Add various uninterpreted bit tensor data types #94992

vkuzo · 2023-02-16T18:56:15Z

Summary:

This PR adds a set of unintrepreted data types on PyTorch which can be used to implement experimental functionality out of core (think fp8, int4, int16 quant, etc).

Note: this is a copy-pasta of #89990 with a bug fix for clang9, easier to just to put up another PR since I'm not sure how comandeering works with Meta-only changes.

@bypass-github-export-checks

Test Plan:

python test/test_quantization.py -k TestBits

Reviewers:

Subscribers:

Tasks:

Tags:

Fixes #ISSUE_NUMBER

pytorch-bot · 2023-02-16T18:56:18Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/94992

📄 Preview Python docs built from this PR
📄 Preview C++ docs built from this PR
❓ Need help or want to give feedback on the CI? Visit the bot commands wiki or our office hours

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

✅ No Failures

As of commit cd20e12:
💚 Looks good so far! There are no failures yet. 💚

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

vkuzo · 2023-02-16T21:03:16Z

c10/core/ScalarType.h

  // are not sure about the correct value for type promotion.
-  static constexpr ScalarType _promoteTypesLookup[static_cast<int>(
-      ScalarType::NumOptions)][static_cast<int>(ScalarType::NumOptions)] = {
+  // clang-format off


note: this is now needed to prevent clang from messing up the table formatting

angelayi

lgtm! Note that you'll have to make internal changes too.

vkuzo · 2023-02-16T21:23:42Z

lgtm! Note that you'll have to make internal changes too.

yep, just waiting for green CI here before I do that, to minimize work

facebook-github-bot · 2023-02-16T23:25:08Z

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

Summary: This PR adds a set of unintrepreted data types on PyTorch which can be used to implement experimental functionality out of core (think fp8, int4, int16 quant, etc). Note: this is a copy-pasta of #89990 with a bug fix for clang9, easier to just to put up another PR since I'm not sure how comandeering works with Meta-only changes. Test Plan: ``` python test/test_quantization.py -k TestBits ``` Reviewers: Subscribers: Tasks: Tags:

facebook-github-bot · 2023-02-17T15:58:23Z

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

facebook-github-bot · 2023-02-18T00:02:41Z

@pytorchbot merge -f 'Landed internally'

(Initiating merge automatically since Phabricator Diff has merged, using force because this PR might not pass merge_rules.json but landed internally)

pytorchmergebot · 2023-02-18T00:04:26Z

Merge started

Your change will be merged immediately since you used the force (-f) flag, bypassing any CI checks (ETA: 1-5 minutes).

Learn more about merging in the wiki.

Questions? Feedback? Please reach out to the PyTorch DevX Team

Advanced Debugging

Check the merge workflow status
here

This reverts commit 9dbfca7.

atalman · 2023-02-23T23:51:34Z

@pytorchmergebot revert -m "breaks libtorch windows builds see: #95406" -c wierd

pytorch-bot · 2023-02-23T23:51:35Z

❌ 🤖 pytorchbot command failed:

@pytorchbot revert: error: argument -c/--classification: invalid choice: 'wierd' (choose from 'nosignal', 'ignoredsignal', 'landrace', 'weird', 'ghfirst')

usage: @pytorchbot revert -m MESSAGE -c
                          {nosignal,ignoredsignal,landrace,weird,ghfirst}

Try @pytorchbot --help for more info.

atalman · 2023-02-23T23:52:37Z

@pytorchbot revert -m "breaks libtorch windows nightly builds see: #95406" -c nosignal

pytorchmergebot · 2023-02-23T23:54:18Z

@pytorchbot successfully started a revert job. Check the current status here.
Questions? Feedback? Please reach out to the PyTorch DevX Team

pytorchmergebot · 2023-02-23T23:54:28Z

@vkuzo your PR has been successfully reverted.

This reverts commit 9dbfca7. Reverted #94992 on behalf of https://github.com/atalman due to breaks libtorch windows nightly builds see: #95406

This reverts commit 9dbfca7. Reverted pytorch/pytorch#94992 on behalf of https://github.com/atalman due to breaks libtorch windows nightly builds see: pytorch/pytorch#95406

Summary: This is a retry of #94992 which was reverted due to CI issues. This PR adds a set of unintrepreted data types on PyTorch which can be used to implement experimental functionality out of core (think fp8, int4, int16 quant, etc). Test Plan: ``` python test/test_quantization.py -k TestBits ``` Reviewers: Subscribers: Tasks: Tags:

Summary: This is a retry of #94992 which was reverted due to CI issues. This PR adds a set of unintrepreted data types on PyTorch which can be used to implement experimental functionality out of core (think fp8, int4, int16 quant, etc). @bypass-github-export-checks Test Plan: ``` python test/test_quantization.py -k TestBits ``` Reviewers: Subscribers: Tasks: Tags: Fixes #ISSUE_NUMBER Pull Request resolved: #95860 Approved by: https://github.com/atalman

This reverts commit 9dbfca7. Reverted pytorch/pytorch#94992 on behalf of https://github.com/atalman due to breaks libtorch windows nightly builds see: pytorch/pytorch#95406

Summary: This is a retry of pytorch/pytorch#94992 which was reverted due to CI issues. This PR adds a set of unintrepreted data types on PyTorch which can be used to implement experimental functionality out of core (think fp8, int4, int16 quant, etc). @bypass-github-export-checks Test Plan: ``` python test/test_quantization.py -k TestBits ``` Reviewers: Subscribers: Tasks: Tags: Fixes #ISSUE_NUMBER Pull Request resolved: pytorch/pytorch#95860 Approved by: https://github.com/atalman

Summary: This is a retry of pytorch#94992 which was reverted due to CI issues. This PR adds a set of unintrepreted data types on PyTorch which can be used to implement experimental functionality out of core (think fp8, int4, int16 quant, etc). @bypass-github-export-checks Test Plan: ``` python test/test_quantization.py -k TestBits ``` Reviewers: Subscribers: Tasks: Tags: Fixes #ISSUE_NUMBER Pull Request resolved: pytorch#95860 Approved by: https://github.com/atalman

…rch#94992)"" This reverts commit 3bafecf.

This reverts commit 9dbfca7.

Summary: This PR adds a set of unintrepreted data types on PyTorch which can be used to implement experimental functionality out of core (think fp8, int4, int16 quant, etc). Note: this is a copy-pasta of pytorch#89990 with a bug fix for clang9, easier to just to put up another PR since I'm not sure how comandeering works with Meta-only changes. @bypass-github-export-checks Test Plan: ``` python test/test_quantization.py -k TestBits ``` Reviewers: Subscribers: Tasks: Tags: Fixes #ISSUE_NUMBER Pull Request resolved: pytorch#94992 Approved by: https://github.com/angelayi

This reverts commit 9dbfca7. Reverted pytorch#94992 on behalf of https://github.com/atalman due to breaks libtorch windows nightly builds see: pytorch#95406

vkuzo requested a review from jerryzh168 as a code owner February 16, 2023 18:56

pytorch-bot bot added the release notes: quantization release notes category label Feb 16, 2023

This was referenced Feb 16, 2023

[not for land] testing bitstensor CI #94941

Closed

Add bits tensor types #89990

Closed

vkuzo force-pushed the bitstensor_0 branch 2 times, most recently from de86298 to c27c21e Compare February 16, 2023 21:02

vkuzo commented Feb 16, 2023

View reviewed changes

vkuzo requested review from Chillee, angelayi and ezyang February 16, 2023 21:07

angelayi approved these changes Feb 16, 2023

View reviewed changes

vkuzo force-pushed the bitstensor_0 branch from c27c21e to cd20e12 Compare February 17, 2023 15:57

pytorchmergebot added the Merged label Feb 18, 2023

pytorchmergebot closed this in 9dbfca7 Feb 18, 2023

atalman added a commit to atalman/pytorch that referenced this pull request Feb 23, 2023

Revert "Add various uninterpreted bit tensor data types (pytorch#94992)"

c29c44e

This reverts commit 9dbfca7.

pytorchmergebot added the Reverted label Feb 23, 2023

vkuzo mentioned this pull request Mar 2, 2023

Add various uninterpreted bit tensor data types (try 2) #95860

Closed

msaroufim mentioned this pull request Mar 3, 2023

Remove mention of dynamo.optimize() in docs #96002

Closed

pruthvistony added a commit to ROCm/pytorch that referenced this pull request May 2, 2023

Revert "Revert "Add various uninterpreted bit tensor data types (pyto…

1ff2f68

…rch#94992)"" This reverts commit 3bafecf.

pruthvistony added a commit to ROCm/pytorch that referenced this pull request May 2, 2023

Revert "Add various uninterpreted bit tensor data types (pytorch#94992)"

c290346

This reverts commit 9dbfca7.

github-actions bot deleted the bitstensor_0 branch August 20, 2024 01:57

Add various uninterpreted bit tensor data types #94992

Add various uninterpreted bit tensor data types #94992

Uh oh!

Conversation

vkuzo commented Feb 16, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pytorch-bot bot commented Feb 16, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/94992

✅ No Failures

Uh oh!

vkuzo Feb 16, 2023

Choose a reason for hiding this comment

Uh oh!

angelayi left a comment

Choose a reason for hiding this comment

Uh oh!

vkuzo commented Feb 16, 2023

Uh oh!

facebook-github-bot commented Feb 16, 2023

Uh oh!

facebook-github-bot commented Feb 17, 2023

Uh oh!

facebook-github-bot commented Feb 18, 2023

Uh oh!

pytorchmergebot commented Feb 18, 2023

Merge started

Uh oh!

atalman commented Feb 23, 2023

Uh oh!

pytorch-bot bot commented Feb 23, 2023

Uh oh!

atalman commented Feb 23, 2023

Uh oh!

pytorchmergebot commented Feb 23, 2023

Uh oh!

pytorchmergebot commented Feb 23, 2023

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

vkuzo commented Feb 16, 2023 •

edited

Loading

pytorch-bot bot commented Feb 16, 2023 •

edited

Loading