You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on May 11, 2023. It is now read-only.
The types returned are inconsistent over operations.
Currently adding LinearOperators to each other return and linear operator. Also adding a jax.numpy.ndarray to a linear operator returns a LinearOperator.
But matrix multiplication does not return a LinearOperator. It returns an jax.numpy.ndarray.
The question is: How should types be coerced?
Should we always return a linear operator? Should operations involving a jax.numpy.ndarray, instead, always return a jax.numpy.ndarray - since these are the least structure preserving.
The text was updated successfully, but these errors were encountered:
The types returned are inconsistent over operations.
Currently adding
LinearOperator
s to each other return and linear operator. Also adding ajax.numpy.ndarray
to a linear operator returns aLinearOperator
.But matrix multiplication does not return a
LinearOperator
. It returns anjax.numpy.ndarray
.The question is: How should types be coerced?
Should we always return a linear operator? Should operations involving a
jax.numpy.ndarray
, instead, always return ajax.numpy.ndarray
- since these are the least structure preserving.The text was updated successfully, but these errors were encountered: