[ATen] Support multi dim any and all reductions #110310

peterbell10 · 2023-09-29T19:56:38Z

Stack from ghstack (oldest at bottom):

This adds a new overload to all and any with support for multiple reduction dims.

all.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor
any.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor

pytorch-bot · 2023-09-29T19:56:41Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/110310

📄 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 aeaa808 with merge base 4f79161 ():
💚 Looks good so far! There are no failures yet. 💚

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

[ghstack-poisoned]

ghstack-source-id: b7c2801 Pull Request resolved: pytorch#110310

[ghstack-poisoned]

ghstack-source-id: b940f74 Pull Request resolved: pytorch#110310

[ghstack-poisoned]

This adds a new overload to `all` and `any` with support for multiple reduction dims. ``` all.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor any.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor ``` [ghstack-poisoned]

peterbell10 · 2023-10-02T12:22:44Z

test/onnx/test_fx_op_consistency.py

        "amax",
        "amin",
-        "any",
+        # "any",    - onnxscript doesn't handle aten::any.dims


Apparently this is testing out-of-tree code, so can't be fixed in this PR
https://github.com/microsoft/onnxscript/blob/f8046e17490a3222d4ee4bca346b4449a2d30ae4/onnxscript/function_libs/torch_lib/ops/core.py#L404

cc @BowenBao what is the best way forward here?

cc @justinchuby

The current best thing to do is to xfail it here: https://github.com/pytorch/pytorch/pull/110310/files/73dd8600a8097f98b675510024e01ea1a3d6af5d#diff-db2f78a51511bb172cbfde1b2f68272b8b33049abe2571cded27bcd0f3ae5fa4R176 and track it with an issue.

e.g.

xfail( "any", reason="reason and link to issue" ),

I will discuss with the team to find a better solution forward. Thanks!

aten/src/ATen/native/ReduceOps.cpp

test/onnx/test_pytorch_onnx_onnxruntime_cuda.py

tools/autograd/derivatives.yaml

tools/pyi/gen_pyi.py

torch/_refs/__init__.py

aten/src/ATen/native/ReduceOps.cpp

aten/src/ATen/native/native_functions.yaml

This adds a new overload to `all` and `any` with support for multiple reduction dims. ``` all.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor any.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor ``` [ghstack-poisoned]

peterbell10 · 2023-10-11T13:02:41Z

Okay if it's a pre-existing issue then I'll just add a skip.

This adds a new overload to `all` and `any` with support for multiple reduction dims. ``` all.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor any.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor ``` [ghstack-poisoned]

albanD · 2023-10-11T16:45:36Z

If the error is a hard crash, you will have to add it to the skip list instead of xfail.

This adds a new overload to `all` and `any` with support for multiple reduction dims. ``` all.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor any.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor ``` [ghstack-poisoned]

albanD

SGTM

This adds a new overload to `all` and `any` with support for multiple reduction dims. ``` all.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor any.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor ``` [ghstack-poisoned]

Implement aten::{all,any}.dims according to pytorch/pytorch#110310 Tests will be enabled after the PyTorch PR is merged

This adds a new overload to `all` and `any` with support for multiple reduction dims. ``` all.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor any.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor ``` [ghstack-poisoned]

peterbell10 · 2023-10-13T20:53:59Z

@pytorchbot rebase

pytorchmergebot · 2023-10-13T20:56:08Z

@pytorchbot started a rebase job onto refs/remotes/origin/viable/strict. Check the current status here

This adds a new overload to `all` and `any` with support for multiple reduction dims. ``` all.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor any.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor ``` [ghstack-poisoned]

pytorchmergebot · 2023-10-13T20:56:27Z

Successfully rebased gh/peterbell10/626/orig onto refs/remotes/origin/viable/strict, please pull locally before adding more changes (for example, via ghstack checkout https://github.com/pytorch/pytorch/pull/110310)

ghstack-source-id: a6bdff4 Pull Request resolved: #110310

This adds a new overload to `all` and `any` with support for multiple reduction dims. ``` all.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor any.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor ``` [ghstack-poisoned]

Pull Request resolved: #110311 Approved by: https://github.com/lezcano ghstack dependencies: #110310

This adds a new overload to `all` and `any` with support for multiple reduction dims. ``` all.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor any.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor ``` Pull Request resolved: pytorch#110310 Approved by: https://github.com/lezcano, https://github.com/albanD, https://github.com/justinchuby

Pull Request resolved: pytorch#110311 Approved by: https://github.com/lezcano ghstack dependencies: pytorch#110310

This adds a new overload to `all` and `any` with support for multiple reduction dims. ``` all.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor any.dims(Tensor self, int[1]? dim=None, bool keepdim=False) -> Tensor ``` Pull Request resolved: pytorch#110310 Approved by: https://github.com/lezcano, https://github.com/albanD, https://github.com/justinchuby

Pull Request resolved: pytorch#110311 Approved by: https://github.com/lezcano ghstack dependencies: pytorch#110310

[ATen] Support multi dim any and all reductions

ab234b7

[ghstack-poisoned]

This was referenced Sep 29, 2023

[inductor] Avoid bool being upcast to int #109913

Closed

[inductor] Decompose boolean min/max into all/any #110311

Closed

Update on "[ATen] Support multi dim any and all reductions"

e150637

[ghstack-poisoned]

pytorchbot added the open source label Sep 29, 2023

peterbell10 added a commit to peterbell10/pytorch that referenced this pull request Sep 30, 2023

[ATen] Support multi dim any and all reductions

0552810

ghstack-source-id: b7c2801 Pull Request resolved: pytorch#110310

peterbell10 added 2 commits September 30, 2023 16:56

Update on "[ATen] Support multi dim any and all reductions"

1e7bf32

[ghstack-poisoned]

Update on "[ATen] Support multi dim any and all reductions"

6683a81

[ghstack-poisoned]

peterbell10 added a commit to peterbell10/pytorch that referenced this pull request Sep 30, 2023

[ATen] Support multi dim any and all reductions

b4bcdc9

ghstack-source-id: b940f74 Pull Request resolved: pytorch#110310

peterbell10 added 3 commits September 30, 2023 22:22

Update on "[ATen] Support multi dim any and all reductions"

45d00ff

[ghstack-poisoned]

Update on "[ATen] Support multi dim any and all reductions"

c2f53bb

[ghstack-poisoned]

Update on "[ATen] Support multi dim any and all reductions"

a9a531a

[ghstack-poisoned]

peterbell10 marked this pull request as ready for review October 2, 2023 12:16

peterbell10 requested review from BowenBao, abock, albanD, mruberry, ngimel, soulitzer, thiagocrepaldi and wschin as code owners October 2, 2023 12:16

peterbell10 requested a review from lezcano October 2, 2023 12:16

pytorch-bot bot added the release notes: onnx torch.onnx related changes that should show up in the release notes label Oct 2, 2023

peterbell10 commented Oct 2, 2023

View reviewed changes

lezcano reviewed Oct 2, 2023

View reviewed changes

aten/src/ATen/native/native_functions.yaml Outdated Show resolved Hide resolved

peterbell10 mentioned this pull request Oct 2, 2023

[inductor] Cast loads from boolean tensors to tl.int1 #110388

Closed

peterbell10 added 2 commits October 11, 2023 14:08

peterbell10 requested a review from kulinseth as a code owner October 11, 2023 20:09

peterbell10 added 2 commits October 11, 2023 21:09

albanD approved these changes Oct 11, 2023

View reviewed changes

justinchuby added a commit to microsoft/onnxscript that referenced this pull request Oct 13, 2023

Implement aten::{all,any}.dims | feat(torchlib) (#1084)

2187973

Implement aten::{all,any}.dims according to pytorch/pytorch#110310 Tests will be enabled after the PyTorch PR is merged

justinchuby approved these changes Oct 13, 2023

View reviewed changes

peterbell10 added 2 commits October 13, 2023 10:48

pytorchmergebot pushed a commit that referenced this pull request Oct 13, 2023

[ATen] Support multi dim any and all reductions

a4f3ae8

ghstack-source-id: a6bdff4 Pull Request resolved: #110310

pytorchmergebot added the Merged label Oct 24, 2023

pytorchmergebot closed this in 46e80ce Oct 24, 2023

pytorchmergebot pushed a commit that referenced this pull request Oct 24, 2023

[inductor] Decompose boolean min/max into all/any (#110311)

1dd5708

Pull Request resolved: #110311 Approved by: https://github.com/lezcano ghstack dependencies: #110310

facebook-github-bot deleted the gh/peterbell10/626/head branch October 28, 2023 14:24

xuhancn pushed a commit to xuhancn/pytorch that referenced this pull request Nov 7, 2023

[inductor] Decompose boolean min/max into all/any (pytorch#110311)

8369897

Pull Request resolved: pytorch#110311 Approved by: https://github.com/lezcano ghstack dependencies: pytorch#110310

Skylion007 pushed a commit to Skylion007/pytorch that referenced this pull request Nov 14, 2023

[inductor] Decompose boolean min/max into all/any (pytorch#110311)

45afb9f

Pull Request resolved: pytorch#110311 Approved by: https://github.com/lezcano ghstack dependencies: pytorch#110310

[ATen] Support multi dim any and all reductions #110310

[ATen] Support multi dim any and all reductions #110310

Uh oh!

Conversation

peterbell10 commented Sep 29, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pytorch-bot bot commented Sep 29, 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/110310

✅ No Failures

Uh oh!

peterbell10 Oct 2, 2023

Choose a reason for hiding this comment

Uh oh!

albanD Oct 2, 2023

Choose a reason for hiding this comment

Uh oh!

peterbell10 Oct 4, 2023

Choose a reason for hiding this comment

Uh oh!

justinchuby Oct 4, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

peterbell10 commented Oct 11, 2023

Uh oh!

albanD commented Oct 11, 2023

Uh oh!

albanD left a comment

Choose a reason for hiding this comment

Uh oh!

peterbell10 commented Oct 13, 2023

Uh oh!

pytorchmergebot commented Oct 13, 2023

Uh oh!

pytorchmergebot commented Oct 13, 2023

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

peterbell10 commented Sep 29, 2023 •

edited

Loading

pytorch-bot bot commented Sep 29, 2023 •

edited

Loading

justinchuby Oct 4, 2023 •

edited

Loading