paddle-inference-deploy-Lib

🔥 Best Deployment Tools for paddleDetection and paddleSeg.

This repository is a universal deployment paradigm of paddleDetection and paddleSeg. All the models export by these two Toolkits were compatible in paddle-inference-deploy-Lib.

Semantic Segmentation

Object Detection

Install Dependence

• PaddlePaddle(>=2.2.0)

Install the PaddlePaddle 2.1 version, please refer to Quick Installation for the specific installation method.

• Paddleseg

pip install paddleseg

Usage

There are two local example models in ./model. You can run it directly.

(1) Run PaddleSeg Deployment

python paddleSegDemo.py 

(2) Run PaddleDet Deployment

python paddleDetDemo.py 

Overview of Kit Compatibale Structures

Object Detection	Semantic Segmentation
Faster RCNN FPN Cascade-RCNN Libra RCNN Hybrid Task RCNN PSS-Det RetinaNet YOLOv3 YOLOv4 PP-YOLOv1/v2 PP-YOLO-Tiny SSD CornerNet-Squeeze FCOS TTFNet PP-PicoDet DETR Deformable DETR Swin Transformer Sparse RCNN	ANN BiSeNetV2 DANet DeepLabV3 DeepLabV3P Fast-SCNN HRNet-FCN GCNet GSCNN HarDNet OCRNet PSPNet U-Net U2-Net Att U-Net U-Net++ U-Net3+ DecoupledSeg EMANet ISANet DNLNet SFNet PP-HumanSeg PortraitNet STDC GINet PointRend SegNet ESPNetV2 HRNet-Contrast DMNet ESPNetV1 ENCNet PFPNNet FastFCN BiSeNetV1 SETR MLA Transformer SegFormer SegMenter

Model Zoo

Detection Model

The relationship between COCO mAP and FPS on Tesla V100 of Detection representative models of each server side architectures and backbones.

Segmentation Model

The relationship between mIoU and FLOPs of representative Segmentation architectures and backbones. See Model Zoo Overview for more details.

Future Work

Star this repository will encourage me to add new features such as quick C++ deployment Lib for the ToolKit.

• C++ PaddleDet Lib
• C++ PaddleSeg Lib