-
Notifications
You must be signed in to change notification settings - Fork 231
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
Support compute_uinit in coreml backend when calling ct.convert #4059
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/4059
Note: Links to docs will display an error until the docs builds have been completed. ✅ You can merge normally! (1 Unrelated Failure)As of commit d088e42 with merge base f538eae ( BROKEN TRUNK - The following job failed but were present on the merge base:👉 Rebase onto the `viable/strict` branch to avoid these failures
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
This pull request was exported from Phabricator. Differential Revision: D58981934 |
I think
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It is still good to expose compute unit control in general cases, though, if you find it necessary to use not only the default compute unit ALL
…rch#4059) Summary: Pull Request resolved: pytorch#4059 It seems currently the compute unit value set in https://www.internalfb.com/code/fbsource/fbcode/executorch/examples/models/llama2/lib/partitioner_lib.py?lines=74 is not used during the convert, adding the support in this diff. However to my surprise, the model name in the actual ios instrument result has the right suffix (see the image below), I couldn't find anywhere in the code how this is set and I have no way to validate if it's just part of the name or it's indeed set correctly through another path. I'll abandon this diff if it's not needed actually... {F1718372833} Reviewed By: cccclai Differential Revision: D58981934
This pull request was exported from Phabricator. Differential Revision: D58981934 |
c2110f7
to
d088e42
Compare
This pull request has been merged in 85cf584. |
Summary:
It seems currently the compute unit value set in executorch/examples/models/llama2/lib/partitioner_lib.py?lines=74 is not used during the convert, adding the support in this change.
However to my surprise, the model name in the actual ios instrument result has the right suffix (see the image below), I couldn't find anywhere in the code how this is set and I have no way to validate if it's just part of the name or it's indeed set correctly through another path. I'll abandon this diff if it's not needed actually...
{F1718372833}
Differential Revision: D58981934