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

cpu: aarch64: non-global stats mode for ACL lnorm #2215

Open
wants to merge 1 commit into
base: main
Choose a base branch
from

Conversation

michalowski-arm
Copy link
Contributor

  • Change acl_layer_normalization to support non-global stats only
  • Remove mean and variance check for ACL lnorm from benchdnn

@michalowski-arm michalowski-arm requested review from a team as code owners November 14, 2024 15:57
  * Change acl_layer_normalization to support non-global stats only --
    ACL does not support global stats
  * Remove mean and variance check for ACL lnorm from benchdnn --
    ACL does not support returning them
"ACL only supports forward training with lnorm if stats "
"are provided (use global stats)");
ACL_CHECK_SUPPORT(use_global_stats(),
"ACL does not support global stats with lnorm");
Copy link
Contributor

Choose a reason for hiding this comment

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

Does it mean that that ACL would compute mean and variance internally but wouldn't return them to the user?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

That's correct

Copy link
Contributor

Choose a reason for hiding this comment

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

Any chance ACL will update their functionality to start returning mean and var to be aligned with oneDNN?

Copy link
Contributor

Choose a reason for hiding this comment

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

As an additional note, it seems that supporting only prop_kind::inference (and dropping prop_kind::training from supported) is a cleaner way to indicate that mean and variance are not going to be returned to the user.
@mgouicem, would you agree with this statement?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants