ORT 1.19.0 Release: Cherry-Pick Round 2

js/react_native/android/build.gradle

@@ -3,7 +3,7 @@ import java.nio.file.Paths
 buildscript {
   repositories {
     google()
-    jcenter()
+    mavenCentral()
   }
 
   dependencies {
@@ -145,7 +145,6 @@ android {
 
 repositories {
   mavenCentral()
-  jcenter()
   google()
 
   def found = false

js/react_native/e2e/android/build.gradle

@@ -10,7 +10,7 @@ buildscript {
     }
     repositories {
         google()
-        jcenter()
+        mavenCentral()
     }
     dependencies {
         classpath('com.android.tools.build:gradle:7.1.1')
@@ -31,13 +31,13 @@ allprojects {
             // Android JSC is installed from npm
             url("$rootDir/../node_modules/jsc-android/dist")
         }
-        maven { 
+        maven {
           // Add Detox as a precompiled native dependency
           url("$rootDir/../node_modules/detox/Detox-android")
         }
 
         google()
-        jcenter()
+        mavenCentral()
         maven { url 'https://www.jitpack.io' }
     }
 }

onnxruntime/contrib_ops/cpu/quantization/matmul_nbits.cc

@@ -105,17 +105,15 @@ class MatMulNBits final : public OpKernel {
     ORT_ENFORCE(nbits_ == 4,
                 "Only 4b quantization is supported for MatMulNBits op, additional bits support is planned.");
     const Tensor* tensor_zero_point = nullptr;
-    has_zp_input_ = info.TryGetConstantInput(3, &tensor_zero_point);
+    has_zp_input_ = info.TryGetConstantInput(InputIndex::zero_points, &tensor_zero_point);
 #ifdef ORT_NEURAL_SPEED
     const Tensor* tensor_B = nullptr;
     const Tensor* tensor_scale = nullptr;
-    const Tensor* tensor_zero_point = nullptr;
     bool B_constant = info.TryGetConstantInput(InputIndex::B, &tensor_B);
     bool scale_constant = info.TryGetConstantInput(InputIndex::scales, &tensor_scale);
-    bool zero_point_constant = info.TryGetConstantInput(InputIndex::zero_points, &tensor_zero_point);
     is_asym_ = zero_point_arg != nullptr;
     all_constant_ = B_constant && scale_constant;
-    all_constant_ = is_asym_ ? all_constant_ && zero_point_constant : all_constant_;
+    all_constant_ = is_asym_ ? all_constant_ && has_zp_input_ : all_constant_;
 #endif
   }
 

tools/ci_build/github/azure-pipelines/bigmodels-ci-pipeline.yml

@@ -282,6 +282,7 @@ stages:
 - stage: Llama2_7B_ONNX
   dependsOn:
   - Build_Onnxruntime_Cuda
+  condition: and (succeeded(), or(eq(variables['Build.SourceBranch'], 'refs/heads/main'), startsWith(variables['Build.SourceBranch'], 'refs/heads/rel-')))
   jobs:
   - job: Llama2_7B_ONNX
     timeoutInMinutes: 120

tools/ci_build/github/azure-pipelines/orttraining-linux-gpu-ci-pipeline.yml

@@ -48,7 +48,7 @@ jobs:
     RunInjectedPipeline: 'true'
     InjectedPipeline: 'orttraining-linux-gpu-test-ci-pipeline.yml'
     DockerImageTag: 'onnxruntime_orttraining_ortmodule_tests_image'
-    TimeoutInMinutes: 140
+    TimeoutInMinutes: 150
     # Enable unreleased onnx opsets in CI builds
     # This facilitates testing the implementation for the new opsets
     AllowReleasedOpsetOnly: '0'

tools/ci_build/github/azure-pipelines/orttraining-py-packaging-pipeline-cuda.yml

@@ -18,7 +18,7 @@ stages:
     torch_version: '2.0.0'
     opset_version: '17'
     cuda_version: '11.8'
-    cmake_cuda_architectures: 60;61;70;75;80;86
+    cmake_cuda_architectures: 70;75;80;86
     docker_file: Dockerfile.manylinux2_28_training_cuda11_8
     agent_pool: Onnxruntime-Linux-GPU
     upload_wheel: 'yes'

tools/ci_build/github/azure-pipelines/orttraining-py-packaging-pipeline-cuda12.yml

@@ -8,7 +8,7 @@ stages:
     torch_version: '2.1.0'
     opset_version: '17'
     cuda_version: '12.2'
-    cmake_cuda_architectures: 70;75;80;86;90
+    cmake_cuda_architectures: 80;86;90
     docker_file: Dockerfile.manylinux2_28_training_cuda12_2
     agent_pool: Onnxruntime-Linux-GPU
     upload_wheel: 'yes'

tools/ci_build/github/azure-pipelines/templates/py-packaging-stage.yml

@@ -516,7 +516,7 @@ stages:
     - stage: Python_Packaging_Windows_x64_QNN
       dependsOn: []
       jobs:
-        - template: py-win-x64-qnn.yml
+        - template: py-win-arm64ec-qnn.yml
           parameters:
             MACHINE_POOL: 'Onnxruntime-QNNEP-Windows-2022-CPU'
             QNN_SDK: ${{ parameters.qnn_sdk_version }}

tools/ci_build/github/azure-pipelines/templates/py-win-arm64ec-qnn.yml

@@ -0,0 +1,165 @@
+parameters:
+
+- name: MACHINE_POOL
+  type: string
+  default: 'Onnxruntime-QNNEP-Windows-2022-CPU'
+
+- name: QNN_SDK
+  displayName: QNN SDK Version
+  type: string
+  default: 2.24.0.240626
+
+- name: ENV_SETUP_SCRIPT
+  type: string
+  default: ''
+
+- name: BUILD_PY_PARAMETERS
+  displayName: >
+    Extra parameters to pass to build.py. Don't put newlines in here.
+  type: string
+  default: ''
+
+jobs:
+- job: Win_py_x64_qnn_Wheels
+  timeoutInMinutes: 210
+  workspace:
+    clean: all
+  pool:
+    name: ${{ parameters.MACHINE_POOL }}
+  strategy:
+    matrix:
+      Python38_x64:
+        PythonVersion: '3.8'
+      Python39_x64:
+        PythonVersion: '3.9'
+      Python310_x64:
+        PythonVersion: '3.10'
+      Python311_x64:
+        PythonVersion: '3.11'
+      Python312_x64:
+        PythonVersion: '3.12'
+  variables:
+    GRADLE_OPTS: '-Dorg.gradle.daemon=false'
+    VSGenerator: 'Visual Studio 17 2022'
+  steps:
+      - checkout: self
+        clean: true
+        submodules: recursive
+
+      - template: telemetry-steps.yml
+
+      - task: UsePythonVersion@0
+        inputs:
+          versionSpec: $(PythonVersion)
+          addToPath: true
+          architecture: 'x64'
+
+      - task: onebranch.pipeline.tsaoptions@1
+        displayName: 'OneBranch TSAOptions'
+        inputs:
+          tsaConfigFilePath: '$(Build.SourcesDirectory)\.config\tsaoptions.json'
+          appendSourceBranchName: false
+
+      - template: download-deps.yml
+
+      - task: PythonScript@0
+        displayName: 'Update deps.txt'
+        inputs:
+          scriptPath: $(Build.SourcesDirectory)/tools/ci_build/replace_urls_in_deps.py
+          arguments: --new_dir $(Build.BinariesDirectory)/deps
+          workingDirectory: $(Build.BinariesDirectory)
+
+      - task: PowerShell@2
+        displayName: 'Install ONNX'
+        inputs:
+          filePath: '$(Build.SourcesDirectory)/tools/ci_build/github/windows/install_third_party_deps.ps1'
+          workingDirectory: '$(Build.BinariesDirectory)'
+          arguments: -cpu_arch x64 -install_prefix $(Build.BinariesDirectory)\RelWithDebInfo\installed -build_config RelWithDebInfo
+
+      - template: set-nightly-build-option-variable-step.yml
+
+      - template: jobs/download_win_qnn_sdk.yml
+        parameters:
+          QnnSDKVersion: ${{ parameters.QNN_SDK }}
+
+      - task: PythonScript@0
+        displayName: 'Generate cmake config'
+        inputs:
+          scriptPath: '$(Build.SourcesDirectory)\tools\ci_build\build.py'
+          arguments: >
+            --config RelWithDebInfo
+            --build_dir $(Build.BinariesDirectory)
+            --skip_submodule_sync
+            --cmake_generator "$(VSGenerator)"
+            --use_qnn
+            --qnn_home $(QnnSDKRootDir)
+            --enable_pybind
+            --parallel --update --arm64ec
+            $(TelemetryOption) ${{ parameters.BUILD_PY_PARAMETERS }}
+          workingDirectory: '$(Build.BinariesDirectory)'
+
+      - task: VSBuild@1
+        displayName: 'Build'
+        inputs:
+          solution: '$(Build.BinariesDirectory)\RelWithDebInfo\onnxruntime.sln'
+          platform: 'arm64ec'
+          configuration: RelWithDebInfo
+          msbuildArchitecture: 'x64'
+          maximumCpuCount: true
+          logProjectEvents: true
+          workingFolder: '$(Build.BinariesDirectory)\RelWithDebInfo'
+          createLogFile: true
+
+      # Esrp signing
+      - template: win-esrp-dll.yml
+        parameters:
+          FolderPath: '$(Build.BinariesDirectory)\RelWithDebInfo\RelWithDebInfo\onnxruntime\capi'
+          DisplayName: 'ESRP - Sign Native dlls'
+          DoEsrp: true
+          Pattern: '*.pyd'
+
+      - task: PythonScript@0
+        displayName: 'Build wheel'
+        inputs:
+          scriptPath: '$(Build.SourcesDirectory)\setup.py'
+          arguments: 'bdist_wheel $(NightlyBuildOption) --wheel_name_suffix=qnn'
+          workingDirectory: '$(Build.BinariesDirectory)\RelWithDebInfo\RelWithDebInfo'
+
+      - task: CopyFiles@2
+        displayName: 'Copy Python Wheel to: $(Build.ArtifactStagingDirectory)'
+        inputs:
+          SourceFolder: '$(Build.BinariesDirectory)\RelWithDebInfo\RelWithDebInfo\dist'
+          Contents: '*.whl'
+          TargetFolder: '$(Build.ArtifactStagingDirectory)'
+
+      - task: PublishBuildArtifacts@1
+        displayName: 'Publish Artifact: ONNXRuntime python wheel'
+        inputs:
+          ArtifactName: onnxruntime_qnn
+
+      - script: |
+          7z x *.whl
+        workingDirectory: '$(Build.ArtifactStagingDirectory)'
+        displayName: 'unzip the package'
+
+      - task: CredScan@3
+        displayName: 'Run CredScan'
+        inputs:
+          debugMode: false
+        continueOnError: true
+
+      - task: BinSkim@4
+        displayName: 'Run BinSkim'
+        inputs:
+          AnalyzeTargetGlob: '+:file|$(Build.ArtifactStagingDirectory)\**\*.dll'
+
+      - task: TSAUpload@2
+        displayName: 'TSA upload'
+        condition: and (succeeded(), eq(variables['Build.SourceBranch'], 'refs/heads/main'))
+        inputs:
+          GdnPublishTsaOnboard: false
+          GdnPublishTsaConfigFile: '$(Build.sourcesDirectory)\.gdn\.gdntsa' 
+
+      - template: component-governance-component-detection-steps.yml
+        parameters:
+          condition: 'succeeded'

tools/ci_build/github/linux/docker/scripts/training/ortmodule/stage2/requirements.txt

@@ -8,7 +8,7 @@ rsa==4.9
 tensorboard==2.13.0
 h5py
 wget
-pytorch-lightning
+pytorch-lightning==2.3.3
 deepspeed==0.9.0
 fairscale==0.4.6
 parameterized>=0.8.1