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
I am trying to write an operator for computing Cholesky decomposition (and gradients).
It will roughly be based on the approach in TensorFlow (which is based on https://homepages.inf.ed.ac.uk/imurray2/pub/16choldiff/choldiff.pdf). I am half way through and
have created a PR tensorflow/tfjs-core#1492 but I think I need some help with a few details for implementing it. In particular
I would like to know
What's the best way to copy the values of a tensor? I have tried tensor.clone() but apparently it does not do deep copying (Clone() doesn't actually clone? #1032).
I would like to know how I can do in-place update to the content of the result tensor. The original code (in Numpy) does in place updates with a[:] = b but I don't know if there is a simple way to do this in tfjs. Right now I am just creating loops to iterate through the indices to do updates but is there a better way?
Thanks!
The text was updated successfully, but these errors were encountered:
Automatically closing due to lack of recent activity. Please update the issue when new information becomes available, and we will reopen the issue. Thanks!
Hi!
I am trying to write an operator for computing Cholesky decomposition (and gradients).
It will roughly be based on the approach in TensorFlow (which is based on https://homepages.inf.ed.ac.uk/imurray2/pub/16choldiff/choldiff.pdf). I am half way through and
have created a PR tensorflow/tfjs-core#1492 but I think I need some help with a few details for implementing it. In particular
I would like to know
a[:] = b
but I don't know if there is a simple way to do this in tfjs. Right now I am just creating loops to iterate through the indices to do updates but is there a better way?Thanks!
The text was updated successfully, but these errors were encountered: