Cherry pick PRs to Rel-1.3.1 by stevenlix · Pull Request #4198 · microsoft/onnxruntime

stevenlix · 2020-06-11T00:46:10Z

Java GPu artifact naming (#4179)
Change group id to com.microsoft.onnxruntime per requirements.
Create Java publishing pipeline (#3944)
Update OnnxRuntime.java for OS X environment. (#3985)
[java] Adds a CUDA test (#3956)
add script to support update nodejs binding version (#4164)
[Node.js binding] add linux and mac package (#4157)
[Node.js binding] upgrade node-addon-api to 3.0 (#4148)
fix build: pipeline Node.js version to 12.16.3 (#4145)
[Nodejs binding] create a new pipeline to generate signed binaries (#4104)
[Node.js binding] add build flag for node.js binding (#3948)
[Node.js binding] fix linux build (#3927)
link to folder instead of READMEs inside folder (#3938)
bump up ORT version to 1.3.1 (#4181)
move back to toolset 14.16 to possibly work around nvcc bug (#4180)

Symbolic shape inference exit on models without onnx opset used (#4090)
Fix Nuphar test failure
Enlarge the read buffer size in C#/Java test code (#4150)
Temporarily disable windows static analysis CI job

otherwise hard to find the source code

…4104) * add yml files * update pipeline * fix yaml syntax * yaml pop BuildCSharp * udpate yaml * do not stage codesign summary

* try mac pipeline * fix path separator * copy prebuilds folder * split esrp yaml for win/mac * disable mac signing temporarily * add linux * fix indent * add nodetool in linux * add nodetool in win-ci-2019 * replace linux build by custom docker scripts * use manylinux as node 12.16 not working on centos6 * try ubuntu * loosen timeout for test case - multiple runs calls

* [java] - adding a cuda enabled test. * Adding --build_java to the windows gpu ci pipeline. * Removing a stray line from the unit tests that always enabled CUDA for Java.

onnxruntime init failure due to wrong path of reading native libraries. In OS X 64 system, the arch name is detected as x86 which generates invalid path to read native libraries. Exception java.lang.UnsatisfiedLinkError: no onnxruntime in java.library.path at java.lang.ClassLoader.loadLibrary(ClassLoader.java:1867) at java.lang.Runtime.loadLibrary0(Runtime.java:870) at java.lang.System.loadLibrary(System.java:1122) at ai.onnxruntime.OnnxRuntime.load(OnnxRuntime.java:174) at ai.onnxruntime.OnnxRuntime.init(OnnxRuntime.java:81) at ai.onnxruntime.OrtEnvironment.<clinit>(OrtEnvironment.java:24)

Create CPU and GPu Java publishing pipelines. Final jars are tested on all platforms. However, signing and publishing to maven are manual steps.

@Craigacp

Modify gradle build so artifactID has _gpu for GPU builds. Pass USE_CUDA flag on CUDA build Adjust publishing pipelines to extract POM from a correct path. Co-Authored-By: @Craigacp

yuslepukhin

snnn · 2020-06-11T02:38:01Z

To make the pipeline pass, please take these changes as well:
2ab3a19
aafe988
d63b905
6f8a4f4

* Symbolic shape inference exit on models without onnx opset used * Temporary fix for ConvTranspose with symbolic input dims Co-authored-by: Changming Sun <me@sunchangming.com>

1. Enlarge the read buffer size further, so that our code can run even faster. TODO: need apply the similar changes to python some other language bindings. 2. Add coreml_VGG16_ImageNet to the test exclusion set of x86_32. It is not a new model but previously we didn't run the test against x86_32.

stevenlix · 2020-06-11T05:19:07Z

@snnn , I solved some conflicts in your PRs. Please take a look .
aafe988
conflict: tools/ci_build/github/azure-pipelines/win-ci-pipeline.yml
Solution: remove windows static analysis CI job

2ab3a19
conflict: onnxruntime/core/graph/model.cc
google::protobuf::io::IstreamInputStream zero_copy_input(&model_istream);
->
google::protobuf::io::IstreamInputStream zero_copy_input(&model_istream, protobuf_block_size_in_bytes);

const bool result = model_proto.ParseFromFileDescriptor(fd);
->
FileInputStream input(fd, protobuf_block_size_in_bytes);
const bool result = model_proto.ParseFromZeroCopyStream(&input) && input.GetErrno() == 0;

conflict: onnxruntime/test/onnx/TestCase.cc
static constexpr int protobuf_block_size_in_bytes = 4 * 1024 * 1024;

snnn · 2020-06-11T06:21:52Z

csharp/test/Microsoft.ML.OnnxRuntime.Tests/InferenceTest.cs

                skipModels["tf_nasnet_large"] = "Get preallocated buffer for initializer ConvBnFusion_BN_B_cell_11/beginning_bn/beta:0_331 failed";
                skipModels["test_zfnet512"] = "System out of memory";
                skipModels["test_bvlc_reference_caffenet"] = "System out of memory";
+                skipModels["coreml_VGG16_ImageNet"] = "System out of memory";


Please insert the line below to:
https://github.com/microsoft/onnxruntime/blob/master/csharp/test/Microsoft.ML.OnnxRuntime.Tests/InferenceTest.cs#L558

{ "coreml_Imputer-LogisticRegression_sklearn_load_breast_cancer", "Can't determine model file name" },

It is picked from 3eaec57

Windows CPU CI Pipeline still failed with the same error after I cherry picked those 4 PRs,
https://dev.azure.com/onnxruntime/onnxruntime/_build/results?buildId=168533&view=logs&j=5b022bb4-70a7-5401-8766-a8a7802c7150&t=23e58576-d346-55e7-f729-22ace67bada2&l=98

##[error]Error: The process 'C:\Program Files\dotnet\dotnet.exe' failed with exit code 1
Info: Azure Pipelines hosted agents have been updated to contain .Net Core 3.x (3.0 and 3.1) SDK/Runtime along with 2.1. Unless you have locked down a SDK version for your project(s), 3.x SDK might be picked up which might have breaking behavior as compared to previous versions.
Some commonly encountered changes are:
If you're using Publish command with -o or --Output argument, you will see that the output folder is now being created at root directory rather than Project File's directory. To learn about more such changes and troubleshoot, refer here: https://docs.microsoft.com/en-us/azure/devops/pipelines/tasks/build/dotnet-core-cli?view=azure-devops#troubleshooting
##[error]Dotnet command failed with non-zero exit code on the following projects : C:\agent_work\1\s\csharp\test\Microsoft.ML.OnnxRuntime.Tests\Microsoft.ML.OnnxRuntime.Tests.csproj

yuslepukhin

yuslepukhin

prasanthpul and others added 15 commits June 10, 2020 14:04

link to folder instead of READMEs inside folder (#3938)

d4c830f

otherwise hard to find the source code

[Node.js binding] fix linux build (#3927)

628dfc4

[Node.js binding] add build flag for node.js binding (#3948)

1c20b57

[Nodejs binding] create a new pipeline to generate signed binaries (#…

b0e2ff4

…4104) * add yml files * update pipeline * fix yaml syntax * yaml pop BuildCSharp * udpate yaml * do not stage codesign summary

fix build: pipeline Node.js version to 12.16.3 (#4145)

1dde49f

[Node.js binding] upgrade node-addon-api to 3.0 (#4148)

a9f7e38

add script to support update nodejs binding version (#4164)

909519b

[java] Adds a CUDA test (#3956)

169cf2b

* [java] - adding a cuda enabled test. * Adding --build_java to the windows gpu ci pipeline. * Removing a stray line from the unit tests that always enabled CUDA for Java.

Create Java publishing pipeline (#3944)

e98abaf

Create CPU and GPu Java publishing pipelines. Final jars are tested on all platforms. However, signing and publishing to maven are manual steps.

Change group id to com.microsoft.onnxruntime per requirements.

a15c04d

Java GPu artifact naming (#4179)

12060d1

Modify gradle build so artifactID has _gpu for GPU builds. Pass USE_CUDA flag on CUDA build Adjust publishing pipelines to extract POM from a correct path. Co-Authored-By: @Craigacp

bump up ORT version to 1.3.1 (#4181)

c0d39a5

move back to toolset 14.16 to possibly work around nvcc bug (#4180)

cc9c639

stevenlix requested review from faxu, fs-eire, jywu-msft, prasanthpul, snnn and yuslepukhin June 11, 2020 00:46

stevenlix requested a review from a team as a code owner June 11, 2020 00:46

yuslepukhin previously approved these changes Jun 11, 2020

View reviewed changes

KeDengMS and others added 4 commits June 10, 2020 21:52

Symbolic shape inference exit on models without onnx opset used (#4090)

bb2e83c

* Symbolic shape inference exit on models without onnx opset used * Temporary fix for ConvTranspose with symbolic input dims Co-authored-by: Changming Sun <me@sunchangming.com>

Fix Nuphar test failure

561a295

Temporarily disable windows static analysis CI job

0f44843

stevenlix dismissed yuslepukhin’s stale review via 0f44843 June 11, 2020 05:06

snnn reviewed Jun 11, 2020

View reviewed changes

skip model coreml_Imputer-LogisticRegression_sklearn_load_breast_cancer

7de4b29

stevenlix closed this Jun 11, 2020

stevenlix reopened this Jun 11, 2020

yuslepukhin previously approved these changes Jun 12, 2020

View reviewed changes

Delete unused variable

1597ec5

stevenlix dismissed yuslepukhin’s stale review via 1597ec5 June 12, 2020 03:13

yuslepukhin approved these changes Jun 12, 2020

View reviewed changes

pranavsharma approved these changes Jun 12, 2020

View reviewed changes

stevenlix merged commit 814638c into rel-1.3.1 Jun 12, 2020

stevenlix deleted the fs-eire/1.3.1 branch June 12, 2020 18:27

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Cherry pick PRs to Rel-1.3.1#4198

Cherry pick PRs to Rel-1.3.1#4198
stevenlix merged 21 commits intorel-1.3.1from
fs-eire/1.3.1

stevenlix commented Jun 11, 2020 •

edited

Loading

Uh oh!

yuslepukhin left a comment

Uh oh!

snnn commented Jun 11, 2020 •

edited

Loading

Uh oh!

stevenlix commented Jun 11, 2020

Uh oh!

snnn Jun 11, 2020

Uh oh!

stevenlix Jun 11, 2020

Uh oh!

yuslepukhin left a comment

Uh oh!

yuslepukhin left a comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

Conversation

stevenlix commented Jun 11, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

yuslepukhin left a comment

Choose a reason for hiding this comment

Uh oh!

snnn commented Jun 11, 2020 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

stevenlix commented Jun 11, 2020

Uh oh!

snnn Jun 11, 2020

Choose a reason for hiding this comment

Uh oh!

stevenlix Jun 11, 2020

Choose a reason for hiding this comment

Uh oh!

yuslepukhin left a comment

Choose a reason for hiding this comment

Uh oh!

yuslepukhin left a comment

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

stevenlix commented Jun 11, 2020 •

edited

Loading

snnn commented Jun 11, 2020 •

edited

Loading