Skip to content

Migrate ExecuTorch's use of pt2e from torch.ao to torchao #10294

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.

Sign up for GitHub

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

Jump to bottom
Closed
wants to merge 1 commit into from
Closed

Migrate ExecuTorch's use of pt2e from torch.ao to torchao #10294

wants to merge 1 commit into from

Conversation

metascroy
Copy link
Contributor

Most code related to PT2E quantization is migrating from torch.ao.quantization to torchao.quantization.pt2e.

torchao.quantization.pt2e contains an exact copy of PT2E code in torch.ao.quantization.

The torchao pin in ExecuTorch has already been bumped pick up these changes.

@pytorch-bot PyTorch Bot
Copy link

pytorch-bot bot commented Apr 18, 2025 •
edited
Loading

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/10294

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

❌ 1 New Failure, 1 Unrelated Failure

As of commit 5f9da89 with merge base 9aaea31 (image):

NEW FAILURE - The following job has failed:

FLAKY - The following job failed but was likely due to flakiness present on trunk:

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

@facebook-github-bot facebook-github-bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Apr 18, 2025
@github-actions GitHub Actions
Copy link

This PR needs a release notes: label

If your changes are user facing and intended to be a part of release notes, please use a label starting with release notes:.

If not, please add the topic: not user facing label.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "topic: not user facing"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.

jerryzh168
jerryzh168 reviewed Apr 18, 2025
View reviewed changes
@metascroy metascroy changed the title Migrate pt2e from torch.ao to torchao [not yet for review] migrate pt2e from torch.ao to torchao Apr 18, 2025
digantdesai
digantdesai reviewed Apr 21, 2025
View reviewed changes
backends/arm/quantizer/arm_quantizer.py Outdated
@@ -34,25 +34,25 @@
is_ethosu,
) # usort: skip
from executorch.exir.backend.compile_spec_schema import CompileSpec
from torch.ao.quantization.fake_quantize import (
from torch.fx import GraphModule, Node
from torchao.quantization.pt2e import _ObserverOrFakeQuantizeConstructor
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Correct me if I am wrong but torchao isn't a mandetory dep today but now it is?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

How do we define mandatory dependencies? It is installed by the install_requirements script?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Seems like we pull in source -

executorch/.gitmodules

Line 60 in 647e1f1

url = https://github.com/pytorch/ao.git

So this submodule is already updated since the tests are passing here.

check (1) if we run tests on et wheels with something quant, (2) if we do are they passing for this diff.

backends/xnnpack/quantizer/xnnpack_quantizer.py
@@ -16,25 +16,25 @@
propagate_annotation,
QuantizationConfig,
)
from torch.ao.quantization.fake_quantize import (
from torchao.quantization.pt2e.fake_quantize import (
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nit: there are some affine quantization stuff which assumes torchao isn't always available, can we fix that with this? here is one example.

cc @mcr229

@cccclai
Copy link
Contributor

cccclai commented Apr 21, 2025

Adding partners for visibility

jerryzh168
jerryzh168 reviewed May 13, 2025
View reviewed changes
backends/example/example_quantizer.py Outdated
from torch.ao.quantization.pt2e.graph_utils import find_sequential_partitions
from torch.ao.quantization.quantizer import QuantizationSpec, Quantizer
from torchao.quantization.pt2e import find_sequential_partitions
from torchao.quantization.pt2e.observer import HistogramObserver, MinMaxObserver
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

we can remove observer here

jerryzh168
jerryzh168 reviewed May 13, 2025
View reviewed changes
jerryzh168
jerryzh168 reviewed May 13, 2025
View reviewed changes
backends/mediatek/quantizer/qconfig.py
Comment on lines +13 to +14
from torchao.quantization.pt2e.fake_quantize import FakeQuantize
from torchao.quantization.pt2e.observer import MinMaxObserver, PerChannelMinMaxObserver
Copy link
Contributor

@jerryzh168 jerryzh168 May 13, 2025 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

can remove observer and fake_quantize from path

@cccclai
Copy link
Contributor

cccclai commented May 13, 2025

@cccclai can you have a look at the qualcomm changes?

Looks reasonable to me. Let's just use trunk to trigger more CI jobs.

@metascroy
Copy link
Contributor Author

@cccclai can you have a look at the qualcomm changes?

Looks reasonable to me. Let's just use trunk to trigger more CI jobs.

trunk is already triggered. Thanks!

@facebook-github-bot
Copy link
Contributor

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

@facebook-github-bot
Copy link
Contributor

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

SS-JIA
SS-JIA approved these changes May 15, 2025
View reviewed changes
Copy link
Contributor

@SS-JIA SS-JIA left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Vulkan changes LGTM!

@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D74694311

metascroy added a commit that referenced this pull request May 16, 2025
@metascroy @facebook-github-bot
Migrate ExecuTorch's use of pt2e from torch.ao to torchao (#10294)
39665db 
Summary:
Most code related to PT2E quantization is migrating from torch.ao.quantization to torchao.quantization.pt2e.

torchao.quantization.pt2e contains an exact copy of PT2E code in torch.ao.quantization.

The torchao pin in ExecuTorch has already been bumped pick up these changes.

Pull Request resolved: #10294

Reviewed By: SS-JIA

Differential Revision: D74694311

Pulled By: metascroy
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D74694311

metascroy added a commit that referenced this pull request May 17, 2025
@metascroy @facebook-github-bot
Migrate ExecuTorch's use of pt2e from torch.ao to torchao (#10294)
2268006 
Summary:
Most code related to PT2E quantization is migrating from torch.ao.quantization to torchao.quantization.pt2e.

torchao.quantization.pt2e contains an exact copy of PT2E code in torch.ao.quantization.

The torchao pin in ExecuTorch has already been bumped pick up these changes.

Pull Request resolved: #10294

Reviewed By: SS-JIA

Differential Revision: D74694311

Pulled By: metascroy
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D74694311

metascroy added a commit that referenced this pull request May 17, 2025
@metascroy @facebook-github-bot
Migrate ExecuTorch's use of pt2e from torch.ao to torchao (#10294)
1727bde 
Summary:
Most code related to PT2E quantization is migrating from torch.ao.quantization to torchao.quantization.pt2e.

torchao.quantization.pt2e contains an exact copy of PT2E code in torch.ao.quantization.

The torchao pin in ExecuTorch has already been bumped pick up these changes.

Pull Request resolved: #10294

Reviewed By: SS-JIA

Differential Revision: D74694311

Pulled By: metascroy
@facebook-github-bot
Copy link
Contributor

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

@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D74694311

metascroy added a commit that referenced this pull request May 17, 2025
@metascroy @facebook-github-bot
Migrate ExecuTorch's use of pt2e from torch.ao to torchao (#10294)
b252af9 
Summary:
Most code related to PT2E quantization is migrating from torch.ao.quantization to torchao.quantization.pt2e.

torchao.quantization.pt2e contains an exact copy of PT2E code in torch.ao.quantization.

The torchao pin in ExecuTorch has already been bumped pick up these changes.

Pull Request resolved: #10294

Reviewed By: SS-JIA

Differential Revision: D74694311

Pulled By: metascroy
@metascroy @facebook-github-bot
Migrate ExecuTorch's use of pt2e from torch.ao to torchao (#10294)
5f9da89 
Summary:
Most code related to PT2E quantization is migrating from torch.ao.quantization to torchao.quantization.pt2e.

torchao.quantization.pt2e contains an exact copy of PT2E code in torch.ao.quantization.

The torchao pin in ExecuTorch has already been bumped pick up these changes.

Pull Request resolved: #10294

Reviewed By: SS-JIA

Differential Revision: D74694311

Pulled By: metascroy
@facebook-github-bot
Copy link
Contributor

This pull request was exported from Phabricator. Differential Revision: D74694311

@robert-kalmar robert-kalmar mentioned this pull request May 23, 2025
Draft
@metascroy metascroy closed this May 29, 2025
@metascroy
Copy link
Contributor Author

Closing this. The migration was instead done in pieces.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@digantdesai digantdesai digantdesai left review comments

@jerryzh168 jerryzh168 jerryzh168 approved these changes

@SS-JIA SS-JIA SS-JIA approved these changes

@cccclai cccclai cccclai approved these changes

@mcr229 mcr229 mcr229 approved these changes

@jackzhxng jackzhxng Awaiting requested review from jackzhxng jackzhxng is a code owner

@iseeyuan iseeyuan Awaiting requested review from iseeyuan

@larryliu0820 larryliu0820 Awaiting requested review from larryliu0820 larryliu0820 is a code owner

@swolchok swolchok Awaiting requested review from swolchok swolchok is a code owner

@JacobSzwejbka JacobSzwejbka Awaiting requested review from JacobSzwejbka JacobSzwejbka is a code owner

@tarun292 tarun292 Awaiting requested review from tarun292

@lucylq lucylq Awaiting requested review from lucylq lucylq is a code owner

@mergennachin mergennachin Awaiting requested review from mergennachin mergennachin is a code owner

@kimishpatel kimishpatel Awaiting requested review from kimishpatel kimishpatel is a code owner

@shoumikhin shoumikhin Awaiting requested review from shoumikhin shoumikhin is a code owner

@chunit-quic chunit-quic Awaiting requested review from chunit-quic

@haowhsu-quic haowhsu-quic Awaiting requested review from haowhsu-quic

@shewu-quic shewu-quic Awaiting requested review from shewu-quic

@winskuo-quic winskuo-quic Awaiting requested review from winskuo-quic

@DannyYuyang-quic DannyYuyang-quic Awaiting requested review from DannyYuyang-quic

@YifanShenSZ YifanShenSZ Awaiting requested review from YifanShenSZ

@neuropilot-captain neuropilot-captain Awaiting requested review from neuropilot-captain

@DenisVieriu97 DenisVieriu97 Awaiting requested review from DenisVieriu97

@GregoryComer GregoryComer Awaiting requested review from GregoryComer

Assignees
No one assigned
Labels
ciflow/trunk CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. fb-exported topic: not user facing
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
7 participants
@metascroy @cccclai @facebook-github-bot @digantdesai @jerryzh168 @SS-JIA @mcr229