax, where the result is just a tuple?
What about e, v = jnp.linalg.eigh(A); return e, v, will this change be compatible with both tuple and namedtuple?

Copy link
Author

@abhinavd abhinavd Jan 18, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Ah, yes, that would work! Shall I update it?

v = result.eigenvectors
return e, v


def jaxeigh_fwd(A: Array) -> Array:
e, v = jnp.linalg.eigh(A)
result = jnp.linalg.eigh(A)
e = result.eigenvalues
v = result.eigenvectors

Check warning on line 163 in tensorcircuit/backends/jax_ops.py

View check run for this annotation

Codecov / codecov/patch

tensorcircuit/backends/jax_ops.py#L161-L163 

Added lines #L161 - L163 were not covered by tests
return (e, v), (A, e, v)


Expand Down
Loading