Fix GroupNorm to support Opset21 #2928

hamptonm1 · 2024-09-03T17:08:40Z

Updating GroupNorm to create a bias tensor and scale tensor using (C) Channel instead of (G) numGroups. The previous understanding was incorrect, I changed the code to reflect the updates made in ONNX.

Right now I included support for Opset18 as well as Opset21.

Here is more information on why the changes were made:
onnx/onnx#5466

…onm1/onnx-mlir into hamptonm/feature/groupnorm

Signed-off-by: hamptonm1 <79232909+hamptonm1@users.noreply.github.com>

hamptonm1 · 2024-09-05T00:57:01Z

@jenkins-droid test this please

Signed-off-by: hamptonm1 <79232909+hamptonm1@users.noreply.github.com>

…onm1/onnx-mlir into hamptonm/feature/groupnorm

Signed-off-by: hamptonm1 <79232909+hamptonm1@users.noreply.github.com>

hamptonm1 · 2024-09-12T23:39:03Z

@AlexandreEichenberger @chentong319 Okay, I took the feedback from you both and made some changes. Let me know what else needs to be fixed and if I got the reshape for biasScaleShape correct. The backend tests for GroupNormalization Passes and I added two new lit tests for Opset21.

hamptonm1 · 2024-09-12T23:41:03Z

src/Dialect/ONNX/ONNXOps/NN/Normalization.cpp

@@ -149,6 +149,21 @@ LogicalResult ONNXInstanceNormalizationOp::verify() {
  return success();
 }

+//===----------------------------------------------------------------------===//
+// GroupNormalizationV18


@AlexandreEichenberger Is this what you were thinking of? I added a print because if I add emitWarning it seems like all opset 18 tests fail and I figure we can still enable support for the meantime. I am fine with removing support in the near future but was not sure if any model still uses GroupNorm Opset 18.

LGTM, in general, we can also test other properties, but since this op is not going to be used, its fine.

hamptonm1 · 2024-09-12T23:41:36Z

FYI- @Sunny-Anand

AlexandreEichenberger

LGTM, nice use of templates. Indicated 2 small nits, otherwise good to go

AlexandreEichenberger · 2024-09-13T13:47:55Z

src/Dialect/ONNX/ONNXOps/NN/Normalization.cpp

@@ -149,6 +149,21 @@ LogicalResult ONNXInstanceNormalizationOp::verify() {
  return success();
 }

+//===----------------------------------------------------------------------===//
+// GroupNormalizationV18


LGTM, in general, we can also test other properties, but since this op is not going to be used, its fine.

AlexandreEichenberger · 2024-09-13T13:55:09Z

src/Dialect/ONNX/Transforms/Decompose.cpp

-    : public OpRewritePattern<ONNXGroupNormalizationOp> {
-  using OpRewritePattern<ONNXGroupNormalizationOp>::OpRewritePattern;
+template <typename OP>
+constexpr bool isNumGroup = std::is_same_v<OP, ONNXGroupNormalizationV18Op>;


Nice C++ construct, did not know we can do something like this without a function.

I would recommend a more explicit name ScaleAndBiasWithNumGroupShape

Okay I can change it!

AlexandreEichenberger · 2024-09-13T13:58:41Z

src/Dialect/ONNX/Transforms/Decompose.cpp

-    rewriter.replaceOp(groupNormOp, Y);
-    return success();
+    Value biasScaleShape = create.onnx.concat(biasScaleShapeType,
+        {NGShape, NGShape, oneDimShape}, /*axis*/


nit: format 0 should be on the previous line.

jenkins-droid · 2024-09-13T20:50:54Z

Jenkins Linux s390x Build #15627 [push] Fix GroupNorm to support... started at 16:50

jenkins-droid · 2024-09-13T20:50:57Z

Jenkins Linux amd64 Build #15624 [push] Fix GroupNorm to support... started at 15:50

jenkins-droid · 2024-09-13T20:50:57Z

Jenkins Linux ppc64le Build #14654 [push] Fix GroupNorm to support... started at 17:01

jenkins-droid · 2024-09-13T21:59:08Z

Jenkins Linux amd64 Build #15624 [push] Fix GroupNorm to support... passed after 1 hr 8 min

jenkins-droid · 2024-09-13T22:20:12Z

Jenkins Linux s390x Build #15627 [push] Fix GroupNorm to support... passed after 1 hr 29 min

jenkins-droid · 2024-09-13T22:52:55Z

Jenkins Linux ppc64le Build #14654 [push] Fix GroupNorm to support... passed after 2 hr 1 min

* Group norm for opset 21 * Testing phase * Fix GroupNorm to support Opset21 --------- Signed-off-by: hamptonm1 <79232909+hamptonm1@users.noreply.github.com> Co-authored-by: Megan Hampton <hamptonm@us.ibm.com> Signed-off-by: Sunny-Anand <sunnyanand.979@gmail.com>

* Change lowering of onnx.IF to Krnl (#2932) * implementation Signed-off-by: chentong319 <chentong@us.ibm.com> * test case change Signed-off-by: chentong319 <chentong@us.ibm.com> * format Signed-off-by: chentong319 <chentong@us.ibm.com> * add test for If back Signed-off-by: chentong319 <chentong@us.ibm.com> * format Signed-off-by: chentong319 <chentong@us.ibm.com> --------- Signed-off-by: chentong319 <chentong@us.ibm.com> Co-authored-by: Tung D. Le <tung@jp.ibm.com> Signed-off-by: Sunny-Anand <sunnyanand.979@gmail.com> * Update c style cast to c++ style cast (#2934) Signed-off-by: Mike Essenmacher <essen@us.ibm.com> Signed-off-by: Sunny-Anand <sunnyanand.979@gmail.com> * Change c style cast to c++ style cast (#2936) Signed-off-by: Mike Essenmacher <essen@us.ibm.com> Signed-off-by: Sunny-Anand <sunnyanand.979@gmail.com> * Add coding practices for onnx-mlir (#2935) Signed-off-by: Mike Essenmacher <essen@us.ibm.com> Signed-off-by: Sunny-Anand <sunnyanand.979@gmail.com> * try to use new buffer deallocation (#2919) * implementation Signed-off-by: Chen Tong <chentong@us.ibm.com> * comments Signed-off-by: Chen Tong <chentong@us.ibm.com> * format Signed-off-by: Chen Tong <chentong@us.ibm.com> --------- Signed-off-by: Chen Tong <chentong@us.ibm.com> Co-authored-by: Tung D. Le <tung@jp.ibm.com> Co-authored-by: Alexandre Eichenberger <alexe@us.ibm.com> Signed-off-by: Sunny-Anand <sunnyanand.979@gmail.com> * fix requirements.txt link Signed-off-by: Sunny-Anand <sunnyanand.979@gmail.com> * Reuse input buffer in lowering to krnl (#2939) * first step Signed-off-by: chentong319 <chentong@us.ibm.com> * cpu Signed-off-by: chentong319 <chentong@us.ibm.com> * options Signed-off-by: chentong319 <chentong@us.ibm.com> * unify Signed-off-by: chentong319 <chentong@us.ibm.com> * simd Signed-off-by: chentong319 <chentong@us.ibm.com> * comments Signed-off-by: chentong319 <chentong@us.ibm.com> * lit test Signed-off-by: chentong319 <chentong@us.ibm.com> * fix test Signed-off-by: chentong319 <chentong@us.ibm.com> * format Signed-off-by: chentong319 <chentong@us.ibm.com> * response Signed-off-by: chentong319 <chentong@us.ibm.com> --------- Signed-off-by: chentong319 <chentong@us.ibm.com> Signed-off-by: Sunny-Anand <sunnyanand.979@gmail.com> * Fix GroupNorm to support Opset21 (#2928) * Group norm for opset 21 * Testing phase * Fix GroupNorm to support Opset21 --------- Signed-off-by: hamptonm1 <79232909+hamptonm1@users.noreply.github.com> Co-authored-by: Megan Hampton <hamptonm@us.ibm.com> Signed-off-by: Sunny-Anand <sunnyanand.979@gmail.com> * Update Ops documentation for ONNX 1.16.2 (#2942) * Update Ops documentation for ONNX 1.16.2 * Fix format --------- Co-authored-by: Megan Hampton <hamptonm@us.ibm.com> Signed-off-by: Sunny-Anand <sunnyanand.979@gmail.com> * LLVM/StableHLO Upgrade eaa95a1 (#2943) Co-authored-by: Megan Hampton <hamptonm@us.ibm.com> Signed-off-by: Sunny-Anand <sunnyanand.979@gmail.com> * added support for no-zero-point quantization (#2938) Signed-off-by: Alexandre Eichenberger <alexe@us.ibm.com> Co-authored-by: Tung D. Le <tung@jp.ibm.com> Signed-off-by: Sunny-Anand <sunnyanand.979@gmail.com> * update with main Signed-off-by: Sunny-Anand <sunnyanand.979@gmail.com> --------- Signed-off-by: chentong319 <chentong@us.ibm.com> Signed-off-by: Sunny-Anand <sunnyanand.979@gmail.com> Signed-off-by: Mike Essenmacher <essen@us.ibm.com> Signed-off-by: Chen Tong <chentong@us.ibm.com> Signed-off-by: hamptonm1 <79232909+hamptonm1@users.noreply.github.com> Signed-off-by: Alexandre Eichenberger <alexe@us.ibm.com> Signed-off-by: Sunny Anand <164108690+Sunny-Anand@users.noreply.github.com> Co-authored-by: Tong Chen <chentong@us.ibm.com> Co-authored-by: Tung D. Le <tung@jp.ibm.com> Co-authored-by: Mike Essenmacher <112431871+mikeessen@users.noreply.github.com> Co-authored-by: Alexandre Eichenberger <alexe@us.ibm.com> Co-authored-by: hamptonm1 <79232909+hamptonm1@users.noreply.github.com> Co-authored-by: Megan Hampton <hamptonm@us.ibm.com>

MegoHam21 and others added 16 commits August 12, 2024 07:36

Group norm for opset 21

c64ad53

Merge branch 'onnx:main' into hamptonm/feature/groupnorm

db003ba

Testing phase

fe10937

Merge branch 'onnx:main' into hamptonm/feature/groupnorm

8979ccb

Fix GroupNorm to support Opset21

fed2948

Merge branch 'hamptonm/feature/groupnorm' of https://github.com/hampt…

b650554

…onm1/onnx-mlir into hamptonm/feature/groupnorm

Merge branch 'main' into hamptonm/feature/groupnorm

35692e8

Fix format

7209660

Merge branch 'hamptonm/feature/groupnorm' of https://github.com/hampt…

f2e87a4

…onm1/onnx-mlir into hamptonm/feature/groupnorm

Still fixing format here

52721fc

Trying this out

32d0132

Logic branches

fb4b39c

Fix linter

c6a0ebb

Try again

b36f32d

Update Dockerfile.llvm-project

a1a15e3

Signed-off-by: hamptonm1 <79232909+hamptonm1@users.noreply.github.com>

Update requirements.txt

3b0bf5f

Signed-off-by: hamptonm1 <79232909+hamptonm1@users.noreply.github.com>

hamptonm1 and others added 13 commits September 4, 2024 20:59

Merge branch 'main' into hamptonm/feature/groupnorm

770b1c4

Update requirements.txt

3f33e0b

Signed-off-by: hamptonm1 <79232909+hamptonm1@users.noreply.github.com>

Testing it

6b7dc63

Merge branch 'hamptonm/feature/groupnorm' of https://github.com/hampt…

fd20266

…onm1/onnx-mlir into hamptonm/feature/groupnorm

Merge branch 'main' into hamptonm/feature/groupnorm

8d50a48

Fix format

76de507

Merge branch 'hamptonm/feature/groupnorm' of https://github.com/hampt…

31e542c

…onm1/onnx-mlir into hamptonm/feature/groupnorm

Testing still

07b4e51

Still figuring this out

a3e8359

Try it out

6b3f4dd

Format fixed

da9c0ab

Merge branch 'main' into hamptonm/feature/groupnorm

41ee2b9

Grr

b9fa3b0

MegoHam21 and others added 7 commits September 11, 2024 17:04

Figure out why it is not working

d0b51e2

My fault

ca8da95

Update Decompose.cpp

d8b26f9

Signed-off-by: hamptonm1 <79232909+hamptonm1@users.noreply.github.com>

Update Decompose.cpp again

6d35971

Signed-off-by: hamptonm1 <79232909+hamptonm1@users.noreply.github.com>

Update Decompose.cpp part 3

327e6ef

Signed-off-by: hamptonm1 <79232909+hamptonm1@users.noreply.github.com>

Still at it

ba471a1

Signed-off-by: hamptonm1 <79232909+hamptonm1@users.noreply.github.com>

Test it!

6c2de2d

Signed-off-by: hamptonm1 <79232909+hamptonm1@users.noreply.github.com>

hamptonm1 marked this pull request as ready for review September 12, 2024 14:31

MegoHam21 added 2 commits September 12, 2024 12:56

I think I got it to work

76e8422

Hmmm

de19219

hamptonm1 requested review from AlexandreEichenberger, tungld and chentong319 September 12, 2024 23:37

hamptonm1 commented Sep 12, 2024

View reviewed changes

AlexandreEichenberger approved these changes Sep 13, 2024

View reviewed changes

MegoHam21 and others added 3 commits September 13, 2024 13:02

Address comments

e459d52

Fix format

0f1b161

Merge branch 'main' into hamptonm/feature/groupnorm

41eb8dd

hamptonm1 merged commit 2f2ccc5 into onnx:main Sep 13, 2024
7 checks passed

hamptonm1 deleted the hamptonm/feature/groupnorm branch September 13, 2024 20:49

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Fix GroupNorm to support Opset21 #2928

Fix GroupNorm to support Opset21 #2928

hamptonm1 commented Sep 3, 2024 •

edited

Loading

hamptonm1 commented Sep 5, 2024 •

edited

Loading

hamptonm1 commented Sep 12, 2024 •

edited

Loading

hamptonm1 Sep 12, 2024

AlexandreEichenberger Sep 13, 2024

hamptonm1 commented Sep 12, 2024

AlexandreEichenberger left a comment

AlexandreEichenberger Sep 13, 2024

AlexandreEichenberger Sep 13, 2024

hamptonm1 Sep 13, 2024

AlexandreEichenberger Sep 13, 2024

hamptonm1 Sep 13, 2024

jenkins-droid commented Sep 13, 2024

jenkins-droid commented Sep 13, 2024

jenkins-droid commented Sep 13, 2024

jenkins-droid commented Sep 13, 2024

jenkins-droid commented Sep 13, 2024

jenkins-droid commented Sep 13, 2024

Fix GroupNorm to support Opset21 #2928

Fix GroupNorm to support Opset21 #2928

Conversation

hamptonm1 commented Sep 3, 2024 • edited Loading

hamptonm1 commented Sep 5, 2024 • edited Loading

hamptonm1 commented Sep 12, 2024 • edited Loading

hamptonm1 Sep 12, 2024

Choose a reason for hiding this comment

AlexandreEichenberger Sep 13, 2024

Choose a reason for hiding this comment

hamptonm1 commented Sep 12, 2024

AlexandreEichenberger left a comment

Choose a reason for hiding this comment

AlexandreEichenberger Sep 13, 2024

Choose a reason for hiding this comment

AlexandreEichenberger Sep 13, 2024

Choose a reason for hiding this comment

hamptonm1 Sep 13, 2024

Choose a reason for hiding this comment

AlexandreEichenberger Sep 13, 2024

Choose a reason for hiding this comment

hamptonm1 Sep 13, 2024

Choose a reason for hiding this comment

jenkins-droid commented Sep 13, 2024

jenkins-droid commented Sep 13, 2024

jenkins-droid commented Sep 13, 2024

jenkins-droid commented Sep 13, 2024

jenkins-droid commented Sep 13, 2024

jenkins-droid commented Sep 13, 2024

hamptonm1 commented Sep 3, 2024 •

edited

Loading

hamptonm1 commented Sep 5, 2024 •

edited

Loading

hamptonm1 commented Sep 12, 2024 •

edited

Loading