LLM extension: add ethosu 8w16a and quantize scope plumbing#19876
Open
xingguo01 wants to merge 1 commit into
Open
LLM extension: add ethosu 8w16a and quantize scope plumbing#19876xingguo01 wants to merge 1 commit into
xingguo01 wants to merge 1 commit into
Conversation
- adds the `ethosu_8w16a` PT2E quantization mode - introduces shared `quantization.quantize_scope` handling for Arm backends - wires the Arm quantize scope through the LLM export path - passes Ethos-U system config and memory mode through the partitioner setup Signed-off-by: Xingguo Li <xingguo.li@arm.com> Change-Id: I3f8446a20fb63670520a6b35484669d8df6f31bf
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/19876
Note: Links to docs will display an error until the docs builds have been completed. ❌ 1 New Failure, 1 Cancelled Job, 4 Unrelated FailuresAs of commit e079d39 with merge base 88faab2 ( NEW FAILURE - The following job has failed:
CANCELLED JOB - The following job was cancelled. Please retry:
FLAKY - The following job failed but was likely due to flakiness present on trunk:
BROKEN TRUNK - The following jobs failed but was 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. |
meta-cla
Bot
added
the
CLA Signed
xingguo01
added
partner: arm
Contributor
There was a problem hiding this comment.
Pull request overview
This PR extends the LLM export PT2E quantization plumbing for Arm backends by adding a new Ethos-U quantization mode and introducing shared “quantize scope” handling across TOSA/Ethos-U/VGF, along with passing Ethos-U compiler flags through the export/partitioner setup.
Changes:
- Added PT2E quantization mode
ethosu_16a8wand wired it through the LLM export path. - Introduced shared Arm quantization-scope application logic (
fullvslinear) for TOSA/Ethos-U/VGF quantizers. - Plumbed Ethos-U
extra_flagsthrough the partitioner and quantizer setup.
Reviewed changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.
| File | Description |
|---|---|
| extension/llm/export/quantizer_lib.py | Adds ethosu_16a8w, adds Arm quantize-scope helper, and plumbs scope/flags into Arm quantizers. |
| extension/llm/export/partitioner_lib.py | Passes Ethos-U extra_flags through to EthosUCompileSpec. |
| extension/llm/export/config/llm_config.py | Adds ethosu_16a8w, introduces shared QuantizeScope, and adds Ethos-U extra_flags config. |
| examples/models/llama/export_llama_lib.py | Exposes ethosu_16a8w via CLI and wires Arm quantize-scope + Ethos-U flags into quantizer/partitioner creation. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+366
to
+371
| compile_spec = EthosUCompileSpec( | ||
| target, | ||
| system_config, | ||
| memory_mode, | ||
| extra_flags=extra_flags, | ||
| ) |
Comment on lines
231
to
236
| "vulkan_8w", | ||
| "tosa_8a8w", | ||
| "ethosu_8a8w", | ||
| "ethosu_16a8w", | ||
| "vgf_8a8w", | ||
| "vgf_16a8w", |
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
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.
ethosu_8w16aPT2E quantization modequantization.quantize_scopehandling for Arm backendscc @digantdesai @freddan80 @per @zingo @oscarandersson8218 @mansnils @Sebastian-Larsson @robell @rascani