Merge pull request #663 from intel/sync_msft_17_4_25

Backmerging with Msft commits

Author: jatinwadhwa921

.github/workflows/linux_openvino_ci.yml

@@ -0,0 +1,45 @@
+name: Linux OpenVINO CI
+
+on:
+  push:
+    branches: [ main, 'rel-*' ]
+  pull_request:
+    branches: [ main, 'rel-*' ]
+  workflow_dispatch:
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
+permissions:
+  contents: read
+  packages: write # Needed if the reusable workflow pushes images
+  attestations: write # Optional: for artifact attestations if enabled
+  id-token: write # Optional: may be needed for OIDC authentication (e.g., ACR)
+
+jobs:
+  build_test_openvino:
+    name: Build and Test OpenVINO EP (AlamLinux8, Py3.12)
+    # Use the reusable workflow as the other Linux CI pipelines
+    uses: ./.github/workflows/reusable_linux_build.yml
+    with:
+      pool_name: "onnxruntime-github-Ubuntu2204-AMD-CPU"
+      build_config: Release
+      # Architecture: OpenVino only supports Intel X64
+      architecture: x64
+      dockerfile_path: tools/ci_build/github/linux/docker/inference/x86_64/python/openvino/Dockerfile
+      docker_image_repo: onnxruntimeopenvino
+
+      execution_providers: 'openvino'
+
+      extra_build_flags: '--use_openvino CPU --enable_generic_interface --build_shared_lib'
+
+      # Python Path Prefix: Set the correct Python 3.12 path inside the manylinux container
+      python_path_prefix: 'PATH=/opt/python/cp312-cp312/bin:$PATH'
+
+      run_tests: true
+      upload_build_output: false
+
+    # Secrets: Pass the necessary GitHub token
+    secrets:
+      GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}

.github/workflows/publish-python-apidocs.yml

@@ -40,6 +40,7 @@ jobs:
       - name: Generate Python docs with Sphinx
         run: |
           cd tools/doc
+          chmod +x *
           ./builddoc.sh /usr/bin ../.. ../../build
       - name: Log source commit
         run: git rev-parse --short HEAD > build/docs/html/source-version.txt

cgmanifests/cgmanifest.json

@@ -18,7 +18,7 @@
         "maven": {
           "GroupId": "com.google.protobuf",
           "ArtifactId": "protobuf-java",
-          "Version": "3.21.7"
+          "Version": "3.25.5"
         },
         "DevelopmentDependency": true
       }

cmake/CMakeLists.txt

@@ -62,7 +62,7 @@ option(onnxruntime_GENERATE_TEST_REPORTS "Enable test report generation" OFF)
 option(onnxruntime_ENABLE_STATIC_ANALYSIS "Enable static analysis" OFF)
 option(onnxruntime_USE_CUSTOM_STATIC_ANALYSIS_RULES "Use a custom SDL Rule. It is mainly for our CI build" OFF)
 option(onnxruntime_REDIRECT_STATIC_ANALYSIS_OUTPUTS_TO_FILE "Use a custom SDL Rule. It is mainly for our CI build" OFF)
-option(onnxruntime_ENABLE_PYTHON "Enable python buildings" OFF)
+option(onnxruntime_ENABLE_PYTHON "Enable python bindings" OFF)
 # Enable it may cause LNK1169 error
 option(onnxruntime_ENABLE_MEMLEAK_CHECKER "Experimental: Enable memory leak checker in Windows debug build" OFF)
 option(onnxruntime_USE_CUDA "Build with CUDA support" OFF)

docs/python/ReadMeOV.rst

@@ -16,7 +16,7 @@ Requirements
 ^^^^^^^^^^^^
 
 - Ubuntu 18.04, 20.04 or Windows 10 - 64 bit
-- Python 3.11, 3.12 and 3.13 for Windows and Linux
+- Python 3.10, 3.11, 3.12 and 3.13 for Windows and Linux
 
 This package supports:
  - Intel® CPUs
@@ -29,7 +29,7 @@ This package supports:
 Please install OpenVINO™ PyPi Package separately for Windows.
 For installation instructions on Windows please refer to  `OpenVINO™ Execution Provider for ONNX Runtime for Windows <https://github.com/intel/onnxruntime/releases/>`_.
 
-**OpenVINO™ Execution Provider for ONNX Runtime** Linux Wheels comes with pre-built libraries of OpenVINO™ version 2025.0.0 eliminating the need to install OpenVINO™ separately.
+**OpenVINO™ Execution Provider for ONNX Runtime** Linux Wheels comes with pre-built libraries of OpenVINO™ version 2025.1.0 eliminating the need to install OpenVINO™ separately.
 
 For more details on build and installation please refer to `Build <https://onnxruntime.ai/docs/build/eps.html#openvino>`_.
 

include/onnxruntime/core/providers/providers.h

@@ -2,12 +2,31 @@
 // Licensed under the MIT License.
 
 #pragma once
+#include <memory>
+
+struct OrtSessionOptions;
+struct OrtLogger;
 
 namespace onnxruntime {
 class IExecutionProvider;
 
 struct IExecutionProviderFactory {
   virtual ~IExecutionProviderFactory() = default;
   virtual std::unique_ptr<IExecutionProvider> CreateProvider() = 0;
+
+  /// <summary>
+  /// Creates an IExecutionProvider instance. Enables initialization of an EP instance using session-level options
+  /// such as session configs (string key/value pairs), graph optimization level, etc.
+  ///
+  /// The default implementation ignores the arguments and calls the above CreateProvider() function,
+  /// which does not take in any arguments.
+  ///
+  /// This version of CreateProvider() is used by InferenceSession when registering EPs.
+  /// </summary>
+  /// <param name="session_options">Options for the session in which the IExecutionProvider is used</param>
+  /// <param name="session_logger">Session logger that should be used by the IExecutionProvider.</param>
+  /// <returns>An IExecutionProvider</returns>
+  virtual std::unique_ptr<IExecutionProvider> CreateProvider(const OrtSessionOptions& session_options,
+                                                             const OrtLogger& session_logger);
 };
 }  // namespace onnxruntime