Skip to content

Fuse_modules in a qat-respecting way #12891

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
Merged
merged 8 commits into from
May 2, 2022
Merged

Fuse_modules in a qat-respecting way #12891

merged 8 commits into from
May 2, 2022

Conversation

@ORippler
Copy link
Contributor

@ORippler ORippler commented Apr 26, 2022 •
edited by akihironitta
Loading

What does this PR do?

For QAT, we should fuse modules in a qat-respecting way

Fixes #12890

Does your PR introduce any breaking changes? If yes, please list them.

None

Before submitting

  • Was this discussed/approved via a GitHub issue? (not for typos and docs)
  • Did you read the contributor guideline, Pull Request section?
  • Did you make sure your PR does only one thing, instead of bundling different changes together?
  • Did you make sure to update the documentation with your changes? (if necessary)
  • Did you write any new necessary tests? (not for typos and docs)
  • Did you verify new and existing tests pass locally with your changes?
  • Did you list all the breaking changes introduced by this pull request?
  • Did you update the CHANGELOG? (not for typos, docs, test updates, or minor internal changes/refactors)

PR review

Anyone in the community is welcome to review the PR.
Before you start reviewing, make sure you have read the review guidelines. In short, see the following bullet-list:

  • Is this pull request ready for review? (if not, please submit in draft mode)
  • Check that all items from Before submitting are resolved
  • Make sure the title is self-explanatory and the description concisely explains the PR
  • Add labels and milestones (and optionally projects) to the PR so it can be classified

Did you have fun?

Make sure you had fun coding 🙃

@carmocca carmocca added bug Something isn't working callback: quantization (removed) The QAT callback is only available up to 1.9.x (LTS) community This PR is from the community labels Apr 26, 2022
@carmocca carmocca added this to the 1.6.x milestone Apr 26, 2022
carmocca
carmocca approved these changes Apr 26, 2022
View reviewed changes
Copy link
Contributor

@carmocca carmocca left a comment

Choose a reason for hiding this comment

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

Do you think we could test this change?

akihironitta
akihironitta reviewed Apr 27, 2022
View reviewed changes
ORippler added 2 commits April 27, 2022 08:57
@ORippler
Add compatibility for PyTorch <1.11
923ae89 
In older pytorch versions, `fuse_modules` used the `Module.training`
flag to determine wheter fusion should be QAT-compliant or not, refer
https://github.com/pytorch/pytorch/releases/tag/v1.11.0
@ORippler
Add CHANGELOG for pull #12891
7f16d6a
@carmocca carmocca requested a review from akihironitta April 27, 2022 14:28
@carmocca carmocca enabled auto-merge (squash) April 27, 2022 14:28
@ORippler
Fix conditional import of fuse_modules_qat
c122790 
`torch.ao.quantization.fuse_modules_qat` was actually added in
torch 1.11.
auto-merge was automatically disabled April 27, 2022 14:52

Head branch was pushed to by a user without write access

akihironitta
akihironitta reviewed Apr 27, 2022
View reviewed changes
akihironitta
akihironitta approved these changes Apr 27, 2022
View reviewed changes
@mergify mergify bot added the ready PRs ready to be merged label Apr 27, 2022
@akihironitta akihironitta enabled auto-merge (squash) April 27, 2022 19:36
@mergify mergify bot removed the ready PRs ready to be merged label Apr 28, 2022
@mergify mergify bot added ready PRs ready to be merged and removed has conflicts ready PRs ready to be merged labels Apr 29, 2022
@akihironitta akihironitta merged commit 456cc87 into Lightning-AI:master May 2, 2022
@ORippler ORippler deleted the fix/12890 branch May 2, 2022 12:37
carmocca pushed a commit that referenced this pull request May 2, 2022
@ORippler @akihironitta @carmocca
çFuse_modules in a qat-respecting way (#12891)
fc36e4a 
* Fuse_modules in a qat-respecting way

* Add compatibility for PyTorch <1.11

In older pytorch versions, `fuse_modules` used the `Module.training`
flag to determine wheter fusion should be QAT-compliant or not, refer
https://github.com/pytorch/pytorch/releases/tag/v1.11.0

* Add CHANGELOG for pull #12891

* Fix conditional import of fuse_modules_qat

`torch.ao.quantization.fuse_modules_qat` was actually added in
torch 1.11.

* Update CHANGELOG.md

Co-authored-by: Akihiro Nitta <nitta@akihironitta.com>
carmocca pushed a commit that referenced this pull request May 3, 2022
@ORippler @akihironitta @carmocca
Fuse_modules in a qat-respecting way (#12891)
a44c140 
* Fuse_modules in a qat-respecting way

* Add compatibility for PyTorch <1.11

In older pytorch versions, `fuse_modules` used the `Module.training`
flag to determine wheter fusion should be QAT-compliant or not, refer
https://github.com/pytorch/pytorch/releases/tag/v1.11.0

* Add CHANGELOG for pull #12891

* Fix conditional import of fuse_modules_qat

`torch.ao.quantization.fuse_modules_qat` was actually added in
torch 1.11.

* Update CHANGELOG.md

Co-authored-by: Akihiro Nitta <nitta@akihironitta.com>
carmocca pushed a commit that referenced this pull request May 3, 2022
@ORippler @akihironitta @carmocca
Fuse_modules in a qat-respecting way (#12891)
60a1680 
* Fuse_modules in a qat-respecting way

* Add compatibility for PyTorch <1.11

In older pytorch versions, `fuse_modules` used the `Module.training`
flag to determine wheter fusion should be QAT-compliant or not, refer
https://github.com/pytorch/pytorch/releases/tag/v1.11.0

* Add CHANGELOG for pull #12891

* Fix conditional import of fuse_modules_qat

`torch.ao.quantization.fuse_modules_qat` was actually added in
torch 1.11.

* Update CHANGELOG.md

Co-authored-by: Akihiro Nitta <nitta@akihironitta.com>
lexierule pushed a commit that referenced this pull request May 3, 2022
@ORippler @akihironitta @lexierule
Fuse_modules in a qat-respecting way (#12891)
163084a 
* Fuse_modules in a qat-respecting way

* Add compatibility for PyTorch <1.11

In older pytorch versions, `fuse_modules` used the `Module.training`
flag to determine wheter fusion should be QAT-compliant or not, refer
https://github.com/pytorch/pytorch/releases/tag/v1.11.0

* Add CHANGELOG for pull #12891

* Fix conditional import of fuse_modules_qat

`torch.ao.quantization.fuse_modules_qat` was actually added in
torch 1.11.

* Update CHANGELOG.md

Co-authored-by: Akihiro Nitta <nitta@akihironitta.com>
@rohitgr7 rohitgr7 mentioned this pull request Jul 1, 2022
Merged
12 tasks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@williamFalcon williamFalcon Awaiting requested review from williamFalcon

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

@Borda Borda Awaiting requested review from Borda

@kaushikb11 kaushikb11 Awaiting requested review from kaushikb11

@SeanNaren SeanNaren Awaiting requested review from SeanNaren

@awaelchli awaelchli Awaiting requested review from awaelchli

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

@rohitgr7 rohitgr7 Awaiting requested review from rohitgr7

3 more reviewers

@otaj otaj otaj approved these changes

@carmocca carmocca carmocca approved these changes

@akihironitta akihironitta akihironitta approved these changes

Reviewers whose approvals may not affect merge requirements

Assignees

@Borda Borda

Labels

bug Something isn't working callback: quantization (removed) The QAT callback is only available up to 1.9.x (LTS) community This PR is from the community ready PRs ready to be merged

Projects

None yet

Milestone

pl:1.6.x

Development

Successfully merging this pull request may close these issues.

QuantizationAwareTraining does not use torch.ao.quantization.fuse_modules_qat

5 participants

@ORippler @Borda @otaj @carmocca @akihironitta