Rewrite all examples using standardization

.github/workflows/linux-cpu-x64-build.yml

@@ -96,7 +96,7 @@ jobs:
           cmake --build --preset linux_gcc_cpu_release
           cmake --build --preset linux_gcc_cpu_release --target PyPackageBuild
 
-      - name: Install the python wheel and test dependencies
+      - name: Install the Python wheel and test dependencies
         run: |
           python3 -m pip install -r test/python/requirements.txt --user
           python3 -m pip install -r test/python/cpu/torch/requirements.txt --user
@@ -110,9 +110,14 @@ jobs:
           ls -l ${{ github.workspace }}/build/cpu
           ls -l ${{ github.workspace }}/build/cpu/wheel
 
+      - name: Build the Java API and Run the Java Tests
+        run: |
+          set -e -x
+          python3 build.py --config=Release --build_dir build/cpu --build_java --parallel --cmake_generator "Ninja"
+
       # This will also download all the test models to the test/test_models directory
       # These models are used by the python tests as well as C#, C++ and others.
-      - name: Run the python tests
+      - name: Run the Python tests
         run: |
           export ORTGENAI_LOG_ORT_LIB=1
           python3 test/python/test_onnxruntime_genai.py --cwd test/python --test_models test/test_models
@@ -123,10 +128,19 @@ jobs:
           cd test/csharp
           dotnet test /p:Configuration=Release /p:NativeBuildOutputDir="../../build/cpu/" /p:OrtLibDir="../../ort/lib/" --verbosity normal
 
-      - name: Build the Java API and Run the Java Tests
+      - name: Build the C# Examples
         run: |
-          set -e -x
-          python3 build.py --config=Release --build_dir build/cpu --build_java --parallel --cmake_generator "Ninja"
+          export ORTGENAI_LOG_ORT_LIB=1
+          cd examples/csharp/ModelChat
+          dotnet build -c Release
+          cd ../ModelMM
+          dotnet build -c Release
+
+      - name: Test the C# LLM Example with Tool Calling
+        run: |
+          export ORTGENAI_LOG_ORT_LIB=1
+          python3 test/python/special_tokens.py -p test/test_models/qwen-2.5-0.5b/int4/cpu/tokenizer.json -s "<tool_call>" -e "</tool_call>"
+          ./examples/csharp/ModelChat/bin/Release/net8.0/ModelChat -m test/test_models/qwen-2.5-0.5b/int4/cpu/ -e cpu --response_format lark_grammar --tools_file test/test_models/tool-definitions/weather.json --tool_call_start "<tool_call>" --tool_call_end "</tool_call>" --user_prompt "What is the weather in Redmond, WA?" --tool_output --non_interactive --verbose
 
       - name: Run tests
         run: |

.github/workflows/linux-cpu-x64-nightly-build.yml

@@ -66,7 +66,7 @@ jobs:
       - name: Run Q&A Example
         run: |
           python3 -m onnxruntime_genai.models.builder -i /data/ortgenai/pytorch/qwen2.5-0.5b-instruct -e cpu -p int4 -o ./example-models/qwen2.5-0.5b-instruct
-          python3 examples/python/model-qa.py -m ./example-models/qwen2.5-0.5b-instruct -e cpu --input_prompt "what is 10+4?" > output.log 2>&1
+          python3 examples/python/model-qa.py -m ./example-models/qwen2.5-0.5b-instruct -e cpu --user_prompt "what is 10+4?" --non_interactive > output.log 2>&1
           if cat output.log | grep -Eq "14|fourteen"; then
             echo "Result seems correct"
           else

.github/workflows/linux-gpu-x64-build.yml

@@ -112,7 +112,7 @@ jobs:
             --multiple_repos \
             --repository onnxruntimecudabuildx64
 
-      - name: Config with Cmake in Docker
+      - name: Config with CMake in Docker
         run: |
           set -e -x
           docker run \
@@ -125,7 +125,7 @@ jobs:
                 -DMANYLINUX=ON \
                 -DPYTHON_EXECUTABLE=${{ env.PYTHON_EXECUTABLE }} "
 
-      - name: Build with Cmake in Docker
+      - name: Build with CMake in Docker
         run: |
           set -e -x
           docker run \
@@ -136,7 +136,23 @@ jobs:
             bash -c " \
               /usr/bin/cmake --build --preset linux_gcc_cuda_release && /usr/bin/cmake --build --preset linux_gcc_cuda_release --target PyPackageBuild"
 
-      - name: Install the onnxruntime-genai Python wheel and run python test
+      - name: Build the Java API and Run the Java Tests in Docker
+        run: |
+          set -e -x
+          docker run \
+            --gpus all \
+            --rm \
+            --user 0 \
+            --volume $GITHUB_WORKSPACE:/ort_genai_src \
+            -w /ort_genai_src onnxruntimecudabuildx64 bash -c " \
+              alias python3=${{ env.PYTHON_EXECUTABLE }} && \
+              dnf -y update && dnf install -y python3.11-devel && dnf install -y python3-pip python3-setuptools python3-wheel && \
+              ${{ env.PYTHON_EXECUTABLE }} -m pip install -r test/python/requirements.txt --user && \
+              ${{ env.PYTHON_EXECUTABLE }} -m pip install -r test/python/cuda/torch/requirements.txt --user && \
+              ${{ env.PYTHON_EXECUTABLE }} -m pip install -r test/python/cuda/ort/requirements.txt --user && \
+              ${{ env.PYTHON_EXECUTABLE }} build.py --config=Release --build_dir build/cuda --build_java --parallel --cmake_generator Ninja --cmake_extra_defines PYTHON_EXECUTABLE=${{ env.PYTHON_EXECUTABLE }}"
+
+      - name: Install the onnxruntime-genai Python wheel and run Python tests
         run: |
           echo "Installing the onnxruntime-genai Python wheel and running the Python tests"
           docker run \

.github/workflows/mac-cpu-arm64-build.yml

@@ -108,7 +108,7 @@ jobs:
           cmake --build --preset macos_arm64_cpu_release --target PyPackageBuild
         continue-on-error: false
 
-      - name: Install the python wheel and test dependencies
+      - name: Install the Python wheel and test dependencies
         run: |
           python3 -m venv genai-macos-venv
           source genai-macos-venv/bin/activate
@@ -117,6 +117,12 @@ jobs:
           python3 -m pip install -r test/python/macos/ort/requirements.txt
           python3 -m pip install build/cpu/osx-arm64/wheel/onnxruntime_genai*.whl --no-deps
 
+      - name: Build the Java API and Run the Java Tests
+        run: |
+          set -e -x
+          source genai-macos-venv/bin/activate
+          python3 build.py --config=Release --build_dir build/cpu/osx-arm64 --build_java --parallel --cmake_generator "Unix Makefiles" --macos MacOSX --osx_arch arm64 --apple_deploy_target 12.0 --apple_sysroot macosx
+
       - name: Remove the ort lib and header files
         run: |
           rm -rf ort
@@ -130,7 +136,7 @@ jobs:
 
       # This will also download all the test models to the test/test_models directory
       # These models are used by the python tests as well as C#, C++ and others.
-      - name: Run the python tests
+      - name: Run the Python tests
         run: |
           source genai-macos-venv/bin/activate
           export HF_TOKEN="12345"
@@ -144,11 +150,19 @@ jobs:
           cd test/csharp
           dotnet test /p:Configuration=Release /p:NativeBuildOutputDir="../../build/cpu/osx-arm64" --verbosity normal
 
-      - name: Build the Java API and Run the Java Tests
+      - name: Build the C# Examples
         run: |
-          set -e -x
-          source genai-macos-venv/bin/activate
-          python3 build.py --config=Release --build_dir build/cpu/osx-arm64 --build_java --parallel --cmake_generator "Unix Makefiles" --macos MacOSX --osx_arch arm64 --apple_deploy_target 12.0 --apple_sysroot macosx
+          export ORTGENAI_LOG_ORT_LIB=1
+          cd examples/csharp/ModelChat
+          dotnet build -c Release
+          cd ../ModelMM
+          dotnet build -c Release
+
+      - name: Test the C# LLM Example with Tool Calling
+        run: |
+          export ORTGENAI_LOG_ORT_LIB=1
+          python3 test/python/special_tokens.py -p test/test_models/qwen-2.5-0.5b/int4/cpu/tokenizer.json -s "<tool_call>" -e "</tool_call>"
+          ./examples/csharp/ModelChat/bin/Release/net8.0/ModelChat -m test/test_models/qwen-2.5-0.5b/int4/cpu/ -e cpu --response_format lark_grammar --tools_file test/test_models/tool-definitions/weather.json --tool_call_start "<tool_call>" --tool_call_end "</tool_call>" --user_prompt "What is the weather in Redmond, WA?" --tool_output --non_interactive --verbose
 
       - name: Run tests
         run: |

.github/workflows/win-cpu-arm64-build.yml

@@ -93,10 +93,16 @@ jobs:
         run: |
           # Uninstalling LLVM/Clang as it is no longer required and causes issues with numpy installation
           choco uninstall llvm --yes
-          python -m pip install "numpy<2" coloredlogs flatbuffers packaging protobuf sympy pytest
+          python -m pip install -r test\python\requirements.txt --user
+          python -m pip install -r test\python\cpu\torch\requirements.txt --user
+          python -m pip install -r test\python\cpu\ort\requirements.txt --user
           python -m pip install onnxruntime-qnn
           python -m pip install (Get-ChildItem ("$env:binaryDir\wheel\*.whl")) --no-deps
 
+      - name: Build the Java API and Run the Java Tests
+        run: |
+          python build.py --config=Release --build_dir $env:binaryDir --build_java --parallel
+
       - name: Run the Python Tests
         run: |
           python test/python/test_onnxruntime_genai.py --cwd "test\python" --test_models "test\test_models"
@@ -106,9 +112,17 @@ jobs:
           cd test\csharp
           dotnet test /p:NativeBuildOutputDir="$env:GITHUB_WORKSPACE\$env:binaryDir\Release" /p:OrtLibDir="$env:GITHUB_WORKSPACE\ort\lib"
 
-      - name: Build the Java API and Run the Java Tests
+      - name: Build the C# Examples
         run: |
-          python build.py --config=Release --build_dir $env:binaryDir --build_java --parallel
+          cd examples\csharp\ModelChat
+          dotnet build -c Release
+          cd ..\ModelMM
+          dotnet build -c Release
+
+      - name: Test the C# LLM Example with Tool Calling
+        run: |
+          python3 test\python\special_tokens.py -p test\test_models\qwen-2.5-0.5b\int4\cpu\tokenizer.json -s "<tool_call>" -e "</tool_call>"
+          .\examples\csharp\ModelChat\bin\Release\net8.0\ModelChat.exe -m test\test_models\qwen-2.5-0.5b\int4\cpu\ -e cpu --response_format lark_grammar --tools_file test\test_models\tool-definitions\weather.json --tool_call_start "<tool_call>" --tool_call_end "</tool_call>" --user_prompt "What is the weather in Redmond, WA?" --tool_output --non_interactive --verbose
 
       - name: Verify Build Artifacts
         if: always()

.github/workflows/win-cpu-x64-build.yml

@@ -102,13 +102,17 @@ jobs:
           cmake --build --preset windows_x64_cpu_release --parallel
           cmake --build --preset windows_x64_cpu_release --target PyPackageBuild
 
-      - name: Install the python wheel and test dependencies
+      - name: Install the Python wheel and test dependencies
         run: |
           python3 -m pip install -r test\python\requirements.txt --user
           python3 -m pip install -r test\python\cpu\torch\requirements.txt --user
           python3 -m pip install -r test\python\cpu\ort\requirements.txt --user
           python3 -m pip install (Get-ChildItem ("$env:binaryDir\wheel\*.whl")) --no-deps
 
+      - name: Build the Java API and Run the Java Tests
+        run: |
+          python3 build.py --config=Release --build_dir $env:binaryDir --build_java --parallel
+
       - name: Run the Python Tests
         run: |
           python test/python/test_onnxruntime_genai.py --cwd "test\python" --test_models "test\test_models"
@@ -118,9 +122,17 @@ jobs:
           cd test\csharp
           dotnet test /p:NativeBuildOutputDir="$env:GITHUB_WORKSPACE\$env:binaryDir\Release" /p:OrtLibDir="$env:GITHUB_WORKSPACE\ort\lib" --verbosity normal
 
-      - name: Build the Java API and Run the Java Tests
+      - name: Build the C# Examples
         run: |
-          python3 build.py --config=Release --build_dir $env:binaryDir --build_java --parallel
+          cd examples\csharp\ModelChat
+          dotnet build -c Release
+          cd ..\ModelMM
+          dotnet build -c Release
+
+      - name: Test the C# LLM Example with Tool Calling
+        run: |
+          python3 test\python\special_tokens.py -p test\test_models\qwen-2.5-0.5b\int4\cpu\tokenizer.json -s "<tool_call>" -e "</tool_call>"
+          .\examples\csharp\ModelChat\bin\Release\net8.0\ModelChat.exe -m test\test_models\qwen-2.5-0.5b\int4\cpu\ -e cpu --response_format lark_grammar --tools_file test\test_models\tool-definitions\weather.json --tool_call_start "<tool_call>" --tool_call_end "</tool_call>" --user_prompt "What is the weather in Redmond, WA?" --tool_output --non_interactive --verbose
 
       - name: Verify Build Artifacts
         if: always()

.github/workflows/win-cuda-x64-build.yml

@@ -98,6 +98,10 @@ jobs:
           python -m pip install -r test\python\cuda\ort\requirements.txt
           python -m pip install (Get-ChildItem ("$env:binaryDir\wheel\*.whl")) --no-deps
 
+      - name: Build the Java API and Run the Java Tests
+        run: |
+          python build.py --config=Release --build_dir $env:binaryDir --build_java --parallel
+
       - name: Run the Python Tests
         run: |
           python test/python/test_onnxruntime_genai.py --cwd "test\python" --test_models "test\test_models" --e2e
@@ -115,6 +119,18 @@ jobs:
           cd test\csharp
           dotnet test /p:Configuration=release /p:NativeBuildOutputDir="$env:GITHUB_WORKSPACE\$env:binaryDir\Release" /p:OrtLibDir="$env:GITHUB_WORKSPACE\ort\lib"
 
+      - name: Build the C# Examples
+        run: |
+          cd examples\csharp\ModelChat
+          dotnet build -c Release
+          cd ..\ModelMM
+          dotnet build -c Release
+
+      - name: Test the C# LLM Example with Tool Calling
+        run: |
+          python test\python\special_tokens.py -p test\test_models\qwen-2.5-0.5b\int4\cpu\tokenizer.json -s "<tool_call>" -e "</tool_call>"
+          .\examples\csharp\ModelChat\bin\Release\net8.0\ModelChat.exe -m test\test_models\qwen-2.5-0.5b\int4\cpu\ -e cpu --response_format lark_grammar --tools_file test\test_models\tool-definitions\weather.json --tool_call_start "<tool_call>" --tool_call_end "</tool_call>" --user_prompt "What is the weather in Redmond, WA?" --tool_output --non_interactive --verbose
+
       - name: Prepend CUDA to PATH and Run tests
         run: |-
           $env:PATH = "${{ env.cuda_dir }}\\v${{ env.cuda_version }}\\bin;" + $env:PATH 

.github/workflows/win-directml-x64-build.yml

@@ -114,6 +114,10 @@ jobs:
           python -m pip install -r test\python\directml\ort\requirements.txt
           python -m pip install (Get-ChildItem ("$env:binaryDir\wheel\*.whl")) --no-deps
 
+      - name: Build the Java API and Run the Java Tests
+        run: |
+          python build.py --config=Release --build_dir $env:binaryDir --build_java --parallel
+
       - name: Run the Python Tests
         run: |
           python test/python/test_onnxruntime_genai.py --cwd "test\python" --test_models "test\test_models" --e2e

.gitignore

@@ -27,7 +27,7 @@ models_outputs_cpu
 benchmark/python/output
 examples/python/genai_models
 examples/python/hf_cache
-examples/csharp/HelloPhi/models
+examples/csharp/ModelChat/models
 
 !test/test_models/hf-internal-testing/
 !test/test_models/hf-internal-testing/tiny-random-gpt2*/*.onnx

.pipelines/stages/jobs/custom-nuget-packaging-job.yml

@@ -167,4 +167,3 @@ jobs:
     inputs:
       targetPath: '$(Build.ArtifactStagingDirectory)\nuget'
       artifactName: $(genai_nuget_package_name)
-

.pipelines/stages/jobs/nuget-validation-job.yml

@@ -98,6 +98,16 @@ jobs:
     ${{ else }}:
       value: 'cpu_and_mobile/cpu-int4-rtn-block-32-acc-level-4'
 
+  - name: prebuild_phi4_mm_model_folder
+    ${{ if eq(parameters.ep, 'cpu') }}:
+      value: 'cpu_and_mobile/cpu-int4-rtn-block-32-acc-level-4'
+    ${{ elseif eq(parameters.ep, 'cuda') }}:
+      value: 'gpu/gpu-int4-rtn-block-32'
+    ${{ elseif eq(parameters.ep, 'directml')}}:
+      value: 'gpu/gpu-int4-rtn-block-32'
+    ${{ else }}:
+      value: 'cpu_and_mobile/cpu-int4-rtn-block-32-acc-level-4'
+
   - name: cuda_docker_image
     ${{ if eq(parameters.cuda_version, '11.8') }}:
       value: onnxruntimebuildcache.azurecr.io/internal/azureml/onnxruntime/build/cuda11_x64_almalinux8_gcc11:20250714.2
@@ -143,14 +153,14 @@ jobs:
       HuggingFaceRepo: 'microsoft/Phi-3-mini-4k-instruct-onnx'
       LocalFolder: 'phi3-mini'
       RepoFolder: $(prebuild_phi3_mini_model_folder)
-      WorkingDirectory: '$(Build.Repository.LocalPath)/examples/csharp/HelloPhi'
+      WorkingDirectory: '$(Build.Repository.LocalPath)/examples/csharp/ModelChat'
       HuggingFaceToken: $(HF_TOKEN)
       os: ${{ parameters.os }}
 
   - template: steps/nuget-validation-step.yml
     parameters:
-      CsprojFolder: "examples/csharp/HelloPhi"
-      CsprojName: "HelloPhi"
+      CsprojFolder: "examples/csharp/ModelChat"
+      CsprojName: "ModelChat"
       CsprojConfiguration: $(csproj_configuration)
       LocalFolder: 'phi3-mini'
       ModelFolder: $(prebuild_phi3_mini_model_folder)
@@ -160,14 +170,14 @@ jobs:
       HuggingFaceRepo: 'microsoft/Phi-3.5-vision-instruct-onnx'
       LocalFolder: 'phi3.5-vision'
       RepoFolder: $(prebuild_phi3_5_vision_model_folder)
-      WorkingDirectory: '$(Build.Repository.LocalPath)/examples/csharp/HelloPhi3V'
+      WorkingDirectory: '$(Build.Repository.LocalPath)/examples/csharp/ModelMM'
       HuggingFaceToken: $(HF_TOKEN)
       os: ${{ parameters.os }}
 
   - template: steps/nuget-validation-step.yml
     parameters:
-      CsprojFolder: "examples/csharp/HelloPhi3V"
-      CsprojName: "HelloPhi3V"
+      CsprojFolder: "examples/csharp/ModelMM"
+      CsprojName: "ModelMM"
       CsprojConfiguration: $(csproj_configuration)
       LocalFolder: 'phi3.5-vision'
       ModelFolder: $(prebuild_phi3_5_vision_model_folder)
@@ -177,14 +187,14 @@ jobs:
       HuggingFaceRepo: 'microsoft/Phi-4-multimodal-instruct-onnx'
       LocalFolder: 'phi4-mm'
       RepoFolder: $(prebuild_phi4_mm_model_folder)
-      WorkingDirectory: '$(Build.Repository.LocalPath)/examples/csharp/HelloPhi4MM'
+      WorkingDirectory: '$(Build.Repository.LocalPath)/examples/csharp/ModelMM'
       HuggingFaceToken: $(HF_TOKEN)
       os: ${{ parameters.os }}
 
   - template: steps/nuget-validation-step.yml
     parameters:
-      CsprojFolder: "examples/csharp/HelloPhi4MM"
-      CsprojName: "HelloPhi4MM"
+      CsprojFolder: "examples/csharp/ModelMM"
+      CsprojName: "ModelMM"
       CsprojConfiguration: $(csproj_configuration)
       LocalFolder: 'phi4-mm'
       ModelFolder: $(prebuild_phi4_mm_model_folder)

.pipelines/stages/jobs/py-validation-job.yml

@@ -211,7 +211,7 @@ jobs:
   - template: steps/python-validation-step.yml
     parameters:
       PythonScriptFolder: "examples/python"
-      PythonScriptName: "phi4-mm.py"
+      PythonScriptName: "model-mm.py"
       LocalFolder: 'phi4-mm'
       ModelFolder: $(prebuild_phi4_mm_model_folder)