Skip to content
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

Remove tiling in gauss_kl when tensorflow supports proper broadcasting #1321

Closed
st-- opened this issue Mar 13, 2020 · 0 comments · Fixed by #1585
Closed

Remove tiling in gauss_kl when tensorflow supports proper broadcasting #1321

st-- opened this issue Mar 13, 2020 · 0 comments · Fixed by #1585

Comments

@st--
Copy link
Member

st-- commented Mar 13, 2020

Currently, kullback_leiblers.gauss_kl() explicitly tiles as tf.linalg.triangular_solve does not broadcast (worse, it segfaults). Once TensorFlow fixes this issue (tensorflow/tensorflow#37584), we should remove this workaround from our code, as it is very inefficient (lots of unnecessary memory allocation).

st-- added a commit that referenced this issue Oct 5, 2020
@st-- st-- closed this as completed in #1585 Oct 6, 2020
st-- added a commit that referenced this issue Oct 6, 2020
…low versions (#1585)

We used to tile in gauss_kl() to work around TensorFlow's lack of broadcasting. Since TensorFlow 2.2, this is finally working, and this PR removes the tiling for TensorFlow>=2.2. Resolves #1321.

Co-authored-by: ST John <st@secondmind.ai>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
GPflow Features
  
Awaiting triage
Development

Successfully merging a pull request may close this issue.

1 participant