Refactor arm quantization

pytorch / executorch

On-device AI across mobile, embedded and edge for PyTorch

https://pytorch.org/executorch/

Other

1.35k stars 218 forks source link

Refactor arm quantization #3616

Closed Erik-Lundell closed 1 week ago

Erik-Lundell commented 2 weeks ago

Move quantization files to own folder
Breakout annotators into one file/ annotator in new folder quantizer/quantization_annotation
Make QuantizationConfig not optional in annotators There is a None-check before calling them.
Move all Quantization config logic into one class
Remove unsupported qat functionality
Remove unsupported dynamic annotation
Remove unused code
Add docstrings

Change-Id: Ie802780bd0eb7f946985e09de13c822645a78e21

pytorch-bot[bot] commented 2 weeks ago

:link: Helpful Links

:test_tube: See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/3616

:page_facing_up: Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

:white_check_mark: No Failures

As of commit 6585650eabf2f77409c84161c8c72f6f49718358 with merge base 60c94e81bc0ea31f63977da57a47e2791b909e9d (): :green_heart: Looks good so far! There are no failures yet. :green_heart:

This comment was automatically generated by Dr. CI and updates every 15 minutes.

facebook-github-bot commented 1 week ago

@digantdesai has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

facebook-github-bot commented 1 week ago

@digantdesai merged this pull request in pytorch/executorch@455f5cc1f6b84f6eeac064c24c00a2149d8d7bf7.

Erik-Lundell commented 1 week ago

I am assuming no functional changes?

Yes, no functional changes.