Add 16A8W support and test for add operation #13568
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Summary: This diff implements a 16A8W (16-bit activations, 8-bit weights) quantization configuration utility for the ExecutorTorch ARM backend, following the feedback from D79746479. ## Key Changes **1. New Quantization Configuration Function** - Add `get_symmetric_a16w8_quantization_config()` in `fbcode/executorch/backends/arm/quantizer/arm_quantizer.py` - Provides 16-bit activations with HistogramObserver (better precision than 8A8W) - Maintains 8-bit weights with MinMaxObserver/PerChannelMinMaxObserver (memory efficient) - **Technically supported by TOSA through [EXT-INT16 extension/profile](https://www.mlplatform.org/tosa/tosa_spec.html#_conv2d)** ## Benefits - **Better Precision**: 16-bit activations provide higher precision than 8-bit. Useful for carrying precision for recurring neural nets. Reviewed By: 3l1 Differential Revision: D79763381
Summary: - Adds linear ops test using the 16A8W config in INT16 profile. - Adds support in view ops validation for INT16 Dtype. - Validated with TOSA pipeline test. - Checked earlier marked flaky tests no longer flaky and remove markers. Note: Not verified with tosa reference model run. Differential Revision: D80308822
Summary: Add 16A8W quantization support and test for the add operation in ExecutorTorch ARM backend. This follows the pattern established for linear operations, extending int16 support to add operations. Changes: - Add INT16 dtype validation support in op_add.py - Add test_add_tensor_16a8w_tosa_INT test function - Enable test_add.py in test targets configuration The 16A8W configuration uses 16-bit activations with 8-bit weights, enabling higher precision for activations while maintaining weight efficiency. Differential Revision: D80510463
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/13568
Note: Links to docs will display an error until the docs builds have been completed. ❌ 2 New FailuresAs of commit 994b904 with merge base 624b38e ( NEW FAILURES - The following jobs have failed:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D80510463 |
This PR needs a
|
facebook-github-bot
pushed a commit
that referenced
this pull request
Aug 21, 2025
Summary: Add 16A8W quantization support and test for the add operation in ExecutorTorch ARM backend. This follows the pattern established for linear operations, extending int16 support to add operations. Changes: - Add INT16 dtype validation support in op_add.py - Add test_add_tensor_16a8w_tosa_INT test function - Enable test_add.py in test targets configuration The 16A8W configuration uses 16-bit activations with 8-bit weights, enabling higher precision for activations while maintaining weight efficiency. Differential Revision: D80510463
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary:
Add 16A8W quantization support and test for the add operation in ExecutorTorch ARM backend.
This follows the pattern established for linear operations, extending int16 support to add operations.
Changes:
The 16A8W configuration uses 16-bit activations with 8-bit weights, enabling higher precision for activations while maintaining weight efficiency.
Differential Revision: D80510463