apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2025-10-06 17:17:14 +08:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

[Doc] Add docs for ppocr ppseg examples (#1429)

Browse Source 
* add docs for examples

* add english doc

* fix

* fix docs
This commit is contained in:
chenjian
2023-02-28 20:13:01 +08:00
committed by GitHub
parent 010f12db3d
commit c8bcada1a2
22 changed files with 2983 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

184
examples/vision/segmentation/paddleseg/cpu-gpu/c/README.md Executable file
View File

@@ -0,0 +1,184 @@
English | [简体中文](README_CN.md)
# PaddleSeg C Deployment Example
This directory provides `infer.c` to finish the deployment of PaddleSeg on CPU/GPU.
Before deployment, two steps require confirmation
- 1. Software and hardware should meet the requirements. Please refer to [FastDeploy Environment Requirements](../../../../../docs/en/build_and_install/download_prebuilt_libraries.md)
- 2. Download the precompiled deployment library and samples code according to your development environment. Refer to [FastDeploy Precompiled Library](../../../../../docs/en/build_and_install/download_prebuilt_libraries.md)
Taking inference on Linux as an example, the compilation test can be completed by executing the following command in this directory. FastDeploy version 1.0.4 or above (x.x.x>=1.0.4) is required to support this model.
```bash
mkdir build
cd build
# Download the FastDeploy precompiled library. Users can choose your appropriate version in the `FastDeploy Precompiled Library` mentioned above
wget https://bj.bcebos.com/fastdeploy/release/cpp/fastdeploy-linux-x64-x.x.x.tgz
tar xvf fastdeploy-linux-x64-x.x.x.tgz
cmake .. -DFASTDEPLOY_INSTALL_DIR=${PWD}/fastdeploy-linux-x64-x.x.x
make -j
# Download model, image files
wget https://bj.bcebos.com/paddlehub/fastdeploy/PP_LiteSeg_B_STDC2_cityscapes_without_argmax_infer.tgz
tar -xvf PP_LiteSeg_B_STDC2_cityscapes_without_argmax_infer.tgz
wget https://paddleseg.bj.bcebos.com/dygraph/demo/cityscapes_demo.png
# CPU inference
./infer_demo PP_LiteSeg_B_STDC2_cityscapes_without_argmax_infer cityscapes_demo.png 0
# GPU inference
./infer_demo PP_LiteSeg_B_STDC2_cityscapes_without_argmax_infer cityscapes_demo.png 1
```
The above command works for Linux or MacOS. For SDK in Windows, refer to:
- [How to use FastDeploy C++ SDK in Windows](../../../../../docs/en/faq/use_sdk_on_windows.md)
The visualized result after running is as follows
<div align="center">
<img src="https://user-images.githubusercontent.com/16222477/191712880-91ae128d-247a-43e0-b1e3-cafae78431e0.jpg", width=512px, height=256px />
</div>
## PaddleSeg C Interface
### RuntimeOption
```c
FD_C_RuntimeOptionWrapper* FD_C_CreateRuntimeOptionWrapper()
```
> Create a RuntimeOption object, and return a pointer to manipulate it.
>
> **Return**
>
> * **fd_c_runtime_option_wrapper**(FD_C_RuntimeOptionWrapper*): Pointer to manipulate RuntimeOption object.
```c
void FD_C_RuntimeOptionWrapperUseCpu(
FD_C_RuntimeOptionWrapper* fd_c_runtime_option_wrapper)
```
> Enable Cpu inference.
>
> **Params**
>
> * **fd_c_runtime_option_wrapper**(FD_C_RuntimeOptionWrapper*): Pointer to manipulate RuntimeOption object.
```c
void FD_C_RuntimeOptionWrapperUseGpu(
FD_C_RuntimeOptionWrapper* fd_c_runtime_option_wrapper,
int gpu_id)
```
> Enable Gpu inference.
>
> **Params**
>
> * **fd_c_runtime_option_wrapper**(FD_C_RuntimeOptionWrapper*): Pointer to manipulate RuntimeOption object.
> * **gpu_id**(int): gpu id
### Model
```c
FD_C_PaddleSegWrapper* FD_C_CreatePaddleSegWrapper(
const char* model_file, const char* params_file, const char* config_file,
FD_C_RuntimeOptionWrapper* fd_c_runtime_option_wrapper,
const FD_C_ModelFormat model_format
)
```
> Create a PaddleSeg model object, and return a pointer to manipulate it.
>
> **Params**
>
> * **model_file**(const char*): Model file path
> * **params_file**(const char*): Parameter file path
> * **config_file**(const char*): config file path
> * **fd_c_runtime_option_wrapper**(FD_C_RuntimeOptionWrapper*): Backend inference configuration. None by default, which is the default configuration
> * **model_format**(FD_C_ModelFormat): Model format
>
> **Return**
>
> * **fd_c_ppseg_wrapper**(FD_C_PaddleSegWrapper*): Pointer to manipulate PaddleSeg object.
#### Read and write image
```c
FD_C_Mat FD_C_Imread(const char* imgpath)
```
> Read an image, and return a pointer to cv::Mat.
>
> **Params**
>
> * **imgpath**(const char*): image path
>
> **Return**
>
> * **imgmat**(FD_C_Mat): pointer to cv::Mat object which holds the image.
```c
FD_C_Bool FD_C_Imwrite(const char* savepath, FD_C_Mat img);
```
> Write image to a file.
>
> **Params**
>
> * **savepath**(const char*): save path
> * **img**(FD_C_Mat): pointer to cv::Mat object
>
> **Return**
>
> * **result**(FD_C_Bool): bool to indicate success or failure
#### Prediction
```c
FD_C_Bool FD_C_PaddleSegWrapperPredict(
FD_C_PaddleSegWrapper* fd_c_ppseg_wrapper,
FD_C_Mat img,
FD_C_SegmentationResult* result)
```
>
> Predict an image, and generate result.
>
> **Params**
> * **fd_c_ppseg_wrapper**(FD_C_PaddleSegWrapper*): Pointer to manipulate PaddleSeg object.
> * **img**（FD_C_Mat）: pointer to cv::Mat object, which can be obained by FD_C_Imread interface
> * **result**(FD_C_SegmentationResult*): Segmentation prediction results, Refer to [Vision Model Prediction Results](../../../../../docs/api/vision_results/) for SegmentationResult
#### Result
```c
FD_C_Mat FD_C_VisSegmentation(FD_C_Mat im,
FD_C_SegmentationResult* result,
float weight)
```
>
> Visualize segmentation results and return visualization image.
>
> **Params**
> * **im**(FD_C_Mat): pointer to input image
> * **segmentation_result**(FD_C_SegmentationResult*): pointer to C FD_C_SegmentationResult structure
> * **weight**(float): weight transparent weight of visualized result image
>
> **Return**
> * **vis_im**(FD_C_Mat): pointer to visualization image.
## Other Documents
- [PPSegmentation Model Description](../../)
- [PaddleSeg Python Deployment](../python)
- [Model Prediction Results](../../../../../docs/api/vision_results/)
- [How to switch the model inference backend engine](../../../../../docs/cn/faq/how_to_change_backend.md)

185
examples/vision/segmentation/paddleseg/cpu-gpu/c/README_CN.md Normal file
View File

@@ -0,0 +1,185 @@
[English](README.md) | 简体中文
# PaddleSeg C部署示例
本目录下提供`infer.c`来调用C API快速完成PaddleSeg模型在CPU/GPU上部署的示例。
在部署前，需确认以下两个步骤
- 1. 软硬件环境满足要求，参考[FastDeploy环境要求](../../../../../docs/cn/build_and_install/download_prebuilt_libraries.md)
- 2. 根据开发环境，下载预编译部署库和samples代码，参考[FastDeploy预编译库](../../../../../docs/cn/build_and_install/download_prebuilt_libraries.md)
以Linux上推理为例，在本目录执行如下命令即可完成编译测试，支持此模型需保证FastDeploy版本1.0.4以上(x.x.x>=1.0.4)
```bash
mkdir build
cd build
# 下载FastDeploy预编译库，用户可在上文提到的`FastDeploy预编译库`中自行选择合适的版本使用
wget https://bj.bcebos.com/fastdeploy/release/cpp/fastdeploy-linux-x64-x.x.x.tgz
tar xvf fastdeploy-linux-x64-x.x.x.tgz
cmake .. -DFASTDEPLOY_INSTALL_DIR=${PWD}/fastdeploy-linux-x64-x.x.x
make -j
# 下载PP-LiteSeg模型文件和测试图片
wget https://bj.bcebos.com/paddlehub/fastdeploy/PP_LiteSeg_B_STDC2_cityscapes_without_argmax_infer.tgz
tar -xvf PP_LiteSeg_B_STDC2_cityscapes_without_argmax_infer.tgz
wget https://paddleseg.bj.bcebos.com/dygraph/demo/cityscapes_demo.png
# CPU推理
./infer_demo PP_LiteSeg_B_STDC2_cityscapes_without_argmax_infer cityscapes_demo.png 0
# GPU推理
./infer_demo PP_LiteSeg_B_STDC2_cityscapes_without_argmax_infer cityscapes_demo.png 1
```
以上命令只适用于Linux或MacOS, Windows下SDK的使用方式请参考:
- [如何在Windows中使用FastDeploy C++ SDK](../../../../../docs/cn/faq/use_sdk_on_windows.md)
如果用户使用华为昇腾NPU部署, 请参考以下方式在部署前初始化部署环境:
- [如何使用华为昇腾NPU部署](../../../../../docs/cn/faq/use_sdk_on_ascend.md)
运行完成可视化结果如下图所示
<div align="center">
<img src="https://user-images.githubusercontent.com/16222477/191712880-91ae128d-247a-43e0-b1e3-cafae78431e0.jpg", width=512px, height=256px />
</div>
## PaddleSeg C API接口
### 配置
```c
FD_C_RuntimeOptionWrapper* FD_C_CreateRuntimeOptionWrapper()
```
> 创建一个RuntimeOption的配置对象，并且返回操作它的指针。
>
> **返回**
>
> * **fd_c_runtime_option_wrapper**(FD_C_RuntimeOptionWrapper*): 指向RuntimeOption对象的指针
```c
void FD_C_RuntimeOptionWrapperUseCpu(
FD_C_RuntimeOptionWrapper* fd_c_runtime_option_wrapper)
```
> 开启CPU推理
>
> **参数**
>
> * **fd_c_runtime_option_wrapper**(FD_C_RuntimeOptionWrapper*): 指向RuntimeOption对象的指针
```c
void FD_C_RuntimeOptionWrapperUseGpu(
FD_C_RuntimeOptionWrapper* fd_c_runtime_option_wrapper,
int gpu_id)
```
> 开启GPU推理
>
> **参数**
>
> * **fd_c_runtime_option_wrapper**(FD_C_RuntimeOptionWrapper*): 指向RuntimeOption对象的指针
> * **gpu_id**(int): 显卡号
### 模型
```c
FD_C_PaddleSegWrapper* FD_C_CreatePaddleSegWrapper(
const char* model_file, const char* params_file, const char* config_file,
FD_C_RuntimeOptionWrapper* fd_c_runtime_option_wrapper,
const FD_C_ModelFormat model_format
)
```
> 创建一个PaddleSeg的模型，并且返回操作它的指针。
>
> **参数**
>
> * **model_file**(const char*): 模型文件路径
> * **params_file**(const char*): 参数文件路径
> * **config_file**(const char*): 配置文件路径
> * **fd_c_runtime_option_wrapper**(FD_C_RuntimeOptionWrapper*): 指向RuntimeOption的指针，表示后端推理配置
> * **model_format**(FD_C_ModelFormat): 模型格式
>
> **返回**
>
> * **fd_c_ppseg_wrapper**(FD_C_PaddleSegWrapper*): 指向PaddleSeg模型对象的指针
#### 读写图像
```c
FD_C_Mat FD_C_Imread(const char* imgpath)
```
> 读取一个图像，并且返回cv::Mat的指针。
>
> **参数**
>
> * **imgpath**(const char*): 图像文件路径
>
> **返回**
>
> * **imgmat**(FD_C_Mat): 指向图像数据cv::Mat的指针。
```c
FD_C_Bool FD_C_Imwrite(const char* savepath, FD_C_Mat img);
```
> 将图像写入文件中。
>
> **参数**
>
> * **savepath**(const char*): 保存图像的路径
> * **img**(FD_C_Mat): 指向图像数据的指针
>
> **返回**
>
> * **result**(FD_C_Bool): 表示操作是否成功
#### Predict函数
```c
FD_C_Bool FD_C_PaddleSegWrapperPredict(
FD_C_PaddleSegWrapper* fd_c_ppseg_wrapper,
FD_C_Mat img,
FD_C_SegmentationResult* result)
```
>
> 模型预测接口，输入图像直接并生成分类结果。
>
> **参数**
> * **fd_c_ppseg_wrapper**(FD_C_PaddleSegWrapper*): 指向PaddleSeg模型的指针
> * **img**（FD_C_Mat）: 输入图像的指针，指向cv::Mat对象，可以调用FD_C_Imread读取图像获取
> * **result**FD_C_SegmentationResult*): Segmentation检测结果，SegmentationResult说明参考[视觉模型预测结果](../../../../../docs/api/vision_results/)
#### Predict结果
```c
FD_C_Mat FD_C_VisSegmentation(FD_C_Mat im,
FD_C_SegmentationResult* result,
float weight)
```
>
> 对结果进行可视化，返回可视化的图像。
>
> **参数**
> * **im**(FD_C_Mat): 指向输入图像的指针
> * **segmentation_result**(FD_C_SegmentationResult*): 指向 FD_C_SegmentationResult结构的指针
> * **weight**(float): 透明度权重
>
> **返回**
> * **vis_im**(FD_C_Mat): 指向可视化图像的指针
## 其它文档
- [PPSegmentation 系列模型介绍](../../)
- [PaddleSeg Python部署](../python)
- [模型预测结果说明](../../../../../docs/api/vision_results/)
- [如何切换模型推理后端引擎](../../../../../docs/cn/faq/how_to_change_backend.md)

104
examples/vision/segmentation/paddleseg/cpu-gpu/csharp/README.md Executable file
View File

@@ -0,0 +1,104 @@
English | [简体中文](README_CN.md)
# PaddleSeg C# Deployment Example
This directory provides `infer.cs` to finish the deployment of PaddleSeg on CPU/GPU.
Before deployment, two steps require confirmation
- 1. Software and hardware should meet the requirements. Please refer to [FastDeploy Environment Requirements](../../../../../docs/en/build_and_install/download_prebuilt_libraries.md)
- 2. Download the precompiled deployment library and samples code according to your development environment. Refer to [FastDeploy Precompiled Library](../../../../../docs/en/build_and_install/download_prebuilt_libraries.md)
Please follow below instructions to compile and test in Windows. FastDeploy version 1.0.4 or above (x.x.x>=1.0.4) is required to support this model.
## 1. Download C# package management tool nuget client
> https://dist.nuget.org/win-x86-commandline/v6.4.0/nuget.exe
Add nuget program into system variable **PATH**
## 2. Download model and image for test
> https://bj.bcebos.com/paddlehub/fastdeploy/PP_LiteSeg_B_STDC2_cityscapes_without_argmax_infer.tgz # (Decompress it)
> https://paddleseg.bj.bcebos.com/dygraph/demo/cityscapes_demo.png
## 3. Compile example code
Open `x64 Native Tools Command Prompt for VS 2019` command tool on Windows, cd to the demo path of ppyoloe and execute commands
```shell
cd D:\Download\fastdeploy-win-x64-gpu-x.x.x\examples\vision\segmentation\paddleseg\cpu-gpu\csharp
mkdir build && cd build
cmake .. -G "Visual Studio 16 2019" -A x64 -DFASTDEPLOY_INSTALL_DIR=D:\Download\fastdeploy-win-x64-gpu-x.x.x -DCUDA_DIRECTORY="C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.2"
nuget restore
msbuild infer_demo.sln /m:4 /p:Configuration=Release /p:Platform=x64
```
For more information about how to use FastDeploy SDK to compile a project with Visual Studio 2019. Please refer to
- [Using the FastDeploy C++ SDK on Windows Platform](../../../../../docs/en/faq/use_sdk_on_windows.md)
## 4. Execute compiled program
fastdeploy.dll and related dynamic libraries are required by the program. FastDeploy provide a script to copy all required dll to your program path.
```shell
cd D:\Download\fastdeploy-win-x64-gpu-x.x.x
fastdeploy_init.bat install %cd% D:\Download\fastdeploy-win-x64-gpu-x.x.x\examples\vision\segmentation\paddleseg\cpu-gpu\csharp\build\Release
```
Then you can run your program and test the model with image
```shell
cd Release
# CPU inference
infer_demo PP_LiteSeg_B_STDC2_cityscapes_without_argmax_infer cityscapes_demo.png 0
# GPU inference
infer_demo PP_LiteSeg_B_STDC2_cityscapes_without_argmax_infer cityscapes_demo.png 1
```
## PaddleSeg C# Interface
### Model Class
```c#
fastdeploy.vision.segmentation.PaddleSeg(
string model_file,
string params_file,
string config_file,
fastdeploy.RuntimeOption runtime_option = null,
fastdeploy.ModelFormat model_format = ModelFormat.PADDLE)
```
> PaddleSeg initialization
> **Params**
>> * **model_file**(str): Model file path
>> * **params_file**(str): Parameter file path
>> * **config_file**(str): Config file path
>> * **runtime_option**(RuntimeOption): Backend inference configuration. null by default, which is the default configuration
>> * **model_format**(ModelFormat): Model format.
#### Predict Function
```c#
fastdeploy.SegmentationResult Predict(OpenCvSharp.Mat im)
```
> Model prediction interface. Input images and output results directly.
>
> **Params**
>
>> * **im**(Mat): Input images in HWC or BGR format
>>
> **Return**
>
>> * **result**: Segmentation prediction results, refer to [Vision Model Prediction Results](../../../../../docs/api/vision_results/) for SegmentationResult
## Other Documents
- [PPSegmentation Model Description](../../)
- [PaddleSeg Python Deployment](../python)
- [Model Prediction Results](../../../../../docs/api/vision_results/)
- [How to switch the model inference backend engine](../../../../../docs/cn/faq/how_to_change_backend.md)

102
examples/vision/segmentation/paddleseg/cpu-gpu/csharp/README_CN.md Normal file
View File

@@ -0,0 +1,102 @@
[English](README.md) | 简体中文
# PaddleSeg C#部署示例
本目录下提供`infer.cs`来调用C# API快速完成PaddleSeg模型在CPU/GPU上部署的示例。
在部署前，需确认以下两个步骤
- 1. 软硬件环境满足要求，参考[FastDeploy环境要求](../../../../../docs/cn/build_and_install/download_prebuilt_libraries.md)
- 2. 根据开发环境，下载预编译部署库和samples代码，参考[FastDeploy预编译库](../../../../../docs/cn/build_and_install/download_prebuilt_libraries.md)
在本目录执行如下命令即可在Windows完成编译测试，支持此模型需保证FastDeploy版本1.0.4以上(x.x.x>=1.0.4)
## 1. 下载C#包管理程序nuget客户端
> https://dist.nuget.org/win-x86-commandline/v6.4.0/nuget.exe
下载完成后将该程序添加到环境变量**PATH**中
## 2. 下载模型文件和测试图片
> https://bj.bcebos.com/paddlehub/fastdeploy/PP_LiteSeg_B_STDC2_cityscapes_without_argmax_infer.tgz # (下载后解压缩)
> https://paddleseg.bj.bcebos.com/dygraph/demo/cityscapes_demo.png
## 3. 编译示例代码
本文档编译的示例代码可在解压的库中找到，编译工具依赖VS 2019的安装，**Windows打开x64 Native Tools Command Prompt for VS 2019命令工具**，通过如下命令开始编译
```shell
cd D:\Download\fastdeploy-win-x64-gpu-x.x.x\examples\vision\segmentation\paddleseg\cpu-gpu\csharp
mkdir build && cd build
cmake .. -G "Visual Studio 16 2019" -A x64 -DFASTDEPLOY_INSTALL_DIR=D:\Download\fastdeploy-win-x64-gpu-x.x.x -DCUDA_DIRECTORY="C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.2"
nuget restore
msbuild infer_demo.sln /m:4 /p:Configuration=Release /p:Platform=x64
```
关于使用Visual Studio 2019创建sln工程，或者CMake工程等方式编译的更详细信息，可参考如下文档
- [在 Windows 使用 FastDeploy C++ SDK](../../../../../docs/cn/faq/use_sdk_on_windows.md)
- [FastDeploy C++库在Windows上的多种使用方式](../../../../../docs/cn/faq/use_sdk_on_windows_build.md)
## 4. 运行可执行程序
注意Windows上运行时，需要将FastDeploy依赖的库拷贝至可执行程序所在目录, 或者配置环境变量。FastDeploy提供了工具帮助我们快速将所有依赖库拷贝至可执行程序所在目录,通过如下命令将所有依赖的dll文件拷贝至可执行程序所在的目录(可能生成的可执行文件在Release下还有一层目录，这里假设生成的可执行文件在Release处)
```shell
cd D:\Download\fastdeploy-win-x64-gpu-x.x.x
fastdeploy_init.bat install %cd% D:\Download\fastdeploy-win-x64-gpu-x.x.x\examples\vision\segmentation\paddleseg\cpu-gpu\csharp\build\Release
```
将dll拷贝到当前路径后，准备好模型和图片，使用如下命令运行可执行程序即可
```shell
cd Release
# CPU推理
infer_demo PP_LiteSeg_B_STDC2_cityscapes_without_argmax_infer cityscapes_demo.png 0
# GPU推理
infer_demo PP_LiteSeg_B_STDC2_cityscapes_without_argmax_infer cityscapes_demo.png 1
```
## PaddleSeg C#接口
### 模型
```c#
fastdeploy.vision.segmentation.PaddleSeg(
string model_file,
string params_file,
string config_file,
fastdeploy.RuntimeOption runtime_option = null,
fastdeploy.ModelFormat model_format = ModelFormat.PADDLE)
```
> PaddleSeg模型加载和初始化。
> **参数**
>> * **model_file**(str): 模型文件路径
>> * **params_file**(str): 参数文件路径
>> * **config_file**(str): 配置文件路径
>> * **runtime_option**(RuntimeOption): 后端推理配置，默认为null，即采用默认配置
>> * **model_format**(ModelFormat): 模型格式，默认为PADDLE格式
#### Predict函数
```c#
fastdeploy.SegmentationResult Predict(OpenCvSharp.Mat im)
```
> 模型预测接口，输入图像直接输出结果。
>
> **参数**
>
>> * **im**(Mat): 输入图像，注意需为HWC，BGR格式
>>
> **返回值**
>
>> * **result**: Segmentation检测结果，SegmentationResult说明参考[视觉模型预测结果](../../../../../docs/api/vision_results/)
- [模型介绍](../../)
- [Python部署](../python)
- [视觉模型预测结果](../../../../../docs/api/vision_results/)
- [如何切换模型推理后端引擎](../../../../../docs/cn/faq/how_to_change_backend.md)