docs(build): add more example (#2378)

* docs(build): add more example * Update linux-x86_64.md * Update linux-x86_64.md * docs(format): update install doc * Update linux-x86_64.md * Update linux-x86_64.md * Update lint.yml * fix lint --------- Co-authored-by: RunningLeon <mnsheng@yeah.net>
open-mmlab · Sep 5, 2023 · 468c423 · 468c423
1 parent 58db0ad
commit 468c423
Show file tree

Hide file tree

Showing 3 changed files with 58 additions and 2 deletions.
diff --git a/.github/workflows/lint.yml b/.github/workflows/lint.yml
@@ -16,7 +16,9 @@ jobs:
           python -m pip install pre-commit
           pre-commit install
       - name: Linting
-        run: pre-commit run --all-files
+        run: |
+          pre-commit run --all-files
+          git diff
       - name: Format c/cuda codes with clang-format
         uses: DoozyX/clang-format-lint-action@v0.11
         with:

diff --git a/docs/en/01-how-to-build/linux-x86_64.md b/docs/en/01-how-to-build/linux-x86_64.md
@@ -395,3 +395,30 @@ You can also activate other engines after the model.
 
   make -j$(nproc) && make install
   ```
+
+- cuda + TensorRT + onnxruntime + openvino + ncnn
+
+  If the [ncnn auto-install script](../../../tools/scripts/build_ubuntu_x64_ncnn.py) is used, protobuf will be installed in mmdeploy-dep/pbinstall in the same directory as mmdeploy.
+
+  ```Bash
+  export PROTO_DIR=/path/to/mmdeploy-dep/pbinstall
+  cmake .. \
+      -DCMAKE_CXX_COMPILER=g++-7 \
+      -DMMDEPLOY_BUILD_SDK=ON \
+      -DMMDEPLOY_BUILD_EXAMPLES=ON \
+      -DMMDEPLOY_BUILD_SDK_PYTHON_API=ON \
+      -DMMDEPLOY_TARGET_DEVICES="cuda;cpu" \
+      -DMMDEPLOY_TARGET_BACKENDS="trt;ort;ncnn;openvino" \
+      -Dpplcv_DIR=${PPLCV_DIR}/cuda-build/install/lib/cmake/ppl \
+      -DTENSORRT_DIR=${TENSORRT_DIR} \
+      -DCUDNN_DIR=${CUDNN_DIR} \
+      -DONNXRUNTIME_DIR=${ONNXRUNTIME_DIR} \
+      -DInferenceEngine_DIR=${OPENVINO_DIR}/runtime/cmake \
+      -Dncnn_DIR=${NCNN_DIR}/build/install/lib/cmake/ncnn \
+      -DProtobuf_LIBRARIES=${PROTO_DIR}/lib/libprotobuf.so \
+      -DProtobuf_PROTOC_EXECUTABLE=${PROTO_DIR}/bin/protoc \
+      -DProtobuf_INCLUDE_DIR=${PROTO_DIR}/pbinstall/include
+  ```
+
+```
+```
diff --git a/docs/zh_cn/01-how-to-build/linux-x86_64.md b/docs/zh_cn/01-how-to-build/linux-x86_64.md
@@ -335,7 +335,7 @@ mim install -e .
 
 #### 编译 SDK 和 Demos
 
-下文展示2个构建SDK的样例，分别用 ONNXRuntime 和 TensorRT 作为推理引擎。您可以参考它们，激活其他的推理引擎。
+下文展示一些构建 SDK 的样例。您可以参考它们，激活其他的推理引擎。
 
 - cpu + ONNXRuntime
 
@@ -390,3 +390,30 @@ mim install -e .
 
   make -j$(nproc) && make install
   ```
+
+- cuda + TensorRT + onnxruntime + openvino + ncnn
+
+  如果使用了 [ncnn 自动安装脚本](../../../tools/scripts/build_ubuntu_x64_ncnn.py), protobuf 会安装在 mmdeploy 同级目录的 mmdeploy-dep/pbinstall 中。
+
+  ```Bash
+  export PROTO_DIR=/path/to/mmdeploy-dep/pbinstall
+  cmake .. \
+      -DCMAKE_CXX_COMPILER=g++-7 \
+      -DMMDEPLOY_BUILD_SDK=ON \
+      -DMMDEPLOY_BUILD_EXAMPLES=ON \
+      -DMMDEPLOY_BUILD_SDK_PYTHON_API=ON \
+      -DMMDEPLOY_TARGET_DEVICES="cuda;cpu" \
+      -DMMDEPLOY_TARGET_BACKENDS="trt;ort;ncnn;openvino" \
+      -Dpplcv_DIR=${PPLCV_DIR}/cuda-build/install/lib/cmake/ppl \
+      -DTENSORRT_DIR=${TENSORRT_DIR} \
+      -DCUDNN_DIR=${CUDNN_DIR} \
+      -DONNXRUNTIME_DIR=${ONNXRUNTIME_DIR} \
+      -DInferenceEngine_DIR=${OPENVINO_DIR}/runtime/cmake \
+      -Dncnn_DIR=${NCNN_DIR}/build/install/lib/cmake/ncnn \
+      -DProtobuf_LIBRARIES=${PROTO_DIR}/lib/libprotobuf.so \
+      -DProtobuf_PROTOC_EXECUTABLE=${PROTO_DIR}/bin/protoc \
+      -DProtobuf_INCLUDE_DIR=${PROTO_DIR}/pbinstall/include
+  ```
+
+```
+```