-
Notifications
You must be signed in to change notification settings - Fork 604
Register quantized ops into quantization example #85
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
Conversation
This pull request was exported from Phabricator. Differential Revision: D48541611 |
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.
thanks! does it work for per channel weight quant as well? (changing is_per_channel to True instead of False in the quant example)
|
||
set -e | ||
|
||
test_buck2_quantization() { |
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.
This is probably a TODO to add a cmake build here later for completeness
This should work for all the existing quantized ops. |
Summary: Pull Request resolved: pytorch/executorch#85 Enables the ability to `export_to_pte`. Previously quantized models can't run `export_to_pte` because some quantized ops are missing out variants. Recently we added support for custom ops and register them into EXIR by loading shared library. This diff adds support for registering quantized ops out variants and add it into CI. Reviewed By: huydhn Differential Revision: D48541611 fbshipit-source-id: 113b9af4bfc3b47673453e6f898d2205f361a879
This pull request was exported from Phabricator. Differential Revision: D48541611 |
662a551
to
30dfc3b
Compare
This pull request has been merged in 0ece196. |
Summary:
Enables the ability to
export_to_pte
. Previously quantized models can't runexport_to_pte
because some quantized ops are missing out variants. Recently we added support for custom ops and register them into EXIR by loading shared library.This diff adds support for registering quantized ops out variants and add it into CI.
Differential Revision: D48541611