更换文档检测模型

2024-08-27 14:42:45 +08:00
parent aea6f19951
commit 1514e09c40
2072 changed files with 254336 additions and 4967 deletions
--- a/paddle_detection/deploy/fastdeploy/cpu-gpu/README.md
+++ b/paddle_detection/deploy/fastdeploy/cpu-gpu/README.md
@@ -0,0 +1,97 @@
+[English](README.md) | 简体中文
+
+# PaddleDetection检测模型在CPU-GPU上的部署方案—FastDeploy
+
+## 1. 说明
+PaddleDetection支持利用FastDeploy在NVIDIA GPU、X86 CPU、飞腾CPU、ARM CPU、Intel GPU(独立显卡/集成显卡)硬件上快速部署检测模型
+
+## 2. 使用预导出的模型列表  
+为了方便开发者的测试，下面提供了PaddleDetection导出的各系列模型，开发者可直接下载使用。其中精度指标来源于PaddleDetection中对各模型的介绍，详情各参考PaddleDetection中的说明。
+
+### 2.1 目标检测及实例分割模型
+| 模型                                                               | 参数大小    | 精度    | 备注 |
+|:---------------------------------------------------------------- |:----- |:----- | :------ |
+| [picodet_l_320_coco_lcnet](https://bj.bcebos.com/paddlehub/fastdeploy/picodet_l_320_coco_lcnet.tgz) |23MB | Box AP 42.6% |
+| [ppyoloe_crn_l_300e_coco](https://bj.bcebos.com/paddlehub/fastdeploy/ppyoloe_crn_l_300e_coco.tgz) |200MB | Box AP 51.4% |
+| [ppyoloe_plus_crn_m_80e_coco](https://bj.bcebos.com/fastdeploy/models/ppyoloe_plus_crn_m_80e_coco.tgz) |83.3MB | Box AP 49.8% |
+| [ppyolo_r50vd_dcn_1x_coco](https://bj.bcebos.com/paddlehub/fastdeploy/ppyolo_r50vd_dcn_1x_coco.tgz) | 180MB | Box AP 44.8% | 暂不支持TensorRT |
+| [ppyolov2_r101vd_dcn_365e_coco](https://bj.bcebos.com/paddlehub/fastdeploy/ppyolov2_r101vd_dcn_365e_coco.tgz) | 282MB | Box AP 49.7% | 暂不支持TensorRT |
+| [yolov3_darknet53_270e_coco](https://bj.bcebos.com/paddlehub/fastdeploy/yolov3_darknet53_270e_coco.tgz) |237MB | Box AP 39.1% | |
+| [yolox_s_300e_coco](https://bj.bcebos.com/paddlehub/fastdeploy/yolox_s_300e_coco.tgz) | 35MB | Box AP 40.4% | |
+| [faster_rcnn_r50_vd_fpn_2x_coco](https://bj.bcebos.com/paddlehub/fastdeploy/faster_rcnn_r50_vd_fpn_2x_coco.tgz) | 160MB | Box AP 40.8%| 暂不支持TensorRT |
+| [mask_rcnn_r50_1x_coco](https://bj.bcebos.com/paddlehub/fastdeploy/mask_rcnn_r50_1x_coco.tgz) | 128M | Box AP 37.4%, Mask AP 32.8%| 暂不支持TensorRT、ORT |
+| [ssd_mobilenet_v1_300_120e_voc](https://bj.bcebos.com/paddlehub/fastdeploy/ssd_mobilenet_v1_300_120e_voc.tgz) | 24.9M | Box AP 73.8%| 暂不支持TensorRT、ORT |
+| [ssd_vgg16_300_240e_voc](https://bj.bcebos.com/paddlehub/fastdeploy/ssd_vgg16_300_240e_voc.tgz) | 106.5M | Box AP 77.8%| 暂不支持TensorRT、ORT |
+| [ssdlite_mobilenet_v1_300_coco](https://bj.bcebos.com/paddlehub/fastdeploy/ssdlite_mobilenet_v1_300_coco.tgz) | 29.1M | | 暂不支持TensorRT、ORT |
+| [rtmdet_l_300e_coco](https://bj.bcebos.com/paddlehub/fastdeploy/rtmdet_l_300e_coco.tgz) | 224M | Box AP 51.2%|  |
+| [rtmdet_s_300e_coco](https://bj.bcebos.com/paddlehub/fastdeploy/rtmdet_s_300e_coco.tgz) | 42M | Box AP 44.5%|  |
+| [yolov5_l_300e_coco](https://bj.bcebos.com/paddlehub/fastdeploy/yolov5_l_300e_coco.tgz) | 183M | Box AP 48.9%|  |
+| [yolov5_s_300e_coco](https://bj.bcebos.com/paddlehub/fastdeploy/yolov5_s_300e_coco.tgz) | 31M | Box AP 37.6%|  |
+| [yolov6_l_300e_coco](https://bj.bcebos.com/paddlehub/fastdeploy/yolov6_l_300e_coco.tgz) | 229M | Box AP 51.0%|  |
+| [yolov6_s_400e_coco](https://bj.bcebos.com/paddlehub/fastdeploy/yolov6_s_400e_coco.tgz) | 68M | Box AP 43.4%|  |
+| [yolov7_l_300e_coco](https://bj.bcebos.com/paddlehub/fastdeploy/yolov7_l_300e_coco.tgz) | 145M | Box AP 51.0%|  |
+| [yolov7_x_300e_coco](https://bj.bcebos.com/paddlehub/fastdeploy/yolov7_x_300e_coco.tgz) | 277M | Box AP 53.0%|  |
+| [cascade_rcnn_r50_fpn_1x_coco](https://bj.bcebos.com/paddlehub/fastdeploy/cascade_rcnn_r50_fpn_1x_coco.tgz) | 271M | Box AP 41.1%|  暂不支持TensorRT、ORT |
+| [cascade_rcnn_r50_vd_fpn_ssld_2x_coco](https://bj.bcebos.com/paddlehub/fastdeploy/cascade_rcnn_r50_vd_fpn_ssld_2x_coco.tgz) | 271M | Box AP 45.0%|  暂不支持TensorRT、ORT |
+| [faster_rcnn_enhance_3x_coco](https://bj.bcebos.com/paddlehub/fastdeploy/faster_rcnn_enhance_3x_coco.tgz) | 119M | Box AP 41.5%|  暂不支持TensorRT、ORT |
+| [fcos_r50_fpn_1x_coco](https://bj.bcebos.com/paddlehub/fastdeploy/fcos_r50_fpn_1x_coco.tgz) | 129M | Box AP 39.6%|  暂不支持TensorRT |
+| [gfl_r50_fpn_1x_coco](https://bj.bcebos.com/paddlehub/fastdeploy/gfl_r50_fpn_1x_coco.tgz) | 128M | Box AP 41.0%|  暂不支持TensorRT |
+| [ppyoloe_crn_l_80e_sliced_visdrone_640_025](https://bj.bcebos.com/paddlehub/fastdeploy/ppyoloe_crn_l_80e_sliced_visdrone_640_025.tgz) | 200M | Box AP 31.9%|  |
+| [retinanet_r101_fpn_2x_coco](https://bj.bcebos.com/paddlehub/fastdeploy/retinanet_r101_fpn_2x_coco.tgz) | 210M | Box AP 40.6%|  暂不支持TensorRT、ORT |
+| [retinanet_r50_fpn_1x_coco](https://bj.bcebos.com/paddlehub/fastdeploy/retinanet_r50_fpn_1x_coco.tgz) | 136M | Box AP 37.5%|  暂不支持TensorRT、ORT |
+| [tood_r50_fpn_1x_coco](https://bj.bcebos.com/paddlehub/fastdeploy/tood_r50_fpn_1x_coco.tgz) | 130M | Box AP 42.5%|  暂不支持TensorRT、ORT |
+| [ttfnet_darknet53_1x_coco](https://bj.bcebos.com/paddlehub/fastdeploy/ttfnet_darknet53_1x_coco.tgz) | 178M | Box AP 33.5%|  暂不支持TensorRT、ORT |  
+| [yolov8_x_500e_coco](https://bj.bcebos.com/paddlehub/fastdeploy/yolov8_x_500e_coco.tgz) | 265M | Box AP 53.8%
+| [yolov8_l_500e_coco](https://bj.bcebos.com/paddlehub/fastdeploy/yolov8_l_500e_coco.tgz) | 173M | Box AP 52.8%
+| [yolov8_m_500e_coco](https://bj.bcebos.com/paddlehub/fastdeploy/yolov8_m_500e_coco.tgz) | 99M | Box AP 50.2%
+| [yolov8_s_500e_coco](https://bj.bcebos.com/paddlehub/fastdeploy/yolov8_s_500e_coco.tgz) | 43M | Box AP 44.9%
+| [yolov8_n_500e_coco](https://bj.bcebos.com/paddlehub/fastdeploy/yolov8_n_500e_coco.tgz) | 13M | Box AP 37.3%
+
+### 2.2 关键点检测模型  
+| 模型 | 说明 | 模型格式 | 版本 |
+| :--- | :--- | :------- | :--- |
+| [PP-TinyPose-128x96](https://bj.bcebos.com/paddlehub/fastdeploy/PP_TinyPose_128x96_infer.tgz) | 单人关键点检测模型 | Paddle | [Release/2.5](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.5/configs/keypoint/tiny_pose) |
+| [PP-TinyPose-256x192](https://bj.bcebos.com/paddlehub/fastdeploy/PP_TinyPose_256x192_infer.tgz) | 单人关键点检测模型 | Paddle | [Release/2.5](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.5/configs/keypoint/tiny_pose) |
+| [PicoDet-S-Lcnet-Pedestrian-192x192](https://bj.bcebos.com/paddlehub/fastdeploy/PP_PicoDet_V2_S_Pedestrian_192x192_infer.tgz) + [PP-TinyPose-128x96](https://bj.bcebos.com/paddlehub/fastdeploy/PP_TinyPose_128x96_infer.tgz) | 单人关键点检测串联配置 | Paddle |[Release/2.5](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.5/configs/keypoint/tiny_pose) |
+| [PicoDet-S-Lcnet-Pedestrian-320x320](https://bj.bcebos.com/paddlehub/fastdeploy/PP_PicoDet_V2_S_Pedestrian_320x320_infer.tgz) + [PP-TinyPose-256x192](https://bj.bcebos.com/paddlehub/fastdeploy/PP_TinyPose_256x192_infer.tgz)  | 多人关键点检测串联配置 | Paddle |[Release/2.5](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.5/configs/keypoint/tiny_pose) |
+
+
+## 3. 自行导出PaddleDetection部署模型  
+### 3.1 模型版本
+支持[PaddleDetection](https://github.com/PaddlePaddle/PaddleDetection)大于等于2.4版本的PaddleDetection模型部署。目前FastDeploy测试过成功部署的模型:   
+
+- [PP-YOLOE(含PP-YOLOE+)系列模型](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.6/configs/ppyoloe)
+- [PicoDet系列模型](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.6/configs/picodet)
+- [PP-YOLO系列模型(含v2)](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.6/configs/ppyolo)
+- [YOLOv3系列模型](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.6/configs/yolov3)
+- [YOLOX系列模型](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.6/configs/yolox)
+- [FasterRCNN系列模型](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.6/configs/faster_rcnn)
+- [MaskRCNN系列模型](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.6/configs/mask_rcnn)
+- [SSD系列模型](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.6/configs/ssd)
+- [YOLOv5系列模型](https://github.com/PaddlePaddle/PaddleYOLO/tree/release/2.6/configs/yolov5)
+- [YOLOv6系列模型](https://github.com/PaddlePaddle/PaddleYOLO/tree/release/2.6/configs/yolov6)
+- [YOLOv7系列模型](https://github.com/PaddlePaddle/PaddleYOLO/tree/release/2.6/configs/yolov7)  
+- [YOLOv8系列模型](https://github.com/PaddlePaddle/PaddleYOLO/tree/release/2.6/configs/yolov8)
+- [RTMDet系列模型](https://github.com/PaddlePaddle/PaddleYOLO/tree/release/2.6/configs/rtmdet)
+- [CascadeRCNN系列模型](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.6/configs/cascade_rcnn)
+- [PSSDet系列模型](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.6/configs/rcnn_enhance)
+- [RetinaNet系列模型](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.6/configs/retinanet)
+- [PPYOLOESOD系列模型](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.6/configs/smalldet)
+- [FCOS系列模型](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.6/configs/fcos)
+- [TTFNet系列模型](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.6/configs/ttfnet)
+- [TOOD系列模型](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.6/configs/tood)
+- [GFL系列模型](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.6/configs/gfl)
+- [PP-PicoDet + PP-TinyPose系列模型](https://github.com/PaddlePaddle/PaddleDetection/tree/release/2.5/configs/keypoint/tiny_pose/README.md)
+
+### 3.2 模型导出   
+PaddleDetection模型导出，请参考其文档说明[模型导出](https://github.com/PaddlePaddle/PaddleDetection/blob/release/2.6/deploy/EXPORT_MODEL.md)，**注意**：PaddleDetection导出的模型包含`model.pdmodel`、`model.pdiparams`和`infer_cfg.yml`三个文件，FastDeploy会从yaml文件中获取模型在推理时需要的预处理信息
+
+### 3.3 导出须知  
+如果您是自行导出PaddleDetection推理模型，请注意以下问题：
+- 在导出模型时不要进行NMS的去除操作，正常导出即可  
+- 如果用于跑原生TensorRT后端（非Paddle Inference后端），不要添加--trt参数
+- 导出模型时，不要添加`fuse_normalize=True`参数
+
+## 4. 详细的部署示例  
+- [Python部署](python)
+- [C++部署](cpp)
--- a/paddle_detection/deploy/fastdeploy/cpu-gpu/cpp/CMakeLists.txt
+++ b/paddle_detection/deploy/fastdeploy/cpu-gpu/cpp/CMakeLists.txt
@@ -0,0 +1,13 @@
+PROJECT(infer_demo C CXX)
+CMAKE_MINIMUM_REQUIRED (VERSION 3.10)
+
+option(FASTDEPLOY_INSTALL_DIR "Path of downloaded fastdeploy sdk.")
+
+include(${FASTDEPLOY_INSTALL_DIR}/FastDeploy.cmake)
+
+include_directories(${FASTDEPLOY_INCS})
+
+add_executable(infer_demo ${PROJECT_SOURCE_DIR}/infer.cc)
+add_executable(infer_tinypose_demo ${PROJECT_SOURCE_DIR}/pptinypose_infer.cc)
+target_link_libraries(infer_demo ${FASTDEPLOY_LIBS})
+target_link_libraries(infer_tinypose_demo ${FASTDEPLOY_LIBS})
--- a/paddle_detection/deploy/fastdeploy/cpu-gpu/cpp/README.md
+++ b/paddle_detection/deploy/fastdeploy/cpu-gpu/cpp/README.md
@@ -0,0 +1,142 @@
+[English](README.md) | 简体中文
+# PaddleDetection CPU-GPU C++部署示例
+
+本目录下提供`infer.cc`快速完成PPYOLOE模型包括PPYOLOE在CPU/GPU，以及GPU上通过Paddle-TensorRT加速部署的示例。 
+
+## 1. 说明  
+PaddleDetection支持利用FastDeploy在NVIDIA GPU、X86 CPU、飞腾CPU、ARM CPU、Intel GPU(独立显卡/集成显卡)硬件上快速部署PaddleDetection模型。FastDeploy目前支持的模型系列，包括但不限于`PPYOLOE`, `PicoDet`, `PaddleYOLOX`, `PPYOLO`, `FasterRCNN`，`SSD`,`PaddleYOLOv5`,`PaddleYOLOv6`,`PaddleYOLOv7`,`RTMDet`,`CascadeRCNN`,`PSSDet`,`RetinaNet`,`PPYOLOESOD`,`FCOS`,`TTFNet`,`TOOD`,`GFL`所有类名的构造函数和预测函数在参数上完全一致。所有模型的调用，只需要参考PPYOLOE的示例，即可快速调用。
+
+## 2. 部署环境准备  
+在部署前，需确认软硬件环境，同时下载预编译部署库，参考[FastDeploy安装文档](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install#FastDeploy预编译库安装)安装FastDeploy预编译库。
+
+## 3. 部署模型准备
+在部署前，请准备好您所需要运行的推理模型，你可以选择使用[预导出的推理模型](../README.md)或者[自行导出PaddleDetection部署模型](../README.md)。  
+
+## 4. 运行部署示例
+以Linux上推理为例，在本目录执行如下命令即可完成编译测试，支持此模型需保证FastDeploy版本1.0.4以上(x.x.x>=1.0.4)
+
+### 4.1 目标检测示例
+```bash
+# 下载FastDeploy预编译库，用户可在上文提到的`FastDeploy预编译库`中自行选择合适的版本使用
+wget https://bj.bcebos.com/fastdeploy/release/cpp/fastdeploy-linux-x64-gpu-x.x.x.tgz
+tar xvf fastdeploy-linux-x64-gpu-x.x.x.tgz
+
+# 下载部署示例代码
+git clone https://github.com/PaddlePaddle/PaddleDetection.git
+cd PaddleDetection/deploy/fastdeploy/cpu-gpu/cpp
+# 注意：如果当前分支找不到下面的fastdeploy测试代码，请切换到develop分支
+# git checkout develop
+
+# 编译部署示例
+mkdir build && cd build
+mv ../fastdeploy-linux-x64-gpu-x.x.x .
+cmake .. -DFASTDEPLOY_INSTALL_DIR=${PWD}/fastdeploy-linux-x64-gpu-x.x.x
+make -j
+
+# 下载PPYOLOE模型文件和测试图片
+wget https://bj.bcebos.com/paddlehub/fastdeploy/ppyoloe_crn_l_300e_coco.tgz
+wget https://gitee.com/paddlepaddle/PaddleDetection/raw/release/2.4/demo/000000014439.jpg
+tar xvf ppyoloe_crn_l_300e_coco.tgz
+
+# 运行部署示例
+# CPU推理
+./infer_demo ./ppyoloe_crn_l_300e_coco 000000014439.jpg 0
+# GPU推理
+./infer_demo ./ppyoloe_crn_l_300e_coco 000000014439.jpg 1
+# GPU上Paddle-TensorRT推理（注意：TensorRT推理第一次运行，有序列化模型的操作，有一定耗时，需要耐心等待）
+./infer_demo ./ppyoloe_crn_l_300e_coco 000000014439.jpg 2
+```
+
+运行完成可视化结果如下图所示
+<div  align="center">  
+<img src="https://user-images.githubusercontent.com/19339784/184326520-7075e907-10ed-4fad-93f8-52d0e35d4964.jpg", width=480px, height=320px />
+</div> 
+
+### 4.2 关键点检测示例 
+```bash
+# 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-gpu-x.x.x.tgz
+tar xvf fastdeploy-linux-x64-gpu-x.x.x.tgz
+
+# 下载部署示例代码
+git clone https://github.com/PaddlePaddle/PaddleDetection.git
+cd PaddleDetection/deploy/fastdeploy/cpu-gpu/cpp
+# 注意：如果当前分支找不到下面的fastdeploy测试代码，请切换到develop分支
+# git checkout develop
+
+# 编译部署示例
+mkdir build && cd build
+mv ../fastdeploy-linux-x64-gpu-x.x.x .
+cmake .. -DFASTDEPLOY_INSTALL_DIR=${PWD}/fastdeploy-linux-x64-gpu-x.x.x
+make -j
+
+# 下载PP-TinyPose模型文件和测试图片
+wget https://bj.bcebos.com/paddlehub/fastdeploy/PP_TinyPose_256x192_infer.tgz
+tar -xvf PP_TinyPose_256x192_infer.tgz
+wget https://bj.bcebos.com/paddlehub/fastdeploy/hrnet_demo.jpg
+
+# 运行部署示例
+# CPU推理
+./infer_tinypose_demo PP_TinyPose_256x192_infer hrnet_demo.jpg 0
+# GPU推理
+./infer_tinypose_demo PP_TinyPose_256x192_infer hrnet_demo.jpg 1
+# GPU上Paddle-TensorRT推理（注意：TensorRT推理第一次运行，有序列化模型的操作，有一定耗时，需要耐心等待）
+./infer_tinypose_demo PP_TinyPose_256x192_infer hrnet_demo.jpg 2
+```
+
+运行完成可视化结果如下图所示
+<div  align="center">  
+<img src="https://user-images.githubusercontent.com/16222477/196386764-dd51ad56-c410-4c54-9580-643f282f5a83.jpeg", width=359px, height=423px />
+</div>
+
+关于如何进行多人关键点检测，请参考[PPTinyPose Pipeline示例](./det_keypoint_unite/)
+
+- 注意，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](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/faq/use_sdk_on_windows.md)  
+- 关于如何通过FastDeploy使用更多不同的推理后端，以及如何使用不同的硬件，请参考文档：[如何切换模型推理后端引擎](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/faq/how_to_change_backend.md)  
+
+## 5. PaddleDetection C++接口
+FastDeploy目前支持的模型系列，包括但不限于`PPYOLOE`, `PicoDet`, `PaddleYOLOX`, `PPYOLO`, `FasterRCNN`，`SSD`,`PaddleYOLOv5`,`PaddleYOLOv6`,`PaddleYOLOv7`,`RTMDet`,`CascadeRCNN`,`PSSDet`,`RetinaNet`,`PPYOLOESOD`,`FCOS`,`TTFNet`,`TOOD`,`GFL`所有类名的构造函数和预测函数在参数上完全一致。所有模型的调用，只需要参考PPYOLOE的示例，即可快速调用。 
+
+### 5.1 目标检测及实例分割模型
+```c++
+fastdeploy::vision::detection::PicoDet(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::SOLOv2(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::PPYOLOE(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::PPYOLO(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::YOLOv3(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::PaddleYOLOX(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::FasterRCNN(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::MaskRCNN(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::SSD(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::PaddleYOLOv5(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::PaddleYOLOv6(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::PaddleYOLOv7(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::PaddleYOLOv8(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::CascadeRCNN(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::PSSDet(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::RetinaNet(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::PPYOLOESOD(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::FCOS(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::TOOD(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+fastdeploy::vision::detection::GFL(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+``` 
+
+### 5.2 关键点检测模型  
+```C++
+fastdeploy::vision::keypointdetection::PPTinyPose(const string& model_file, const string& params_file, const string& config_file, const RuntimeOption& runtime_option = RuntimeOption(), const ModelFormat& model_format = ModelFormat::PADDLE);
+```
+
+PaddleDetection模型加载和初始化，其中model_file， params_file为导出的Paddle部署模型格式, config_file为PaddleDetection同时导出的部署配置yaml文件  
+
+## 6. 更多指南
+- [PaddleDetection C++ API文档](https://www.paddlepaddle.org.cn/fastdeploy-api-doc/cpp/html/namespacefastdeploy_1_1vision_1_1detection.html)
+- [FastDeploy部署PaddleDetection模型概览](../../)
+- [Python部署](../python)
+
+## 7. 常见问题
+- [如何切换模型推理后端引擎](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/faq/how_to_change_backend.md)
+- [Intel GPU(独立显卡/集成显卡)的使用](https://github.com/PaddlePaddle/FastDeploy/blob/develop/tutorials/intel_gpu/README.md)
+- [编译CPU部署库](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install/cpu.md)
+- [编译GPU部署库](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install/gpu.md)
+- [编译Jetson部署库](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install/jetson.md)
+
--- a/paddle_detection/deploy/fastdeploy/cpu-gpu/cpp/det_keypoint_unite/CMakeLists.txt
+++ b/paddle_detection/deploy/fastdeploy/cpu-gpu/cpp/det_keypoint_unite/CMakeLists.txt
@@ -0,0 +1,11 @@
+PROJECT(infer_demo C CXX)
+CMAKE_MINIMUM_REQUIRED (VERSION 3.12)
+
+option(FASTDEPLOY_INSTALL_DIR "Path of downloaded fastdeploy sdk.")
+
+include(${FASTDEPLOY_INSTALL_DIR}/FastDeploy.cmake)
+
+include_directories(${FASTDEPLOY_INCS})
+
+add_executable(infer_demo ${PROJECT_SOURCE_DIR}/det_keypoint_unite_infer.cc)
+target_link_libraries(infer_demo ${FASTDEPLOY_LIBS})
--- a/paddle_detection/deploy/fastdeploy/cpu-gpu/cpp/det_keypoint_unite/README.md
+++ b/paddle_detection/deploy/fastdeploy/cpu-gpu/cpp/det_keypoint_unite/README.md
@@ -0,0 +1,74 @@
+[English](README.md) | 简体中文
+# PP-PicoDet + PP-TinyPose (Pipeline) CPU-GPU C++部署示例
+
+本目录下提供`det_keypoint_unite_infer.cc`快速完成多人模型配置 PP-PicoDet + PP-TinyPose 在CPU/GPU，以及GPU上通过TensorRT加速部署的`单图多人关键点检测`示例。执行如下脚本即可完成。**注意**: PP-TinyPose单模型独立部署，请参考[PP-TinyPose 单模型](../README.md)
+
+## 1. 部署环境准备  
+在部署前，需确认软硬件环境，同时下载预编译部署库，参考[FastDeploy安装文档](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install#FastDeploy预编译库安装)安装FastDeploy预编译库。
+
+## 2. 部署模型准备
+在部署前，请准备好您所需要运行的推理模型，你可以选择使用[预导出的推理模型](../../README.md)或者[自行导出PaddleDetection部署模型](../../README.md)。  
+
+## 3. 运行部署示例
+以Linux上推理为例，在本目录执行如下命令即可完成编译测试，支持此模型需保证FastDeploy版本1.0.4以上(x.x.x>=1.0.4)
+
+```bash
+mkdir build
+cd build
+# 下载FastDeploy预编译库，用户可在上文提到的`FastDeploy预编译库`中自行选择合适的版本使用
+wget https://bj.bcebos.com/fastdeploy/release/cpp/fastdeploy-linux-x64-x.x.x.tgz
+tar xvf fastdeploy-linux-x64-x.x.x.tgz
+cmake .. -DFASTDEPLOY_INSTALL_DIR=${PWD}/fastdeploy-linux-x64-x.x.x
+make -j
+
+# 下载部署示例代码
+git clone https://github.com/PaddlePaddle/PaddleDetection.git
+cd PaddleDetection/deploy/fastdeploy/cpu-gpu/cpp/det_keypoint_unite
+# 注意：如果当前分支找不到下面的fastdeploy测试代码，请切换到develop分支
+# git checkout develop
+
+# 下载PP-TinyPose和PP-PicoDet模型文件和测试图片
+wget https://bj.bcebos.com/paddlehub/fastdeploy/PP_TinyPose_256x192_infer.tgz
+tar -xvf PP_TinyPose_256x192_infer.tgz
+wget https://bj.bcebos.com/paddlehub/fastdeploy/PP_PicoDet_V2_S_Pedestrian_320x320_infer.tgz
+tar -xvf PP_PicoDet_V2_S_Pedestrian_320x320_infer.tgz
+wget https://bj.bcebos.com/paddlehub/fastdeploy/000000018491.jpg
+
+# CPU推理
+./infer_demo PP_PicoDet_V2_S_Pedestrian_320x320_infer PP_TinyPose_256x192_infer 000000018491.jpg 0
+# GPU推理
+./infer_demo PP_PicoDet_V2_S_Pedestrian_320x320_infer PP_TinyPose_256x192_infer 000000018491.jpg 1
+# GPU上Paddle-TensorRT推理（注意：TensorRT推理第一次运行，有序列化模型的操作，有一定耗时，需要耐心等待）
+./infer_demo PP_PicoDet_V2_S_Pedestrian_320x320_infer PP_TinyPose_256x192_infer 000000018491.jpg 2
+```
+
+运行完成可视化结果如下图所示
+<div  align="center">  
+<img src="https://user-images.githubusercontent.com/16222477/196393343-eeb6b68f-0bc6-4927-871f-5ac610da7293.jpeg", width=359px, height=423px />
+</div>  
+
+- 注意，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](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/faq/use_sdk_on_windows.md)  
+- 关于如何通过FastDeploy使用更多不同的推理后端，以及如何使用不同的硬件，请参考文档：[如何切换模型推理后端引擎](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/faq/how_to_change_backend.md) 
+
+## 4. PP-TinyPose 模型串联 C++ 接口
+
+```c++
+fastdeploy::pipeline::PPTinyPose(
+        fastdeploy::vision::detection::PicoDet* det_model,
+        fastdeploy::vision::keypointdetection::PPTinyPose* pptinypose_model)
+```
+
+PPTinyPose Pipeline模型加载和初始化。det_model表示初始化后的检测模型，pptinypose_model表示初始化后的关键点检测模型。
+
+
+## 5. 更多指南
+- [PaddleDetection C++ API文档](https://www.paddlepaddle.org.cn/fastdeploy-api-doc/cpp/html/namespacefastdeploy_1_1vision_1_1detection.html)
+- [FastDeploy部署PaddleDetection模型概览](../../../)
+- [Python部署](../../python/det_keypoint_unite/)
+
+## 6. 常见问题
+- [如何切换模型推理后端引擎](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/faq/how_to_change_backend.md)
+- [Intel GPU(独立显卡/集成显卡)的使用](https://github.com/PaddlePaddle/FastDeploy/blob/develop/tutorials/intel_gpu/README.md)
+- [编译CPU部署库](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install/cpu.md)
+- [编译GPU部署库](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install/gpu.md)
+- [编译Jetson部署库](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install/jetson.md)
--- a/paddle_detection/deploy/fastdeploy/cpu-gpu/cpp/det_keypoint_unite/det_keypoint_unite_infer.cc
+++ b/paddle_detection/deploy/fastdeploy/cpu-gpu/cpp/det_keypoint_unite/det_keypoint_unite_infer.cc
@@ -0,0 +1,205 @@
+// Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#include "fastdeploy/vision.h"
+#include "fastdeploy/pipeline.h"
+
+#ifdef WIN32
+const char sep = '\\';
+#else
+const char sep = '/';
+#endif
+
+void CpuInfer(const std::string& det_model_dir,
+              const std::string& tinypose_model_dir,
+              const std::string& image_file) {
+  auto det_model_file = det_model_dir + sep + "model.pdmodel";
+  auto det_params_file = det_model_dir + sep + "model.pdiparams";
+  auto det_config_file = det_model_dir + sep + "infer_cfg.yml";
+  auto det_model = fastdeploy::vision::detection::PicoDet(
+      det_model_file, det_params_file, det_config_file);
+  if (!det_model.Initialized()) {
+    std::cerr << "Detection Model Failed to initialize." << std::endl;
+    return;
+  }
+
+  auto tinypose_model_file = tinypose_model_dir + sep + "model.pdmodel";
+  auto tinypose_params_file = tinypose_model_dir + sep + "model.pdiparams";
+  auto tinypose_config_file = tinypose_model_dir + sep + "infer_cfg.yml";
+  auto tinypose_model = fastdeploy::vision::keypointdetection::PPTinyPose(
+      tinypose_model_file, tinypose_params_file, tinypose_config_file);
+  if (!tinypose_model.Initialized()) {
+    std::cerr << "TinyPose Model Failed to initialize." << std::endl;
+    return;
+  }
+
+  auto im = cv::imread(image_file);
+  fastdeploy::vision::KeyPointDetectionResult res;
+
+  auto pipeline =fastdeploy::pipeline::PPTinyPose(&det_model, &tinypose_model);
+  pipeline.detection_model_score_threshold = 0.5;
+  if (!pipeline.Predict(&im, &res)) {
+    std::cerr << "TinyPose Prediction Failed." << std::endl;
+    return;
+  } else {
+    std::cout << "TinyPose Prediction Done!" << std::endl;
+  }
+  
+  std::cout << res.Str() << std::endl;
+
+  auto vis_im =
+      fastdeploy::vision::VisKeypointDetection(im, res, 0.2);
+  cv::imwrite("vis_result.jpg", vis_im);
+  std::cout << "TinyPose visualized result saved in ./vis_result.jpg"
+            << std::endl;
+}
+
+void GpuInfer(const std::string& det_model_dir,
+              const std::string& tinypose_model_dir,
+              const std::string& image_file) {
+  auto option = fastdeploy::RuntimeOption();
+  option.UseGpu();
+  auto det_model_file = det_model_dir + sep + "model.pdmodel";
+  auto det_params_file = det_model_dir + sep + "model.pdiparams";
+  auto det_config_file = det_model_dir + sep + "infer_cfg.yml";
+  auto det_model = fastdeploy::vision::detection::PicoDet(
+      det_model_file, det_params_file, det_config_file, option);
+  if (!det_model.Initialized()) {
+    std::cerr << "Detection Model Failed to initialize." << std::endl;
+    return;
+  }
+
+  auto tinypose_model_file = tinypose_model_dir + sep + "model.pdmodel";
+  auto tinypose_params_file = tinypose_model_dir + sep + "model.pdiparams";
+  auto tinypose_config_file = tinypose_model_dir + sep + "infer_cfg.yml";
+  auto tinypose_model = fastdeploy::vision::keypointdetection::PPTinyPose(
+      tinypose_model_file, tinypose_params_file, tinypose_config_file, option);
+  if (!tinypose_model.Initialized()) {
+    std::cerr << "TinyPose Model Failed to initialize." << std::endl;
+    return;
+  }
+
+  auto im = cv::imread(image_file);
+  fastdeploy::vision::KeyPointDetectionResult res;
+
+  auto pipeline =
+      fastdeploy::pipeline::PPTinyPose(
+          &det_model, &tinypose_model);
+  pipeline.detection_model_score_threshold = 0.5;
+  if (!pipeline.Predict(&im, &res)) {
+    std::cerr << "TinyPose Prediction Failed." << std::endl;
+    return;
+  } else {
+    std::cout << "TinyPose Prediction Done!" << std::endl;
+  }
+  
+  std::cout << res.Str() << std::endl;
+
+  auto vis_im =
+      fastdeploy::vision::VisKeypointDetection(im, res, 0.2);
+  cv::imwrite("vis_result.jpg", vis_im);
+  std::cout << "TinyPose visualized result saved in ./vis_result.jpg"
+            << std::endl;
+}
+
+void TrtInfer(const std::string& det_model_dir,
+              const std::string& tinypose_model_dir,
+              const std::string& image_file) {
+  auto det_model_file = det_model_dir + sep + "model.pdmodel";
+  auto det_params_file = det_model_dir + sep + "model.pdiparams";
+  auto det_config_file = det_model_dir + sep + "infer_cfg.yml";
+
+  auto det_option = fastdeploy::RuntimeOption();
+  det_option.UseGpu();
+  det_option.UsePaddleInferBackend();
+  // If use original Tensorrt, not Paddle-TensorRT,
+  // please try `option.UseTrtBackend()`
+  det_option.paddle_infer_option.enable_trt = true;
+  det_option.paddle_infer_option.collect_trt_shape = true;
+  det_option.trt_option.SetShape("image", {1, 3, 320, 320}, {1, 3, 320, 320}, 
+                                 {1, 3, 320, 320});
+  det_option.trt_option.SetShape("scale_factor", {1, 2}, {1, 2}, {1, 2});    
+  auto det_model = fastdeploy::vision::detection::PicoDet(
+      det_model_file, det_params_file, det_config_file, det_option);
+  if (!det_model.Initialized()) {
+    std::cerr << "Detection Model Failed to initialize." << std::endl;
+    return;
+  }
+
+  auto tinypose_model_file = tinypose_model_dir + sep + "model.pdmodel";
+  auto tinypose_params_file = tinypose_model_dir + sep + "model.pdiparams";
+  auto tinypose_config_file = tinypose_model_dir + sep + "infer_cfg.yml";
+  auto tinypose_option = fastdeploy::RuntimeOption();
+  
+  tinypose_option.UseGpu();
+  tinypose_option.UsePaddleInferBackend();
+  // If use original Tensorrt, not Paddle-TensorRT,
+  // please try `option.UseTrtBackend()`
+  tinypose_option.paddle_infer_option.enable_trt = true;
+  tinypose_option.paddle_infer_option.collect_trt_shape = true;
+  tinypose_option.trt_option.SetShape("image", {1, 3, 256, 192}, {1, 3, 256, 192}, 
+                                      {1, 3, 256, 192});
+  auto tinypose_model = fastdeploy::vision::keypointdetection::PPTinyPose(
+      tinypose_model_file, tinypose_params_file, tinypose_config_file,
+      tinypose_option);
+  if (!tinypose_model.Initialized()) {
+    std::cerr << "TinyPose Model Failed to initialize." << std::endl;
+    return;
+  }
+
+  auto im = cv::imread(image_file);
+  fastdeploy::vision::KeyPointDetectionResult res;
+
+  auto pipeline =
+      fastdeploy::pipeline::PPTinyPose(
+          &det_model, &tinypose_model);
+  pipeline.detection_model_score_threshold = 0.5;
+  if (!pipeline.Predict(&im, &res)) {
+    std::cerr << "TinyPose Prediction Failed." << std::endl;
+    return;
+  } else {
+    std::cout << "TinyPose Prediction Done!" << std::endl;
+  }
+  
+  std::cout << res.Str() << std::endl;
+
+  auto vis_im =
+      fastdeploy::vision::VisKeypointDetection(im, res, 0.2);
+  cv::imwrite("vis_result.jpg", vis_im);
+  std::cout << "TinyPose visualized result saved in ./vis_result.jpg"
+            << std::endl;
+}
+
+int main(int argc, char* argv[]) {
+  if (argc < 5) {
+    std::cout << "Usage: infer_demo path/to/detection_model_dir "
+                 "path/to/pptinypose_model_dir path/to/image run_option, "
+                 "e.g ./infer_model ./picodet_model_dir ./pptinypose_model_dir "
+                 "./test.jpeg 0"
+              << std::endl;
+    std::cout << "The data type of run_option is int, 0: run with cpu; 1: run "
+                 "with gpu; 2: run with gpu and use tensorrt backend;"
+              << std::endl;
+    return -1;
+  }
+
+  if (std::atoi(argv[4]) == 0) {
+    CpuInfer(argv[1], argv[2], argv[3]);
+  } else if (std::atoi(argv[4]) == 1) {
+    GpuInfer(argv[1], argv[2], argv[3]);
+  } else if (std::atoi(argv[4]) == 2) {
+    TrtInfer(argv[1], argv[2], argv[3]);
+  }
+  return 0;
+}
--- a/paddle_detection/deploy/fastdeploy/cpu-gpu/cpp/infer.cc
+++ b/paddle_detection/deploy/fastdeploy/cpu-gpu/cpp/infer.cc
@@ -0,0 +1,134 @@
+// Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#include "fastdeploy/vision.h"
+
+#ifdef WIN32
+const char sep = '\\';
+#else
+const char sep = '/';
+#endif
+
+void CpuInfer(const std::string& model_dir, const std::string& image_file) {
+  auto model_file = model_dir + sep + "model.pdmodel";
+  auto params_file = model_dir + sep + "model.pdiparams";
+  auto config_file = model_dir + sep + "infer_cfg.yml";
+  auto option = fastdeploy::RuntimeOption();
+  option.UseCpu();
+  auto model = fastdeploy::vision::detection::PPYOLOE(model_file, params_file,
+                                                      config_file, option);
+  if (!model.Initialized()) {
+    std::cerr << "Failed to initialize." << std::endl;
+    return;
+  }
+
+  auto im = cv::imread(image_file);
+
+  fastdeploy::vision::DetectionResult res;
+  if (!model.Predict(im, &res)) {
+    std::cerr << "Failed to predict." << std::endl;
+    return;
+  }
+
+  std::cout << res.Str() << std::endl;
+  auto vis_im = fastdeploy::vision::VisDetection(im, res, 0.5);
+  cv::imwrite("vis_result.jpg", vis_im);
+  std::cout << "Visualized result saved in ./vis_result.jpg" << std::endl;
+}
+
+void GpuInfer(const std::string& model_dir, const std::string& image_file) {
+  auto model_file = model_dir + sep + "model.pdmodel";
+  auto params_file = model_dir + sep + "model.pdiparams";
+  auto config_file = model_dir + sep + "infer_cfg.yml";
+
+  auto option = fastdeploy::RuntimeOption();
+  option.UseGpu();
+  auto model = fastdeploy::vision::detection::PPYOLOE(model_file, params_file,
+                                                      config_file, option);
+  if (!model.Initialized()) {
+    std::cerr << "Failed to initialize." << std::endl;
+    return;
+  }
+
+  auto im = cv::imread(image_file);
+
+  fastdeploy::vision::DetectionResult res;
+  if (!model.Predict(im, &res)) {
+    std::cerr << "Failed to predict." << std::endl;
+    return;
+  }
+
+  std::cout << res.Str() << std::endl;
+  auto vis_im = fastdeploy::vision::VisDetection(im, res, 0.5);
+  cv::imwrite("vis_result.jpg", vis_im);
+  std::cout << "Visualized result saved in ./vis_result.jpg" << std::endl;
+}
+
+void TrtInfer(const std::string& model_dir, const std::string& image_file) {
+  auto model_file = model_dir + sep + "model.pdmodel";
+  auto params_file = model_dir + sep + "model.pdiparams";
+  auto config_file = model_dir + sep + "infer_cfg.yml";
+
+  auto option = fastdeploy::RuntimeOption();
+  option.UseGpu();
+  option.UsePaddleInferBackend();
+  // If use original Tensorrt, not Paddle-TensorRT,
+  // please try `option.UseTrtBackend()`
+  option.paddle_infer_option.enable_trt = true;
+  option.paddle_infer_option.collect_trt_shape = true;
+  option.trt_option.SetShape("image", {1, 3, 640, 640}, {1, 3, 640, 640}, 
+                                      {1, 3, 640, 640});
+  option.trt_option.SetShape("scale_factor", {1, 2}, {1, 2}, {1, 2});                        
+  auto model = fastdeploy::vision::detection::PPYOLOE(model_file, params_file,
+                                                      config_file, option);
+  if (!model.Initialized()) {
+    std::cerr << "Failed to initialize." << std::endl;
+    return;
+  }
+
+  auto im = cv::imread(image_file);
+
+  fastdeploy::vision::DetectionResult res;
+  if (!model.Predict(im, &res)) {
+    std::cerr << "Failed to predict." << std::endl;
+    return;
+  }
+
+  std::cout << res.Str() << std::endl;
+  auto vis_im = fastdeploy::vision::VisDetection(im, res, 0.5);
+  cv::imwrite("vis_result.jpg", vis_im);
+  std::cout << "Visualized result saved in ./vis_result.jpg" << std::endl;
+}
+
+int main(int argc, char* argv[]) {
+  if (argc < 4) {
+    std::cout
+        << "Usage: infer_demo path/to/model_dir path/to/image run_option, "
+           "e.g ./infer_demo ./ppyoloe_model_dir ./test.jpeg 0"
+        << std::endl;
+    std::cout << "The data type of run_option is int, 0: run with cpu; 1: run "
+                 "with gpu; 2: run with gpu and use tensorrt backend"
+              << std::endl;
+    return -1;
+  }
+
+  if (std::atoi(argv[3]) == 0) {
+    CpuInfer(argv[1], argv[2]);
+  } else if (std::atoi(argv[3]) == 1) {
+    GpuInfer(argv[1], argv[2]);
+  } else if (std::atoi(argv[3]) == 2) {
+    TrtInfer(argv[1], argv[2]);
+  }
+  return 0;
+}
--- a/paddle_detection/deploy/fastdeploy/cpu-gpu/cpp/pptinypose_infer.cc
+++ b/paddle_detection/deploy/fastdeploy/cpu-gpu/cpp/pptinypose_infer.cc
@@ -0,0 +1,149 @@
+// Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#include "fastdeploy/vision.h"
+
+#ifdef WIN32
+const char sep = '\\';
+#else
+const char sep = '/';
+#endif
+
+void CpuInfer(const std::string& tinypose_model_dir,
+              const std::string& image_file) {
+  auto tinypose_model_file = tinypose_model_dir + sep + "model.pdmodel";
+  auto tinypose_params_file = tinypose_model_dir + sep + "model.pdiparams";
+  auto tinypose_config_file = tinypose_model_dir + sep + "infer_cfg.yml";
+  auto option = fastdeploy::RuntimeOption();
+  option.UseCpu();
+  auto tinypose_model = fastdeploy::vision::keypointdetection::PPTinyPose(
+      tinypose_model_file, tinypose_params_file, tinypose_config_file, option);
+  if (!tinypose_model.Initialized()) {
+    std::cerr << "TinyPose Model Failed to initialize." << std::endl;
+    return;
+  }
+
+  auto im = cv::imread(image_file);
+  fastdeploy::vision::KeyPointDetectionResult res;
+  if (!tinypose_model.Predict(&im, &res)) {
+    std::cerr << "TinyPose Prediction Failed." << std::endl;
+    return;
+  } else {
+    std::cout << "TinyPose Prediction Done!" << std::endl;
+  }
+ 
+  std::cout << res.Str() << std::endl;
+
+  auto tinypose_vis_im =
+      fastdeploy::vision::VisKeypointDetection(im, res, 0.5);
+  cv::imwrite("tinypose_vis_result.jpg", tinypose_vis_im);
+  std::cout << "TinyPose visualized result saved in ./tinypose_vis_result.jpg"
+            << std::endl;
+}
+
+void GpuInfer(const std::string& tinypose_model_dir,
+              const std::string& image_file) {
+  auto option = fastdeploy::RuntimeOption();
+  option.UseGpu();
+
+  auto tinypose_model_file = tinypose_model_dir + sep + "model.pdmodel";
+  auto tinypose_params_file = tinypose_model_dir + sep + "model.pdiparams";
+  auto tinypose_config_file = tinypose_model_dir + sep + "infer_cfg.yml";
+  auto tinypose_model = fastdeploy::vision::keypointdetection::PPTinyPose(
+      tinypose_model_file, tinypose_params_file, tinypose_config_file, option);
+  if (!tinypose_model.Initialized()) {
+    std::cerr << "TinyPose Model Failed to initialize." << std::endl;
+    return;
+  }
+
+  auto im = cv::imread(image_file);
+  fastdeploy::vision::KeyPointDetectionResult res;
+  if (!tinypose_model.Predict(&im, &res)) {
+    std::cerr << "TinyPose Prediction Failed." << std::endl;
+    return;
+  } else {
+    std::cout << "TinyPose Prediction Done!" << std::endl;
+  }
+  
+  std::cout << res.Str() << std::endl;
+
+  auto tinypose_vis_im =
+      fastdeploy::vision::VisKeypointDetection(im, res, 0.5);
+  cv::imwrite("tinypose_vis_result.jpg", tinypose_vis_im);
+  std::cout << "TinyPose visualized result saved in ./tinypose_vis_result.jpg"
+            << std::endl;
+}
+
+void TrtInfer(const std::string& tinypose_model_dir,
+              const std::string& image_file) {
+  auto tinypose_model_file = tinypose_model_dir + sep + "model.pdmodel";
+  auto tinypose_params_file = tinypose_model_dir + sep + "model.pdiparams";
+  auto tinypose_config_file = tinypose_model_dir + sep + "infer_cfg.yml";
+  auto tinypose_option = fastdeploy::RuntimeOption();
+  tinypose_option.UseGpu();
+  tinypose_option.UsePaddleInferBackend();
+  // If use original Tensorrt, not Paddle-TensorRT,
+  // please try `option.UseTrtBackend()`
+  tinypose_option.paddle_infer_option.enable_trt = true;
+  tinypose_option.paddle_infer_option.collect_trt_shape = true;
+  tinypose_option.trt_option.SetShape("image", {1, 3, 256, 192}, {1, 3, 256, 192}, 
+                                      {1, 3, 256, 192});
+
+  auto tinypose_model = fastdeploy::vision::keypointdetection::PPTinyPose(
+      tinypose_model_file, tinypose_params_file, tinypose_config_file,
+      tinypose_option);
+  if (!tinypose_model.Initialized()) {
+    std::cerr << "TinyPose Model Failed to initialize." << std::endl;
+    return;
+  }
+
+  auto im = cv::imread(image_file);
+  fastdeploy::vision::KeyPointDetectionResult res;
+  if (!tinypose_model.Predict(&im, &res)) {
+    std::cerr << "TinyPose Prediction Failed." << std::endl;
+    return;
+  } else {
+    std::cout << "TinyPose Prediction Done!" << std::endl;
+  }
+  
+  std::cout << res.Str() << std::endl;
+
+  auto tinypose_vis_im =
+      fastdeploy::vision::VisKeypointDetection(im, res, 0.5);
+  cv::imwrite("tinypose_vis_result.jpg", tinypose_vis_im);
+  std::cout << "TinyPose visualized result saved in ./tinypose_vis_result.jpg"
+            << std::endl;
+}
+
+int main(int argc, char* argv[]) {
+  if (argc < 4) {
+    std::cout << "Usage: infer_demo path/to/pptinypose_model_dir path/to/image "
+                 "run_option, "
+                 "e.g ./infer_demo ./pptinypose_model_dir ./test.jpeg 0"
+              << std::endl;
+    std::cout << "The data type of run_option is int, 0: run with cpu; 1: run "
+                 "with gpu; 2: run with gpu and use tensorrt backend;"
+              << std::endl;
+    return -1;
+  }
+
+  if (std::atoi(argv[3]) == 0) {
+    CpuInfer(argv[1], argv[2]);
+  } else if (std::atoi(argv[3]) == 1) {
+    GpuInfer(argv[1], argv[2]);
+  } else if (std::atoi(argv[3]) == 2) {
+    TrtInfer(argv[1], argv[2]);
+  }
+  return 0;
+}
--- a/paddle_detection/deploy/fastdeploy/cpu-gpu/python/README.md
+++ b/paddle_detection/deploy/fastdeploy/cpu-gpu/python/README.md
@@ -0,0 +1,126 @@
+[English](README.md) | 简体中文
+# PaddleDetection CPU-GPU Python部署示例
+
+本目录下提供`infer.py`快速完成PPYOLOE模型包括PPYOLOE在CPU/GPU，以及GPU上通过Paddle-TensorRT加速部署的示例。 
+
+## 1. 说明  
+PaddleDetection支持利用FastDeploy在NVIDIA GPU、X86 CPU、飞腾CPU、ARM CPU、Intel GPU(独立显卡/集成显卡)硬件上快速部署PaddleDetection模型。FastDeploy目前支持的模型系列，包括但不限于`PPYOLOE`, `PicoDet`, `PaddleYOLOX`, `PPYOLO`, `FasterRCNN`，`SSD`,`PaddleYOLOv5`,`PaddleYOLOv6`,`PaddleYOLOv7`,`RTMDet`,`CascadeRCNN`,`PSSDet`,`RetinaNet`,`PPYOLOESOD`,`FCOS`,`TTFNet`,`TOOD`,`GFL`所有类名的构造函数和预测函数在参数上完全一致。所有模型的调用，只需要参考PPYOLOE的示例，即可快速调用。
+
+## 2. 部署环境准备  
+在部署前，需确认软硬件环境，同时下载预编译部署库，参考[FastDeploy安装文档](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install#FastDeploy预编译库安装)安装FastDeploy预编译库。
+
+## 3. 部署模型准备
+在部署前，请准备好您所需要运行的推理模型，你可以选择使用[预导出的推理模型](../README.md)或者[自行导出PaddleDetection部署模型](../README.md)。    
+
+## 4. 运行部署示例
+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)
+
+### 4.1 目标检测示例
+```bash
+# 下载部署示例代码
+git clone https://github.com/PaddlePaddle/PaddleDetection.git
+cd PaddleDetection/deploy/fastdeploy/cpu-gpu/python
+# 注意：如果当前分支找不到下面的fastdeploy测试代码，请切换到develop分支
+# git checkout develop
+
+# 下载PPYOLOE模型文件和测试图片
+wget https://bj.bcebos.com/paddlehub/fastdeploy/ppyoloe_crn_l_300e_coco.tgz
+wget https://gitee.com/paddlepaddle/PaddleDetection/raw/release/2.4/demo/000000014439.jpg
+tar xvf ppyoloe_crn_l_300e_coco.tgz
+
+# 运行部署示例
+# CPU推理
+python infer.py --model_dir ppyoloe_crn_l_300e_coco --image_file 000000014439.jpg --device cpu
+# GPU推理
+python infer.py --model_dir ppyoloe_crn_l_300e_coco --image_file 000000014439.jpg --device gpu
+# GPU上Paddle-TensorRT推理 （注意：TensorRT推理第一次运行，有序列化模型的操作，有一定耗时，需要耐心等待）
+python infer.py --model_dir ppyoloe_crn_l_300e_coco --image_file 000000014439.jpg --device gpu --use_trt True
+```
+
+运行完成可视化结果如下图所示
+<div  align="center">  
+<img src="https://user-images.githubusercontent.com/19339784/184326520-7075e907-10ed-4fad-93f8-52d0e35d4964.jpg", width=480px, height=320px />
+</div>
+
+### 4.2 关键点检测示例  
+```bash
+# 下载部署示例代码
+git clone https://github.com/PaddlePaddle/PaddleDetection.git
+cd PaddleDetection/deploy/fastdeploy/cpu-gpu/python
+# 注意：如果当前分支找不到下面的fastdeploy测试代码，请切换到develop分支
+# git checkout develop
+
+# 下载PP-TinyPose模型文件和测试图片
+wget https://bj.bcebos.com/paddlehub/fastdeploy/PP_TinyPose_256x192_infer.tgz
+tar -xvf PP_TinyPose_256x192_infer.tgz
+wget https://bj.bcebos.com/paddlehub/fastdeploy/hrnet_demo.jpg
+
+# 运行部署示例
+# CPU推理
+python pptinypose_infer.py --model_dir PP_TinyPose_256x192_infer --image_file hrnet_demo.jpg --device cpu
+# GPU推理
+python pptinypose_infer.py --model_dir PP_TinyPose_256x192_infer --image_file hrnet_demo.jpg --device gpu
+# GPU上Paddle-TensorRT推理 （注意：TensorRT推理第一次运行，有序列化模型的操作，有一定耗时，需要耐心等待）
+python pptinypose_infer.py --model_dir PP_TinyPose_256x192_infer --image_file hrnet_demo.jpg --device gpu --use_trt True
+```
+
+运行完成可视化结果如下图所示
+<div  align="center">  
+<img src="https://user-images.githubusercontent.com/16222477/196386764-dd51ad56-c410-4c54-9580-643f282f5a83.jpeg", width=359px, height=423px />
+</div>  
+
+关于如何进行多人关键点检测，请参考[PPTinyPose Pipeline示例](./det_keypoint_unite/)
+
+## 5. 部署示例选项说明  
+
+|参数|含义|默认值
+|---|---|---|  
+|--model_dir|指定模型文件夹所在的路径|None|
+|--image_file|指定测试图片所在的路径|None|  
+|--device|指定即将运行的硬件类型，支持的值为`[cpu, gpu]`，当设置为cpu时，可运行在x86 cpu/arm cpu等cpu上|cpu|
+|--use_trt|是否使用trt，该项只在device为gpu时有效|False|  
+
+## 6. PaddleDetection Python接口
+FastDeploy目前支持的模型系列，包括但不限于`PPYOLOE`, `PicoDet`, `PaddleYOLOX`, `PPYOLO`, `FasterRCNN`，`SSD`,`PaddleYOLOv5`,`PaddleYOLOv6`,`PaddleYOLOv7`,`RTMDet`,`CascadeRCNN`,`PSSDet`,`RetinaNet`,`PPYOLOESOD`,`FCOS`,`TTFNet`,`TOOD`,`GFL`所有类名的构造函数和预测函数在参数上完全一致。所有模型的调用，只需要参考PPYOLOE的示例，即可快速调用。 
+
+### 6.1 目标检测及实例分割模型
+```python
+fastdeploy.vision.detection.PPYOLOE(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.PicoDet(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.PaddleYOLOX(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.YOLOv3(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.PPYOLO(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.FasterRCNN(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.MaskRCNN(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.SSD(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.PaddleYOLOv5(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.PaddleYOLOv6(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.PaddleYOLOv7(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.RTMDet(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.CascadeRCNN(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.PSSDet(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.RetinaNet(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.PPYOLOESOD(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.FCOS(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.TTFNet(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.TOOD(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+fastdeploy.vision.detection.GFL(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+```
+### 6.2 关键点检测模型  
+```python
+fd.vision.keypointdetection.PPTinyPose(model_file, params_file, config_file, runtime_option=None, model_format=ModelFormat.PADDLE)
+```
+
+PaddleDetection模型加载和初始化，其中model_file， params_file为导出的Paddle部署模型格式, config_file为PaddleDetection同时导出的部署配置yaml文件
+
+## 7. 更多指南
+- [PaddleDetection Python API文档](https://www.paddlepaddle.org.cn/fastdeploy-api-doc/python/html/object_detection.html)
+- [FastDeploy部署PaddleDetection模型概览](../../)
+- [C++部署](../cpp)
+
+## 8. 常见问题
+- [如何切换模型推理后端引擎](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/faq/how_to_change_backend.md)
+- [Intel GPU(独立显卡/集成显卡)的使用](https://github.com/PaddlePaddle/FastDeploy/blob/develop/tutorials/intel_gpu/README.md)
+- [编译CPU部署库](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install/cpu.md)
+- [编译GPU部署库](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install/gpu.md)
+- [编译Jetson部署库](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install/jetson.md)
--- a/paddle_detection/deploy/fastdeploy/cpu-gpu/python/det_keypoint_unite/README.md
+++ b/paddle_detection/deploy/fastdeploy/cpu-gpu/python/det_keypoint_unite/README.md
@@ -0,0 +1,70 @@
+[English](README.md) | 简体中文
+# PP-PicoDet + PP-TinyPose (Pipeline) CPU-GPU Python部署示例
+
+本目录下提供`det_keypoint_unite_infer.py`快速完成多人模型配置 PP-PicoDet + PP-TinyPose 在CPU/GPU，以及GPU上通过TensorRT加速部署的`单图多人关键点检测`示例。执行如下脚本即可完成.**注意**: PP-TinyPose单模型独立部署，请参考[PP-TinyPose 单模型](../README.md)
+
+## 1. 部署环境准备  
+在部署前，需确认软硬件环境，同时下载预编译部署库，参考[FastDeploy安装文档](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install#FastDeploy预编译库安装)安装FastDeploy预编译库。
+
+## 2. 部署模型准备
+在部署前，请准备好您所需要运行的推理模型，你可以选择使用[预导出的推理模型](../../README.md)或者[自行导出PaddleDetection部署模型](../../README.md)。  
+
+## 3. 运行部署示例
+
+```bash
+# 下载部署示例代码
+git clone https://github.com/PaddlePaddle/PaddleDetection.git
+cd PaddleDetection/deploy/fastdeploy/cpu-gpu/python/det_keypoint_unite
+# 注意：如果当前分支找不到下面的fastdeploy测试代码，请切换到develop分支
+# git checkout develop
+
+# 下载PP-TinyPose模型文件和测试图片
+wget https://bj.bcebos.com/paddlehub/fastdeploy/PP_TinyPose_256x192_infer.tgz
+tar -xvf PP_TinyPose_256x192_infer.tgz
+wget https://bj.bcebos.com/paddlehub/fastdeploy/PP_PicoDet_V2_S_Pedestrian_320x320_infer.tgz
+tar -xvf PP_PicoDet_V2_S_Pedestrian_320x320_infer.tgz
+wget https://bj.bcebos.com/paddlehub/fastdeploy/000000018491.jpg
+# CPU推理
+python det_keypoint_unite_infer.py --tinypose_model_dir PP_TinyPose_256x192_infer --det_model_dir PP_PicoDet_V2_S_Pedestrian_320x320_infer --image_file 000000018491.jpg --device cpu
+# GPU推理
+python det_keypoint_unite_infer.py --tinypose_model_dir PP_TinyPose_256x192_infer --det_model_dir PP_PicoDet_V2_S_Pedestrian_320x320_infer --image_file 000000018491.jpg --device gpu
+# GPU上Paddle-TensorRT推理（注意：TensorRT推理第一次运行，有序列化模型的操作，有一定耗时，需要耐心等待）
+python det_keypoint_unite_infer.py --tinypose_model_dir PP_TinyPose_256x192_infer --det_model_dir PP_PicoDet_V2_S_Pedestrian_320x320_infer --image_file 000000018491.jpg --device gpu --use_trt True
+```
+
+运行完成可视化结果如下图所示
+<div  align="center">  
+<img src="https://user-images.githubusercontent.com/16222477/196393343-eeb6b68f-0bc6-4927-871f-5ac610da7293.jpeg", width=640px, height=427px />
+</div>
+
+- 关于如何通过FastDeploy使用更多不同的推理后端，以及如何使用不同的硬件，请参考文档：[如何切换模型推理后端引擎](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/faq/how_to_change_backend.md)
+
+## 4. 部署示例选项说明  
+
+|参数|含义|默认值
+|---|---|---|  
+|--tinypose_model_dir|指定关键点模型文件夹所在的路径|None|
+|--det_model_dir|指定目标模型文件夹所在的路径|None|
+|--image_file|指定测试图片所在的路径|None|  
+|--device|指定即将运行的硬件类型，支持的值为`[cpu, gpu]`，当设置为cpu时，可运行在x86 cpu/arm cpu等cpu上|cpu|
+|--use_trt|是否使用trt，该项只在device为gpu时有效|False|  
+
+## 5. PPTinyPose 模型串联 Python接口
+
+```python
+fd.pipeline.PPTinyPose(det_model=None, pptinypose_model=None)
+```
+
+PPTinyPose Pipeline 模型加载和初始化，其中det_model是使用`fd.vision.detection.PicoDet`初始化的检测模型，pptinypose_model是使用`fd.vision.keypointdetection.PPTinyPose`初始化的关键点检测模型。
+
+## 6. 更多指南
+- [PaddleDetection Python API文档](https://www.paddlepaddle.org.cn/fastdeploy-api-doc/python/html/object_detection.html)
+- [FastDeploy部署PaddleDetection模型概览](../../../)
+- [C++部署](../../cpp/)
+
+## 7. 常见问题
+- [如何切换模型推理后端引擎](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/faq/how_to_change_backend.md)
+- [Intel GPU(独立显卡/集成显卡)的使用](https://github.com/PaddlePaddle/FastDeploy/blob/develop/tutorials/intel_gpu/README.md)
+- [编译CPU部署库](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install/cpu.md)
+- [编译GPU部署库](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install/gpu.md)
+- [编译Jetson部署库](https://github.com/PaddlePaddle/FastDeploy/blob/develop/docs/cn/build_and_install/jetson.md)
--- a/paddle_detection/deploy/fastdeploy/cpu-gpu/python/det_keypoint_unite/det_keypoint_unite_infer.py
+++ b/paddle_detection/deploy/fastdeploy/cpu-gpu/python/det_keypoint_unite/det_keypoint_unite_infer.py
@@ -0,0 +1,101 @@
+import fastdeploy as fd
+import cv2
+import os
+
+
+def parse_arguments():
+    import argparse
+    import ast
+    parser = argparse.ArgumentParser()
+    parser.add_argument(
+        "--tinypose_model_dir",
+        required=True,
+        help="path of paddletinypose model directory")
+    parser.add_argument(
+        "--det_model_dir", help="path of paddledetection model directory")
+    parser.add_argument(
+        "--image_file", required=True, help="path of test image file.")
+    parser.add_argument(
+        "--device",
+        type=str,
+        default='cpu',
+        help="type of inference device, support 'cpu' or 'gpu'.")
+    parser.add_argument(
+        "--use_trt",
+        type=ast.literal_eval,
+        default=False,
+        help="wether to use tensorrt.")
+    return parser.parse_args()
+
+
+def build_picodet_option(args):
+    option = fd.RuntimeOption()
+
+    if args.device.lower() == "gpu":
+        option.use_gpu()
+
+    if args.use_trt:
+        option.use_paddle_infer_backend()
+        # If use original Tensorrt, not Paddle-TensorRT,
+        # please try `option.use_trt_backend()`
+        option.paddle_infer_option.enable_trt = True
+        option.paddle_infer_option.collect_trt_shape = True
+        option.trt_option.set_shape("image", [1, 3, 320, 320], [1, 3, 320, 320],
+                                             [1, 3, 320, 320])
+        option.trt_option.set_shape("scale_factor", [1, 2], [1, 2], [1, 2])                                     
+    return option
+
+
+def build_tinypose_option(args):
+    option = fd.RuntimeOption()
+
+    if args.device.lower() == "gpu":
+        option.use_gpu()
+
+    if args.use_trt:
+        option.use_paddle_infer_backend()
+        # If use original Tensorrt, not Paddle-TensorRT,
+        # please try `option.use_trt_backend()`
+        option.paddle_infer_option.enable_trt = True
+        option.paddle_infer_option.collect_trt_shape = True
+        option.trt_option.set_shape("image", [1, 3, 256, 192], [1, 3, 256, 192],
+                                             [1, 3, 256, 192])
+    return option
+
+
+args = parse_arguments()
+picodet_model_file = os.path.join(args.det_model_dir, "model.pdmodel")
+picodet_params_file = os.path.join(args.det_model_dir, "model.pdiparams")
+picodet_config_file = os.path.join(args.det_model_dir, "infer_cfg.yml")
+
+# setup runtime
+runtime_option = build_picodet_option(args)
+det_model = fd.vision.detection.PicoDet(
+    picodet_model_file,
+    picodet_params_file,
+    picodet_config_file,
+    runtime_option=runtime_option)
+
+tinypose_model_file = os.path.join(args.tinypose_model_dir, "model.pdmodel")
+tinypose_params_file = os.path.join(args.tinypose_model_dir, "model.pdiparams")
+tinypose_config_file = os.path.join(args.tinypose_model_dir, "infer_cfg.yml")
+# setup runtime
+runtime_option = build_tinypose_option(args)
+tinypose_model = fd.vision.keypointdetection.PPTinyPose(
+    tinypose_model_file,
+    tinypose_params_file,
+    tinypose_config_file,
+    runtime_option=runtime_option)
+
+# predict
+im = cv2.imread(args.image_file)
+pipeline = fd.pipeline.PPTinyPose(det_model, tinypose_model)
+pipeline.detection_model_score_threshold = 0.5
+pipeline_result = pipeline.predict(im)
+print("Paddle TinyPose Result:\n", pipeline_result)
+
+# visualize
+vis_im = fd.vision.vis_keypoint_detection(
+    im, pipeline_result, conf_threshold=0.2)
+cv2.imwrite("visualized_result.jpg", vis_im)
+print("TinyPose visualized result save in ./visualized_result.jpg")
--- a/paddle_detection/deploy/fastdeploy/cpu-gpu/python/infer.py
+++ b/paddle_detection/deploy/fastdeploy/cpu-gpu/python/infer.py
@@ -0,0 +1,74 @@
+import cv2
+import os
+
+import fastdeploy as fd
+
+
+def parse_arguments():
+    import argparse
+    import ast
+    parser = argparse.ArgumentParser()
+    parser.add_argument(
+        "--model_dir", required=True, help="Path of PaddleDetection model.")
+    parser.add_argument(
+        "--image_file", type=str, required=True, help="Path of test image file.")
+    parser.add_argument(
+        "--device",
+        type=str,
+        default='cpu',
+        help="Type of inference device, support, 'cpu' or 'gpu'.")
+    parser.add_argument(
+        "--use_trt",
+        type=ast.literal_eval,
+        default=False,
+        help="Wether to use tensorrt.")
+    return parser.parse_args()
+
+
+def build_option(args):
+    option = fd.RuntimeOption()
+
+    if args.device.lower() == "gpu":
+        option.use_gpu()
+
+    if args.use_trt:
+        option.use_paddle_infer_backend()
+        # If use original Tensorrt, not Paddle-TensorRT,
+        # please try `option.use_trt_backend()`
+        option.paddle_infer_option.enable_trt = True
+        option.paddle_infer_option.collect_trt_shape = True
+        option.trt_option.set_shape("image", [1, 3, 640, 640], [1, 3, 640, 640],
+                                             [1, 3, 640, 640])
+        option.trt_option.set_shape("scale_factor", [1, 2], [1, 2], [1, 2])
+    return option
+
+
+args = parse_arguments()
+
+if args.model_dir is None:
+    model_dir = fd.download_model(name='ppyoloe_crn_l_300e_coco')
+else:
+    model_dir = args.model_dir
+
+model_file = os.path.join(model_dir, "model.pdmodel")
+params_file = os.path.join(model_dir, "model.pdiparams")
+config_file = os.path.join(model_dir, "infer_cfg.yml")
+
+# settting for runtime
+runtime_option = build_option(args)
+model = fd.vision.detection.PPYOLOE(
+    model_file, params_file, config_file, runtime_option=runtime_option)
+
+# predict
+if args.image_file is None:
+    image_file = fd.utils.get_detection_test_image()
+else:
+    image_file = args.image_file
+im = cv2.imread(image_file)
+result = model.predict(im)
+print(result)
+
+# visualize
+vis_im = fd.vision.vis_detection(im, result, score_threshold=0.5)
+cv2.imwrite("visualized_result.jpg", vis_im)
+print("Visualized result save in ./visualized_result.jpg")
--- a/paddle_detection/deploy/fastdeploy/cpu-gpu/python/pptinypose_infer.py
+++ b/paddle_detection/deploy/fastdeploy/cpu-gpu/python/pptinypose_infer.py
@@ -0,0 +1,67 @@
+import fastdeploy as fd
+import cv2
+import os
+
+
+def parse_arguments():
+    import argparse
+    import ast
+    parser = argparse.ArgumentParser()
+    parser.add_argument(
+        "--model_dir",
+        required=True,
+        help="path of PP-TinyPose model directory")
+    parser.add_argument(
+        "--image_file", required=True, help="path of test image file.")
+    parser.add_argument(
+        "--device",
+        type=str,
+        default='cpu',
+        help="type of inference device, support 'cpu', or 'gpu'.")
+    parser.add_argument(
+        "--use_trt",
+        type=ast.literal_eval,
+        default=False,
+        help="wether to use tensorrt.")
+    return parser.parse_args()
+
+
+def build_option(args):
+    option = fd.RuntimeOption()
+
+    if args.device.lower() == "gpu":
+        option.use_gpu()
+
+    if args.use_trt:
+        option.use_paddle_infer_backend()
+        # If use original Tensorrt, not Paddle-TensorRT,
+        # please try `option.use_trt_backend()`
+        option.paddle_infer_option.enable_trt = True
+        option.paddle_infer_option.collect_trt_shape = True
+        option.trt_option.set_shape("image", [1, 3, 256, 192], [1, 3, 256, 192],
+                                             [1, 3, 256, 192])
+    return option
+
+
+args = parse_arguments()
+
+tinypose_model_file = os.path.join(args.model_dir, "model.pdmodel")
+tinypose_params_file = os.path.join(args.model_dir, "model.pdiparams")
+tinypose_config_file = os.path.join(args.model_dir, "infer_cfg.yml")
+# setup runtime 
+runtime_option = build_option(args)
+tinypose_model = fd.vision.keypointdetection.PPTinyPose(
+    tinypose_model_file,
+    tinypose_params_file,
+    tinypose_config_file,
+    runtime_option=runtime_option)
+# predict
+im = cv2.imread(args.image_file)
+tinypose_result = tinypose_model.predict(im)
+print("Paddle TinyPose Result:\n", tinypose_result)
+
+# visualize
+vis_im = fd.vision.vis_keypoint_detection(
+    im, tinypose_result, conf_threshold=0.5)
+cv2.imwrite("visualized_result.jpg", vis_im)
+print("TinyPose visualized result save in ./visualized_result.jpg")