[Model] Add PIPNet and FaceLandmark1000 Support (#548)

* first commit for yolov7 * pybind for yolov7 * CPP README.md * CPP README.md * modified yolov7.cc * README.md * python file modify * delete license in fastdeploy/ * repush the conflict part * README.md modified * README.md modified * file path modified * file path modified * file path modified * file path modified * file path modified * README modified * README modified * move some helpers to private * add examples for yolov7 * api.md modified * api.md modified * api.md modified * YOLOv7 * yolov7 release link * yolov7 release link * yolov7 release link * copyright * change some helpers to private * change variables to const and fix documents. * gitignore * Transfer some funtions to private member of class * Transfer some funtions to private member of class * Merge from develop (#9) * Fix compile problem in different python version (#26) * fix some usage problem in linux * Fix compile problem Co-authored-by: root <root@bjyz-sys-gpu-kongming3.bjyz.baidu.com> * Add PaddleDetetion/PPYOLOE model support (#22) * add ppdet/ppyoloe * Add demo code and documents * add convert processor to vision (#27) * update .gitignore * Added checking for cmake include dir * fixed missing trt_backend option bug when init from trt * remove un-need data layout and add pre-check for dtype * changed RGB2BRG to BGR2RGB in ppcls model * add model_zoo yolov6 c++/python demo * fixed CMakeLists.txt typos * update yolov6 cpp/README.md * add yolox c++/pybind and model_zoo demo * move some helpers to private * fixed CMakeLists.txt typos * add normalize with alpha and beta * add version notes for yolov5/yolov6/yolox * add copyright to yolov5.cc * revert normalize * fixed some bugs in yolox * fixed examples/CMakeLists.txt to avoid conflicts * add convert processor to vision * format examples/CMakeLists summary * Fix bug while the inference result is empty with YOLOv5 (#29) * Add multi-label function for yolov5 * Update README.md Update doc * Update fastdeploy_runtime.cc fix variable option.trt_max_shape wrong name * Update runtime_option.md Update resnet model dynamic shape setting name from images to x * Fix bug when inference result boxes are empty * Delete detection.py Co-authored-by: Jason <jiangjiajun@baidu.com> Co-authored-by: root <root@bjyz-sys-gpu-kongming3.bjyz.baidu.com> Co-authored-by: DefTruth <31974251+DefTruth@users.noreply.github.com> Co-authored-by: huangjianhui <852142024@qq.com> * first commit for yolor * for merge * Develop (#11) * Fix compile problem in different python version (#26) * fix some usage problem in linux * Fix compile problem Co-authored-by: root <root@bjyz-sys-gpu-kongming3.bjyz.baidu.com> * Add PaddleDetetion/PPYOLOE model support (#22) * add ppdet/ppyoloe * Add demo code and documents * add convert processor to vision (#27) * update .gitignore * Added checking for cmake include dir * fixed missing trt_backend option bug when init from trt * remove un-need data layout and add pre-check for dtype * changed RGB2BRG to BGR2RGB in ppcls model * add model_zoo yolov6 c++/python demo * fixed CMakeLists.txt typos * update yolov6 cpp/README.md * add yolox c++/pybind and model_zoo demo * move some helpers to private * fixed CMakeLists.txt typos * add normalize with alpha and beta * add version notes for yolov5/yolov6/yolox * add copyright to yolov5.cc * revert normalize * fixed some bugs in yolox * fixed examples/CMakeLists.txt to avoid conflicts * add convert processor to vision * format examples/CMakeLists summary * Fix bug while the inference result is empty with YOLOv5 (#29) * Add multi-label function for yolov5 * Update README.md Update doc * Update fastdeploy_runtime.cc fix variable option.trt_max_shape wrong name * Update runtime_option.md Update resnet model dynamic shape setting name from images to x * Fix bug when inference result boxes are empty * Delete detection.py Co-authored-by: Jason <jiangjiajun@baidu.com> Co-authored-by: root <root@bjyz-sys-gpu-kongming3.bjyz.baidu.com> Co-authored-by: DefTruth <31974251+DefTruth@users.noreply.github.com> Co-authored-by: huangjianhui <852142024@qq.com> * Yolor (#16) * Develop (#11) (#12) * Fix compile problem in different python version (#26) * fix some usage problem in linux * Fix compile problem Co-authored-by: root <root@bjyz-sys-gpu-kongming3.bjyz.baidu.com> * Add PaddleDetetion/PPYOLOE model support (#22) * add ppdet/ppyoloe * Add demo code and documents * add convert processor to vision (#27) * update .gitignore * Added checking for cmake include dir * fixed missing trt_backend option bug when init from trt * remove un-need data layout and add pre-check for dtype * changed RGB2BRG to BGR2RGB in ppcls model * add model_zoo yolov6 c++/python demo * fixed CMakeLists.txt typos * update yolov6 cpp/README.md * add yolox c++/pybind and model_zoo demo * move some helpers to private * fixed CMakeLists.txt typos * add normalize with alpha and beta * add version notes for yolov5/yolov6/yolox * add copyright to yolov5.cc * revert normalize * fixed some bugs in yolox * fixed examples/CMakeLists.txt to avoid conflicts * add convert processor to vision * format examples/CMakeLists summary * Fix bug while the inference result is empty with YOLOv5 (#29) * Add multi-label function for yolov5 * Update README.md Update doc * Update fastdeploy_runtime.cc fix variable option.trt_max_shape wrong name * Update runtime_option.md Update resnet model dynamic shape setting name from images to x * Fix bug when inference result boxes are empty * Delete detection.py Co-authored-by: Jason <jiangjiajun@baidu.com> Co-authored-by: root <root@bjyz-sys-gpu-kongming3.bjyz.baidu.com> Co-authored-by: DefTruth <31974251+DefTruth@users.noreply.github.com> Co-authored-by: huangjianhui <852142024@qq.com> Co-authored-by: Jason <jiangjiajun@baidu.com> Co-authored-by: root <root@bjyz-sys-gpu-kongming3.bjyz.baidu.com> Co-authored-by: DefTruth <31974251+DefTruth@users.noreply.github.com> Co-authored-by: huangjianhui <852142024@qq.com> * Develop (#13) * Fix compile problem in different python version (#26) * fix some usage problem in linux * Fix compile problem Co-authored-by: root <root@bjyz-sys-gpu-kongming3.bjyz.baidu.com> * Add PaddleDetetion/PPYOLOE model support (#22) * add ppdet/ppyoloe * Add demo code and documents * add convert processor to vision (#27) * update .gitignore * Added checking for cmake include dir * fixed missing trt_backend option bug when init from trt * remove un-need data layout and add pre-check for dtype * changed RGB2BRG to BGR2RGB in ppcls model * add model_zoo yolov6 c++/python demo * fixed CMakeLists.txt typos * update yolov6 cpp/README.md * add yolox c++/pybind and model_zoo demo * move some helpers to private * fixed CMakeLists.txt typos * add normalize with alpha and beta * add version notes for yolov5/yolov6/yolox * add copyright to yolov5.cc * revert normalize * fixed some bugs in yolox * fixed examples/CMakeLists.txt to avoid conflicts * add convert processor to vision * format examples/CMakeLists summary * Fix bug while the inference result is empty with YOLOv5 (#29) * Add multi-label function for yolov5 * Update README.md Update doc * Update fastdeploy_runtime.cc fix variable option.trt_max_shape wrong name * Update runtime_option.md Update resnet model dynamic shape setting name from images to x * Fix bug when inference result boxes are empty * Delete detection.py Co-authored-by: Jason <jiangjiajun@baidu.com> Co-authored-by: root <root@bjyz-sys-gpu-kongming3.bjyz.baidu.com> Co-authored-by: DefTruth <31974251+DefTruth@users.noreply.github.com> Co-authored-by: huangjianhui <852142024@qq.com> * documents * documents * documents * documents * documents * documents * documents * documents * documents * documents * documents * documents * Develop (#14) * Fix compile problem in different python version (#26) * fix some usage problem in linux * Fix compile problem Co-authored-by: root <root@bjyz-sys-gpu-kongming3.bjyz.baidu.com> * Add PaddleDetetion/PPYOLOE model support (#22) * add ppdet/ppyoloe * Add demo code and documents * add convert processor to vision (#27) * update .gitignore * Added checking for cmake include dir * fixed missing trt_backend option bug when init from trt * remove un-need data layout and add pre-check for dtype * changed RGB2BRG to BGR2RGB in ppcls model * add model_zoo yolov6 c++/python demo * fixed CMakeLists.txt typos * update yolov6 cpp/README.md * add yolox c++/pybind and model_zoo demo * move some helpers to private * fixed CMakeLists.txt typos * add normalize with alpha and beta * add version notes for yolov5/yolov6/yolox * add copyright to yolov5.cc * revert normalize * fixed some bugs in yolox * fixed examples/CMakeLists.txt to avoid conflicts * add convert processor to vision * format examples/CMakeLists summary * Fix bug while the inference result is empty with YOLOv5 (#29) * Add multi-label function for yolov5 * Update README.md Update doc * Update fastdeploy_runtime.cc fix variable option.trt_max_shape wrong name * Update runtime_option.md Update resnet model dynamic shape setting name from images to x * Fix bug when inference result boxes are empty * Delete detection.py Co-authored-by: root <root@bjyz-sys-gpu-kongming3.bjyz.baidu.com> Co-authored-by: DefTruth <31974251+DefTruth@users.noreply.github.com> Co-authored-by: huangjianhui <852142024@qq.com> Co-authored-by: Jason <jiangjiajun@baidu.com> Co-authored-by: root <root@bjyz-sys-gpu-kongming3.bjyz.baidu.com> Co-authored-by: DefTruth <31974251+DefTruth@users.noreply.github.com> Co-authored-by: huangjianhui <852142024@qq.com> Co-authored-by: Jason <928090362@qq.com> * add is_dynamic for YOLO series (#22) * modify ppmatting backend and docs * modify ppmatting docs * fix the PPMatting size problem * fix LimitShort's log * retrigger ci * modify PPMatting docs * modify the way for dealing with LimitShort * add python comments for external models * modify resnet c++ comments * modify C++ comments for external models * modify python comments and add result class comments * fix comments compile error * modify result.h comments * c++ version for FaceLandmark1000 * add pipnet land1000 sigle test and python code * fix facelandmark1000 sigle test * fix python examples for PIPNet and FaceLandmark1000 * fix examples links for PIPNet and FaceLandmark1000 * modify test_vision_colorspace_convert.cc * modify facealign readme * retrigger ci * modify README * test ci * fix download_prebuilt_libraries.md * fix download_prebuilt_libraries.md * modify for comments * modify supported_num_landmarks * retrigger ci * check code style * check code style Co-authored-by: Jason <jiangjiajun@baidu.com> Co-authored-by: root <root@bjyz-sys-gpu-kongming3.bjyz.baidu.com> Co-authored-by: DefTruth <31974251+DefTruth@users.noreply.github.com> Co-authored-by: huangjianhui <852142024@qq.com> Co-authored-by: Jason <928090362@qq.com>
PaddlePaddle · Nov 16, 2022 · 18669b5 · 18669b5
1 parent c8e2f46
commit 18669b5
Show file tree

Hide file tree

Showing 31 changed files with 2,325 additions and 15 deletions.
diff --git a/docs/en/build_and_install/download_prebuilt_libraries.md b/docs/en/build_and_install/download_prebuilt_libraries.md
@@ -3,8 +3,8 @@
 FastDeploy provides pre-built libraries for developers to download and install directly. Meanwhile, FastDeploy also offers easy access to compile so that developers can compile FastDeploy according to their own needs.
 
 This article is divided into two parts:
-- [1.GPU Deployment Environment](##GPU Deployment Environment)
-- [2.CPU Deployment Environment](##CPU Deployment Environment)
+- [1.GPU Deployment Environment](#gpu-deployment-environment)
+- [2.CPU Deployment Environment](#cpu-deployment-environment)
 
 ## GPU Deployment Environment
 

diff --git a/examples/vision/facealign/README.md b/examples/vision/facealign/README.md
@@ -5,3 +5,5 @@ FastDeploy目前支持如下人脸对齐(关键点检测）模型部署
 | 模型 | 说明 | 模型格式 | 版本 |
 | :--- | :--- | :------- | :--- |
 | [Hsintao/pfld_106_face_landmarks](./pfld) | PFLD 系列模型 | ONNX | [CommitID:e150195](https://github.com/Hsintao/pfld_106_face_landmarks/commit/e150195) |
+| [Single430/FaceLandmark1000](./face_landmark_1000) | FaceLandmark1000 系列模型 | ONNX | [CommitID:1a951b6](https://github.com/Single430/FaceLandmark1000/tree/1a951b6) |
+| [jhb86253817/PIPNet](./pipnet) | PIPNet 系列模型 | ONNX | [CommitID:b9eab58](https://github.com/jhb86253817/PIPNet/tree/b9eab58) |
diff --git a/examples/vision/facealign/face_landmark_1000/README.md b/examples/vision/facealign/face_landmark_1000/README.md
@@ -0,0 +1,25 @@
+# FaceLandmark 模型部署
+
+## 模型版本说明
+
+- [FaceLandmark1000](https://github.com/Single430/FaceLandmark1000/tree/1a951b6)
+
+## 支持模型列表
+
+目前FastDeploy支持如下模型的部署
+
+- [FaceLandmark1000 模型](https://github.com/Single430/FaceLandmark1000)
+
+## 下载预训练模型
+
+为了方便开发者的测试，下面提供了FaceLandmark导出的各系列模型，开发者可直接下载使用。
+
+| 模型                                                               | 参数大小    | 精度    | 备注 |
+|:---------------------------------------------------------------- |:----- |:----- | :------ |
+| [FaceLandmark1000](https://bj.bcebos.com/paddlehub/fastdeploy/FaceLandmark1000.onnx) | 2.1M | - |
+
+
+## 详细部署文档
+
+- [Python部署](python)
+- [C++部署](cpp)
diff --git a/examples/vision/facealign/face_landmark_1000/cpp/CMakeLists.txt b/examples/vision/facealign/face_landmark_1000/cpp/CMakeLists.txt
@@ -0,0 +1,18 @@
+PROJECT(infer_demo C CXX)
+CMAKE_MINIMUM_REQUIRED (VERSION 3.10)
+
+# 指定下载解压后的fastdeploy库路径
+option(FASTDEPLOY_INSTALL_DIR "Path of downloaded fastdeploy sdk.")
+include(${FASTDEPLOY_INSTALL_DIR}/utils/gflags.cmake)
+include(${FASTDEPLOY_INSTALL_DIR}/FastDeploy.cmake)
+
+# 添加FastDeploy依赖头文件
+include_directories(${FASTDEPLOY_INCS})
+
+add_executable(infer_demo ${PROJECT_SOURCE_DIR}/infer.cc)
+# 添加FastDeploy库依赖
+if(UNIX AND (NOT APPLE) AND (NOT ANDROID))
+  target_link_libraries(infer_demo ${FASTDEPLOY_LIBS} gflags pthread)
+else()
+  target_link_libraries(infer_demo ${FASTDEPLOY_LIBS} gflags)
+endif()
diff --git a/examples/vision/facealign/face_landmark_1000/cpp/README.md b/examples/vision/facealign/face_landmark_1000/cpp/README.md
@@ -0,0 +1,84 @@
+# FaceLandmark1000 C++部署示例
+
+本目录下提供`infer.cc`快速完成FaceLandmark1000在CPU/GPU，以及GPU上通过TensorRT加速部署的示例。
+
+在部署前，需确认以下两个步骤
+
+- 1. 软硬件环境满足要求，参考[FastDeploy环境要求](../../../../../docs/cn/build_and_install/download_prebuilt_libraries.md)  
+- 2. 根据开发环境，下载预编译部署库和samples代码，参考[FastDeploy预编译库](../../../../../docs/cn/build_and_install/download_prebuilt_libraries.md)
+
+以Linux上CPU推理为例，在本目录执行如下命令即可完成编译测试，保证 FastDeploy 版本0.7.0以上(x.x.x >= 0.7.0)支持FaceLandmark1000模型
+
+```bash
+mkdir build
+cd build
+wget https://bj.bcebos.com/fastdeploy/release/cpp/fastdeploy-linux-x64-x.x.x.tgz
+tar xvf fastdeploy-linux-x64-x.x.x.tgz
+cmake .. -DFASTDEPLOY_INSTALL_DIR=${PWD}/fastdeploy-linux-x64-x.x.x
+make -j
+
+#下载官方转换好的 FaceLandmark1000 模型文件和测试图片
+wget https://bj.bcebos.com/paddlehub/fastdeploy/FaceLandmark1000.onnx
+wget https://bj.bcebos.com/paddlehub/fastdeploy/facealign_input.png
+
+# CPU推理
+./infer_demo --model FaceLandmark1000.onnx --image facealign_input.png --device cpu
+# GPU推理
+./infer_demo --model FaceLandmark1000.onnx --image facealign_input.png --device gpu
+# GPU上TensorRT推理
+./infer_demo --model FaceLandmark1000.onnx --image facealign_input.png --device gpu --backend trt
+```
+
+运行完成可视化结果如下图所示
+
+<div width="500">
+<img width="470" height="384" float="left" src="https://user-images.githubusercontent.com/67993288/200761309-90c096e2-c2f3-4140-8012-32ed84e5f389.jpg">
+</div>
+
+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)
+
+## FaceLandmark1000 C++接口
+
+### FaceLandmark1000 类
+
+```c++
+fastdeploy::vision::facealign::FaceLandmark1000(
+        const string& model_file,
+        const string& params_file = "",
+        const RuntimeOption& runtime_option = RuntimeOption(),
+        const ModelFormat& model_format = ModelFormat::ONNX)
+```
+
+FaceLandmark1000模型加载和初始化，其中model_file为导出的ONNX模型格式。
+
+**参数**
+
+> * **model_file**(str): 模型文件路径
+> * **params_file**(str): 参数文件路径，当模型格式为ONNX时，此参数传入空字符串即可
+> * **runtime_option**(RuntimeOption): 后端推理配置，默认为None，即采用默认配置
+> * **model_format**(ModelFormat): 模型格式，默认为ONNX格式
+
+#### Predict函数
+
+> ```c++
+> FaceLandmark1000::Predict(cv::Mat* im, FaceAlignmentResult* result)
+> ```
+>
+> 模型预测接口，输入图像直接输出landmarks结果。
+>
+> **参数**
+>
+> > * **im**: 输入图像，注意需为HWC，BGR格式
+> > * **result**: landmarks结果, FaceAlignmentResult说明参考[视觉模型预测结果](../../../../../docs/api/vision_results/)
+
+### 类成员变量
+
+用户可按照自己的实际需求，修改下列预处理参数，从而影响最终的推理和部署效果
+
+> > * **size**(vector&lt;int&gt;): 通过此参数修改预处理过程中resize的大小，包含两个整型元素，表示[width, height], 默认值为[128, 128]
+
+- [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)
diff --git a/examples/vision/facealign/face_landmark_1000/cpp/infer.cc b/examples/vision/facealign/face_landmark_1000/cpp/infer.cc
@@ -0,0 +1,110 @@
+// 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 "gflags/gflags.h"
+
+DEFINE_string(model, "", "Directory of the inference model.");
+DEFINE_string(image, "", "Path of the image file.");
+DEFINE_string(device, "cpu",
+              "Type of inference device, support 'cpu' or 'gpu'.");
+DEFINE_string(backend, "default",
+              "The inference runtime backend, support: ['default', 'ort', "
+              "'paddle', 'ov', 'trt', 'paddle_trt']");
+DEFINE_bool(use_fp16, false, "Whether to use FP16 mode, only support 'trt' and 'paddle_trt' backend");
+
+void PrintUsage() {
+  std::cout << "Usage: infer_demo --model model_path --image img_path --device [cpu|gpu] --backend "
+               "[default|ort|paddle|ov|trt|paddle_trt] "
+               "--use_fp16 false"
+            << std::endl;
+  std::cout << "Default value of device: cpu" << std::endl;
+  std::cout << "Default value of backend: default" << std::endl;
+  std::cout << "Default value of use_fp16: false" << std::endl;
+}
+
+bool CreateRuntimeOption(fastdeploy::RuntimeOption* option) {
+  if (FLAGS_device == "gpu") {
+    option->UseGpu();
+    if (FLAGS_backend == "ort") {
+      option->UseOrtBackend();
+    } else if (FLAGS_backend == "paddle") {
+      option->UsePaddleBackend();
+    } else if (FLAGS_backend == "trt" || 
+               FLAGS_backend == "paddle_trt") {
+      option->UseTrtBackend();
+      option->SetTrtInputShape("input", {1, 3, 128, 128});
+      if (FLAGS_backend == "paddle_trt") {
+        option->EnablePaddleToTrt();
+      }
+      if (FLAGS_use_fp16) {
+        option->EnableTrtFP16();
+      }
+    } else if (FLAGS_backend == "default") {
+      return true;
+    } else {
+      std::cout << "While inference with GPU, only support default/ort/paddle/trt/paddle_trt now, " << FLAGS_backend << " is not supported." << std::endl;
+      return false;
+    }
+  } else if (FLAGS_device == "cpu") {
+    if (FLAGS_backend == "ort") {
+      option->UseOrtBackend();
+    } else if (FLAGS_backend == "ov") {
+      option->UseOpenVINOBackend();
+    } else if (FLAGS_backend == "paddle") {
+      option->UsePaddleBackend();
+    } else if (FLAGS_backend == "default") {
+      return true;
+    } else {
+      std::cout << "While inference with CPU, only support default/ort/ov/paddle now, " << FLAGS_backend << " is not supported." << std::endl;
+      return false;
+    }
+  } else {
+    std::cerr << "Only support device CPU/GPU now, "  << FLAGS_device << " is not supported." << std::endl;
+    return false;
+  }
+
+  return true;
+}
+
+int main(int argc, char* argv[]) {
+  google::ParseCommandLineFlags(&argc, &argv, true);
+  auto option = fastdeploy::RuntimeOption();
+  if (!CreateRuntimeOption(&option)) {
+    PrintUsage();
+    return -1;
+  }
+
+  auto model = fastdeploy::vision::facealign::FaceLandmark1000(FLAGS_model, "", option);
+  if (!model.Initialized()) {
+    std::cerr << "Failed to initialize." << std::endl;
+    return -1;
+  }
+
+  auto im = cv::imread(FLAGS_image);
+  auto im_bak = im.clone();
+
+  fastdeploy::vision::FaceAlignmentResult res;
+  if (!model.Predict(&im, &res)) {
+    std::cerr << "Failed to predict." << std::endl;
+    return -1;
+  }
+  std::cout << res.Str() << std::endl;
+
+  auto vis_im = fastdeploy::vision::VisFaceAlignment(im_bak, res);
+  cv::imwrite("vis_result.jpg", vis_im);
+  std::cout << "Visualized result saved in ./vis_result.jpg" << std::endl;
+
+  return 0;
+}
diff --git a/examples/vision/facealign/face_landmark_1000/python/README.md b/examples/vision/facealign/face_landmark_1000/python/README.md
@@ -0,0 +1,71 @@
+# FaceLandmark1000 Python部署示例
+
+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. FastDeploy Python whl包安装，参考[FastDeploy Python安装](../../../../../docs/cn/build_and_install/download_prebuilt_libraries.md)
+
+本目录下提供`infer.py`快速完成FaceLandmark1000在CPU/GPU，以及GPU上通过TensorRT加速部署的示例，保证 FastDeploy 版本 >= 0.7.0 支持FaceLandmark1000模型。执行如下脚本即可完成
+
+```bash
+#下载部署示例代码
+git clone https://github.com/PaddlePaddle/FastDeploy.git
+cd FastDeploy/examples/vision/facealign/facelandmark1000/python
+
+# 下载FaceLandmark1000模型文件和测试图片
+## 原版ONNX模型
+wget https://bj.bcebos.com/paddlehub/fastdeploy/FaceLandmark1000.onnx
+wget https://bj.bcebos.com/paddlehub/fastdeploy/facealign_input.png
+
+# CPU推理
+python infer.py --model FaceLandmark1000.onnx --image facealign_input.png --device cpu
+# GPU推理
+python infer.py --model FaceLandmark1000.onnx --image facealign_input.png --device gpu
+# TRT推理
+python infer.py --model FaceLandmark1000.onnx --image facealign_input.png --device gpu --backend trt
+```
+
+运行完成可视化结果如下图所示
+
+<div width="500">
+<img width="470" height="384" float="left" src="https://user-images.githubusercontent.com/67993288/200761309-90c096e2-c2f3-4140-8012-32ed84e5f389.jpg">
+</div>
+
+## FaceLandmark1000 Python接口
+
+```python
+fd.vision.facealign.FaceLandmark1000(model_file, params_file=None, runtime_option=None, model_format=ModelFormat.ONNX)
+```
+
+FaceLandmark1000模型加载和初始化，其中model_file为导出的ONNX模型格式
+
+**参数**
+
+> * **model_file**(str): 模型文件路径
+> * **params_file**(str): 参数文件路径，当模型格式为ONNX格式时，此参数无需设定
+> * **runtime_option**(RuntimeOption): 后端推理配置，默认为None，即采用默认配置
+> * **model_format**(ModelFormat): 模型格式，默认为ONNX
+
+### predict函数
+
+> ```python
+> FaceLandmark1000.predict(input_image)
+> ```
+>
+> 模型预测结口，输入图像直接输出landmarks坐标结果。
+>
+> **参数**
+>
+> > * **input_image**(np.ndarray): 输入数据，注意需为HWC，BGR格式
+
+> **返回**
+>
+> > 返回`fastdeploy.vision.FaceAlignmentResult`结构体，结构体说明参考文档[视觉模型预测结果](../../../../../docs/api/vision_results/)
+
+
+## 其它文档
+
+- [FaceLandmark1000 模型介绍](..)
+- [FaceLandmark1000 C++部署](../cpp)
+- [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)