# FastDeploy **Repository Path**: SKYMIRRORA/FastDeploy ## Basic Information - **Project Name**: FastDeploy - **Description**: Easy-to-use Deep Learning Model Deployment Toolkit - **Primary Language**: Python - **License**: Apache-2.0 - **Default Branch**: develop - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 0 - **Forks**: 37 - **Created**: 2022-11-01 - **Last Updated**: 2022-11-01 ## Categories & Tags **Categories**: Uncategorized **Tags**: None ## README [English](README.md) | 简体中文 ![⚡️FastDeploy](https://user-images.githubusercontent.com/31974251/185771818-5d4423cd-c94c-4a49-9894-bc7a8d1c29d0.png)

**⚡️FastDeploy**是一款**易用高效**的推理部署开发套件。覆盖业界🔥**热门AI模型**并提供📦**开箱即用**的部署体验，包括图像分类、目标检测、图像分割、人脸检测、人脸识别、人体关键点识别、文字识别、语义理解等多任务，满足开发者**多场景**，**多硬件**、**多平台**的产业部署需求。 | Potrait Segmentation | Image Matting | Semantic Segmentation | Real-Time Matting | |:---------------------------------------------------------------------------------------------------------:|:---------------------------------------------------------------------------------------------------------:|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------:|:-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------:| | | | | | | **OCR** | **Behavior Recognition** | **Object Detection** |**Pose Estimation** | | | | | | **Face Alignment** | **3D Object Detection** | **Face Editing** | **Image Animation** | | | | ## 近期更新 - 🔥 **2022.10.15：Release FastDeploy [release v0.3.0](https://github.com/PaddlePaddle/FastDeploy/tree/release%2F0.3.0)**
- **New server-side deployment upgrade:更快的推理性能，一键量化，更多的视觉和NLP模型** - 集成 OpenVINO 推理引擎，并且保证了使用 OpenVINO 与 使用 TensorRT、ONNX Runtime、 Paddle Inference一致的开发体验； - 提供[一键模型量化工具](tools/quantization)，支持YOLOv7、YOLOv6、YOLOv5等视觉模型，在CPU和GPU推理速度可提升1.5～2倍； - 新增加 PP-OCRv3, PP-OCRv2, PP-Matting, PP-HumanMatting, ModNet 等视觉模型并提供[端到端部署示例](examples/vision)； - 新增加NLP信息抽取模型 UIE 并提供[端到端部署示例](examples/text/uie). - - 🔥 **2022.8.18：发布FastDeploy [release v0.2.0](https://github.com/PaddlePaddle/FastDeploy/tree/release%2F0.2.0)**
- **服务端部署全新升级：更快的推理性能，更多的视觉模型支持** - 发布基于x86 CPU、NVIDIA GPU的高性能推理引擎SDK，推理速度大幅提升 - 集成Paddle Inference、ONNX Runtime、TensorRT等推理引擎并提供统一的部署体验 - 支持YOLOv7、YOLOv6、YOLOv5、PP-YOLOE等全系列目标检测模型并提供[端到端部署示例](examples/vision/detection/) - 支持人脸检测、人脸识别、实时人像抠图、图像分割等40+重点模型及[Demo示例](examples/vision/) - 支持Python和C++两种语言部署 - **边缘移动端部署新增瑞芯微、晶晨、恩智浦等NPU芯片部署能力** - 发布轻量化目标检测[Picodet-NPU部署Demo](https://github.com/PaddlePaddle/Paddle-Lite-Demo/tree/develop/object_detection/linux/picodet_detection)，提供极致的INT8全量化推理能力 ## 目录 * **服务端部署** * [Python SDK快速开始](#fastdeploy-quick-start-python) * [C++ SDK快速开始](#fastdeploy-quick-start-cpp) * [服务端模型支持列表](#fastdeploy-server-models) * **端侧部署** * [EasyEdge边缘端部署](#fastdeploy-edge-sdk-arm-linux) * [EasyEdge移动端部署](#fastdeploy-edge-sdk-ios-android) * [EasyEdge自定义模型部署](#fastdeploy-edge-sdk-custom) * [Paddle Lite NPU部署](#fastdeploy-edge-sdk-npu) * [端侧模型支持列表](#fastdeploy-edge-sdk) * [社区交流](#fastdeploy-community) * [Acknowledge](#fastdeploy-acknowledge) * [License](#fastdeploy-license) ## 服务端部署 ### Python SDK快速开始
#### 快速安装 ##### 前置依赖 - CUDA >= 11.2 - cuDNN >= 8.0 - python >= 3.6 - OS: Linux x86_64/macOS/Windows 10 ##### 安装GPU版本 ```bash pip install numpy opencv-python fastdeploy-gpu-python -f https://www.paddlepaddle.org.cn/whl/fastdeploy.html ``` ##### [Conda安装(推荐)](docs/quick_start/Python_prebuilt_wheels.md) ```bash conda config --add channels conda-forge && conda install cudatoolkit=11.2 cudnn=8.2 ``` ##### 安装CPU版本 ```bash pip install numpy opencv-python fastdeploy-python -f https://www.paddlepaddle.org.cn/whl/fastdeploy.html ``` #### Python 推理示例 * 准备模型和图片 ```bash wget https://bj.bcebos.com/paddlehub/fastdeploy/ppyoloe_crn_l_300e_coco.tgz tar xvf ppyoloe_crn_l_300e_coco.tgz wget https://gitee.com/paddlepaddle/PaddleDetection/raw/release/2.4/demo/000000014439.jpg ``` * 测试推理结果 ```python # GPU/TensorRT部署参考 examples/vision/detection/paddledetection/python import cv2 import fastdeploy.vision as vision model = vision.detection.PPYOLOE("ppyoloe_crn_l_300e_coco/model.pdmodel", "ppyoloe_crn_l_300e_coco/model.pdiparams", "ppyoloe_crn_l_300e_coco/infer_cfg.yml") im = cv2.imread("000000014439.jpg") result = model.predict(im.copy()) print(result) vis_im = vision.vis_detection(im, result, score_threshold=0.5) cv2.imwrite("vis_image.jpg", vis_im) ```
C++ SDK快速开始 ### C++ SDK快速开始
#### 安装 - 参考[C++预编译库下载](docs/quick_start/CPP_prebuilt_libraries.md)文档 #### C++ 推理示例 * 准备模型和图片 ```bash wget https://bj.bcebos.com/paddlehub/fastdeploy/ppyoloe_crn_l_300e_coco.tgz tar xvf ppyoloe_crn_l_300e_coco.tgz wget https://gitee.com/paddlepaddle/PaddleDetection/raw/release/2.4/demo/000000014439.jpg ``` * 测试推理结果 ```C++ // GPU/TensorRT部署参考 examples/vision/detection/paddledetection/cpp #include "fastdeploy/vision.h" int main(int argc, char* argv[]) { namespace vision = fastdeploy::vision; auto model = vision::detection::PPYOLOE("ppyoloe_crn_l_300e_coco/model.pdmodel", "ppyoloe_crn_l_300e_coco/model.pdiparams", "ppyoloe_crn_l_300e_coco/infer_cfg.yml"); auto im = cv::imread("000000014439.jpg"); vision::DetectionResult res; model.Predict(&im, &res); auto vis_im = vision::Visualize::VisDetection(im, res, 0.5); cv::imwrite("vis_image.jpg", vis_im); return 0; } ```
更多部署案例请参考[视觉模型部署示例](examples/vision) . ### 服务端和web模型支持列表 🔥🔥🔥
符号说明: (1) ✅: 已经支持; (2) ❔: 未来支持; (3) ❌: 暂不支持; (4) --: 暂不考虑;
链接说明：「模型列」会跳转到模型推理Demo代码 | 任务场景 | 模型 | API | Linux | Linux | Win | Win | Mac | Mac | Linux | Linux | [web_demo](examples/application/js/web_demo) | [mini_program](examples/application/js/mini_program) | |:-----------------------------:|:---------------------------------------------------------------------------------------:|:---------------------------------------------------------------------------------------------------------------------------------:|:---------------------:|:------------------------:|:------------------------:|:------------------------:|:-----------------------:|:---------------------:|:--------------------------:|:---------------------------:|:--------------------------:|:---------------------------:| | --- | --- | --- | X86 CPU | NVIDIA GPU | Intel CPU | NVIDIA GPU | Intel CPU | Arm CPU | AArch64 CPU | NVIDIA Jetson |[Paddle.js](examples/application/js)| [Paddle.js](examples/application/js)| | Classification | [PaddleClas/ResNet50](./examples/vision/classification/paddleclas) | [Python](./examples/vision/classification/paddleclas/python)/[C++](./examples/vision/classification/paddleclas/cpp) |✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Classification | [PaddleClas/PP-LCNet](./examples/vision/classification/paddleclas) | [Python](./examples/vision/classification/paddleclas/python)/[C++](./examples/vision/classification/paddleclas/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Classification | [PaddleClas/PP-LCNetv2](./examples/vision/classification/paddleclas) | [Python](./examples/vision/classification/paddleclas/python)/[C++](./examples/vision/classification/paddleclas/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Classification | [PaddleClas/EfficientNet](./examples/vision/classification/paddleclas) | [Python](./examples/vision/classification/paddleclas/python)/[C++](./examples/vision/classification/paddleclas/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Classification | [PaddleClas/GhostNet](./examples/vision/classification/paddleclas) | [Python](./examples/vision/classification/paddleclas/python)/[C++](./examples/vision/classification/paddleclas/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Classification | [PaddleClas/MobileNetV1](./examples/vision/classification/paddleclas) | [Python](./examples/vision/classification/paddleclas/python)/[C++](./examples/vision/classification/paddleclas/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Classification | [PaddleClas/MobileNetV2](./examples/vision/classification/paddleclas) | [Python](./examples/vision/classification/paddleclas/python)/[C++](./examples/vision/classification/paddleclas/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |✅|✅| | Classification | [PaddleClas/MobileNetV3](./examples/vision/classification/paddleclas) | [Python](./examples/vision/classification/paddleclas/python)/[C++](./examples/vision/classification/paddleclas/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Classification | [PaddleClas/ShuffleNetV2](./examples/vision/classification/paddleclas) | [Python](./examples/vision/classification/paddleclas/python)/[C++](./examples/vision/classification/paddleclas/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Classification | [PaddleClas/SqueeezeNetV1.1](./examples/vision/classification/paddleclas) | [Python](./examples/vision/classification/paddleclas/python)/[C++](./examples/vision/classification/paddleclas/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Classification | [PaddleClas/Inceptionv3](./examples/vision/classification/paddleclas) | [Python](./examples/vision/classification/paddleclas/python)/[C++](./examples/vision/classification/paddleclas/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Classification | [PaddleClas/PP-HGNet](./examples/vision/classification/paddleclas) | [Python](./examples/vision/classification/paddleclas/python)/[C++](./examples/vision/classification/paddleclas/cpp) |✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Classification | [PaddleClas/SwinTransformer](./examples/vision/classification/paddleclas) | [Python](./examples/vision/classification/paddleclas/python)/[C++](./examples/vision/classification/paddleclas/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Detection | [PaddleDetection/PP-YOLOE](./examples/vision/detection/paddledetection) | [Python](./examples/vision/detection/paddledetection/python)/[C++](./examples/vision/detection/paddledetection/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Detection | [PaddleDetection/PicoDet](./examples/vision/detection/paddledetection) | [Python](./examples/vision/detection/paddledetection/python)/[C++](./examples/vision/detection/paddledetection/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Detection | [PaddleDetection/YOLOX](./examples/vision/detection/paddledetection) | [Python](./examples/vision/detection/paddledetection/python)/[C++](./examples/vision/detection/paddledetection/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Detection | [PaddleDetection/YOLOv3](./examples/vision/detection/paddledetection) | [Python](./examples/vision/detection/paddledetection/python)/[C++](./examples/vision/detection/paddledetection/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Detection | [PaddleDetection/PP-YOLO](./examples/vision/detection/paddledetection) | [Python](./examples/vision/detection/paddledetection/python)/[C++](./examples/vision/detection/paddledetection/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ | ❌ | ❌ |❔|❔| | Detection | [PaddleDetection/PP-YOLOv2](./examples/vision/detection/paddledetection) | [Python](./examples/vision/detection/paddledetection/python)/[C++](./examples/vision/detection/paddledetection/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ | ❌ | ❌ |❔|❔| | Detection | [PaddleDetection/FasterRCNN](./examples/vision/detection/paddledetection) | [Python](./examples/vision/detection/paddledetection/python)/[C++](./examples/vision/detection/paddledetection/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ❌ | ❌ | ❌ |❔|❔| | Detection | [Megvii-BaseDetection/YOLOX](./examples/vision/detection/yolox) | [Python](./examples/vision/detection/yolox/python)/[C++](./examples/vision/detection/yolox/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Detection | [WongKinYiu/YOLOv7](./examples/vision/detection/yolov7) | [Python](./examples/vision/detection/yolov7/python)/[C++](./examples/vision/detection/yolov7/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Detection | [meituan/YOLOv6](./examples/vision/detection/yolov6) | [Python](./examples/vision/detection/yolov6/python)/[C++](./examples/vision/detection/yolov6/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Detection | [ultralytics/YOLOv5](./examples/vision/detection/yolov5) | [Python](./examples/vision/detection/yolov5/python)/[C++](./examples/vision/detection/yolov5/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Detection | [WongKinYiu/YOLOR](./examples/vision/detection/yolor) | [Python](./examples/vision/detection/yolor/python)/[C++](./examples/vision/detection/yolor/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Detection | [WongKinYiu/ScaledYOLOv4](./examples/vision/detection/scaledyolov4) | [Python](./examples/vision/detection/scaledyolov4/python)/[C++](./examples/vision/detection/scaledyolov4/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Detection | [ppogg/YOLOv5Lite](./examples/vision/detection/yolov5lite) | [Python](./examples/vision/detection/yolov5lite/python)/[C++](./examples/vision/detection/yolov5lite/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Detection | [RangiLyu/NanoDetPlus](./examples/vision/detection/nanodet_plus) | [Python](./examples/vision/detection/nanodet_plus/python)/[C++](./examples/vision/detection/nanodet_plus/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅|❔|❔| | OCR | [PaddleOCR/PP-OCRv2](./examples/vision/ocr) | [Python](./examples/vision/detection/nanodet_plus/python)/[C++](./examples/vision/ocr/PP-OCRv3/cpp)| ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅|❔|❔| | OCR | [PaddleOCR/PP-OCRv3](./examples/vision/ocr) | [Python](./examples/vision/ocr/PP-OCRv3/python)/[C++](./examples/vision/ocr/PP-OCRv3/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅|✅|✅| | Segmentation | [PaddleSeg/PP-LiteSeg](./examples/vision/segmentation/paddleseg) | [Python](./examples/vision/segmentation/paddleseg/python)/[C++](./examples/vision/segmentation/paddleseg/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Segmentation | [PaddleSeg/PP-HumanSegLite](./examples/vision/segmentation/paddleseg) | [Python](./examples/vision/segmentation/paddleseg/python)/[C++](./examples/vision/segmentation/paddleseg/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Segmentation | [PaddleSeg/HRNet](./examples/vision/segmentation/paddleseg) | [Python](./examples/vision/segmentation/paddleseg/python)/[C++](./examples/vision/segmentation/paddleseg/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Segmentation | [PaddleSeg/PP-HumanSegServer](./examples/vision/segmentation/paddleseg) | [Python](./examples/vision/segmentation/paddleseg/python)/[C++](./examples/vision/segmentation/paddleseg/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |✅|✅| | Segmentation | [PaddleSeg/Unet](./examples/vision/segmentation/paddleseg) | [Python](./examples/vision/segmentation/paddleseg/python)/[C++](./examples/vision/segmentation/paddleseg/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Segmentation | [PaddleSeg/Deeplabv3](./examples/vision/segmentation/paddleseg) | [Python](./examples/vision/segmentation/paddleseg/python)/[C++](./examples/vision/segmentation/paddleseg/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Face Detection | [biubug6/RetinaFace](./examples/vision/facedet/retinaface) | [Python](./examples/vision/facedet/retinaface/python)/[C++](./examples/vision/facedet/retinaface/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Face Detection | [Linzaer/UltraFace](./examples/vision/facedet/ultraface) | [ Python](./examples/vision/facedet/ultraface/python)/[C++](./examples/vision/facedet/ultraface/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | FaceDetection | [deepcam-cn/YOLOv5Face](./examples/vision/facedet/yolov5face) | [Python](./examples/vision/facedet/yolov5face/python)/[C++](./examples/vision/facedet/yolov5face/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Face Detection | [insightface/SCRFD](./examples/vision/facedet/scrfd) | [Python](./examples/vision/facedet/scrfd/python)/[C++](./examples/vision/facedet/scrfd/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Face Recognition | [insightface/ArcFace](./examples/vision/faceid/insightface) | [Python](./examples/vision/faceid/insightface/python)/[C++](./examples/vision/faceid/insightface/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Face Recognition | [insightface/CosFace](./examples/vision/faceid/insightface) | [Python](./examples/vision/faceid/insightface/python)/[C++](./examples/vision/faceid/insightface/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Face Recognition | [insightface/PartialFC](./examples/vision/faceid/insightface) | [Python](./examples/vision/faceid/insightface/python)/[C++](./examples/vision/faceid/insightface/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Face Recognition | [insightface/VPL](./examples/vision/faceid/insightface) | [Python](./examples/vision/faceid/insightface/python)/[C++](./examples/vision/faceid/insightface/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |❔|❔| | Matting | [ZHKKKe/MODNet](./examples/vision/matting/modnet) | [Python](./examples/vision/matting/modnet/python)/[C++](./examples/vision/matting/modnet/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅|❔|❔| | Matting | [PaddleSeg/PP-Matting](./examples/vision/matting/ppmatting) | [Python](./examples/vision/matting/ppmatting/python)/[C++](./examples/vision/matting/ppmatting/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅|❔|❔| | Matting | [PaddleSeg/PP-HumanMatting](./examples/vision/matting/modnet) | [Python](./examples/vision/matting/ppmatting/python)/[C++](./examples/vision/matting/ppmatting/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅|❔|❔| | Matting | [PaddleSeg/ModNet](./examples/vision/matting/modnet) | [Python](./examples/vision/matting/ppmatting/python)/[C++](./examples/vision/matting/ppmatting/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅|❔|❔| | Information Extraction | [PaddleNLP/UIE](./examples/text/uie) | [Python](./examples/text/uie/python)/[C++](./examples/text/uie/cpp) | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ | ✅|❔|❔| ## 端侧部署
### EasyEdge边缘端部署
- ARM Linux 系统 - [C++ Inference部署（含视频流）](./docs/arm_cpu/arm_linux_cpp_sdk_inference.md) - [C++ 服务化部署](./docs/arm_cpu/arm_linux_cpp_sdk_serving.md) - [Python Inference部署](./docs/arm_cpu/arm_linux_python_sdk_inference.md) - [Python 服务化部署](./docs/arm_cpu/arm_linux_python_sdk_serving.md) ### EasyEdge移动端部署
- [iOS 系统部署](./docs/arm_cpu/ios_sdk.md) - [Android 系统部署](./docs/arm_cpu/android_sdk.md) ### EasyEdge自定义模型部署
- [快速实现个性化模型替换](./docs/arm_cpu/replace_model_with_another_one.md) ### Paddle Lite NPU部署
- [瑞芯微-NPU/晶晨-NPU/恩智浦-NPU](https://github.com/PaddlePaddle/Paddle-Lite-Demo/tree/develop/object_detection/linux/picodet_detection) ### 端侧模型支持列表
| 任务场景 | 模型 | 大小(MB) | Linux | Android | iOS |Linux | Linux | Linux |更新中...| |:------------------:|:----------------------------:|:---------------------:|:---------------------:|:----------------------:|:---------------------:| :------------------:|:----------------------------:|:---------------------:|:---------------------:| | --- | --- | --- | ARM CPU | ARM CPU | ARM CPU |瑞芯微NPU
RV1109
RV1126
RK1808 | 晶晨NPU
A311D
S905D
C308X | 恩智浦NPU
i.MX 8M Plus |更新中...｜ | Classification | PP-LCNet | 11.9 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Classification | PP-LCNetv2 | 26.6 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Classification | EfficientNet | 31.4 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Classification | GhostNet | 20.8 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Classification | MobileNetV1 | 17 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Classification | MobileNetV2 | 14.2 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Classification | MobileNetV3 | 22 | ✅ | ✅ | ✅ |❔ | ❔ | ❔ |❔| | Classification | ShuffleNetV2 | 9.2 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Classification | SqueezeNetV1.1 | 5 | ✅ | ✅ | ✅ | | Classification | Inceptionv3 | 95.5 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Classification | PP-HGNet | 59 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Classification | SwinTransformer_224_win7 | 352.7 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Detection | PP-PicoDet_s_320_coco | 4.1 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Detection | PP-PicoDet_s_320_lcnet | 4.9 | ✅ | ✅ | ✅ |✅ | ✅ | ✅ | ❔| | Detection | CenterNet | 4.8 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Detection | YOLOv3_MobileNetV3 | 94.6 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Detection | PP-YOLO_tiny_650e_coco | 4.4 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Detection | SSD_MobileNetV1_300_120e_voc | 23.3 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Detection | PP-YOLO_ResNet50vd | 188.5 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Detection | PP-YOLOv2_ResNet50vd | 218.7 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Detection | PP-YOLO_crn_l_300e_coco | 209.1 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Detection | YOLOv5s | 29.3 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Face Detection | BlazeFace | 1.5 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Face Detection | RetinaFace | 1.7 | ✅ | ❌ | ❌ |-- | -- | -- |--| | Keypoint Detection | PP-TinyPose | 5.5 | ✅ | ✅ | ✅ |❔ | ❔ | ❔ |❔| | Segmentation | PP-LiteSeg(STDC1) | 32.2 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Segmentation | PP-HumanSeg-Lite | 0.556 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Segmentation | HRNet-w18 | 38.7 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Segmentation | PP-HumanSeg-Server | 107.2 | ✅ | ✅ | ✅ |-- | -- | -- |--| | Segmentation | Unet | 53.7 | ❌ | ✅ | ❌ |-- | -- | -- |--| | OCR | PP-OCRv1 | 2.3+4.4 | ✅ | ✅ | ✅ |-- | -- | -- |--| | OCR | PP-OCRv2 | 2.3+4.4 | ✅ | ✅ | ✅ |-- | -- | -- |--| | OCR | PP-OCRv3 | 2.4+10.6 | ✅ | ✅ | ✅ |❔ | ❔ | ❔ |❔| | OCR | PP-OCRv3-tiny | 2.4+10.7 | ✅ | ✅ | ✅ |-- | -- | -- |--| ## 社区交流
- **加入社区👬：** 微信扫描二维码，进入**FastDeploy技术交流群**
## Acknowledge
本项目中SDK生成和下载使用了[EasyEdge](https://ai.baidu.com/easyedge/app/openSource)中的免费开放能力，在此表示感谢。 ## License
FastDeploy遵循[Apache-2.0开源协议](./LICENSE)。