PP-YOLOE: An evolved version of YOLO (2022-03-30T00:00:00.000000Z)

TL;DR

PP-YOLOE, an industrial state-of-the-art object detector with high performance and friendly deployment is presented, using anchor-free paradigm, more powerful backbone and neck equipped with CSPRepResStage, ET-head and dynamic label assignment algorithm TAL.

Abstract

In this report, we present PP-YOLOE, an industrial state-of-the-art object detector with high performance and friendly deployment. We optimize on the basis of the previous PP-YOLOv2, using anchor-free paradigm, more powerful backbone and neck equipped with CSPRepResStage, ET-head and dynamic label assignment algorithm TAL. We provide s/m/l/x models for different practice scenarios. As a result, PP-YOLOE-l achieves 51.4 mAP on COCO test-dev and 78.1 FPS on Tesla V100, yielding a remarkable improvement of (+1.9 AP, +13.35% speed up) and (+1.3 AP, +24.96% speed up), compared to the previous state-of-the-art industrial models PP-YOLOv2 and YOLOX respectively. Further, PP-YOLOE inference speed achieves 149.2 FPS with TensorRT and FP16-precision. We also conduct extensive experiments to verify the effectiveness of our designs. Source code and pre-trained models are available at https://github.com/PaddlePaddle/PaddleDetection.

Authors

Qingqing Dang

3 papers

Baohua Lai

3 papers

Shangliang Xu

1 papers

TL;DR

Abstract

Authors

References35 items

PP-PicoDet: A Better Real-Time Object Detector on Mobile Devices

RMNet: Equivalently Removing Residual Connection from Networks

TreeNet: A lightweight One-Shot Aggregation Convolutional Network

TOOD: Task-aligned One-stage Object Detection

YOLOX: Exceeding YOLO Series in 2021

PP-YOLOv2: A Practical Object Detector

RepVGG: Making VGG-style ConvNets Great Again

Object Detection

Scaled-YOLOv4: Scaling Cross Stage Partial Network

VarifocalNet: An IoU-aware Dense Object Detector

PP-YOLO: An Effective and Efficient Implementation of Object Detector

Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection

YOLOv4: Optimal Speed and Accuracy of Object Detection

SOLOv2: Dynamic and Fast Instance Segmentation

Bridging the Gap Between Anchor-Based and Anchor-Free Detection via Adaptive Training Sample Selection

CSPNet: A New Backbone that can Enhance Learning Capability of CNN

Learning Spatial Fusion for Single-Shot Object Detection

EfficientDet: Scalable and Efficient Object Detection

Objects365: A Large-Scale, High-Quality Dataset for Object Detection

An Energy and GPU-Computation Efficient Backbone Network for Real-Time Object Detection

FCOS: Fully Convolutional One-Stage Object Detection

Bag of Tricks for Image Classification with Convolutional Neural Networks

DropBlock: A regularization method for convolutional networks

YOLOv3: An Incremental Improvement

Squeeze-and-Excitation Networks

Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour

Deformable Convolutional Networks

YOLO9000: Better, Faster, Stronger

Aggregated Residual Transformations for Deep Neural Networks

Densely Connected Convolutional Networks

Deep Residual Learning for Image Recognition

You Only Look Once: Unified, Real-Time Object Detection

Rethinking the Aligned and Misaligned Features in One-stage Object Detection

De-formable convnets v2: More deformable

PaddleDetection, object detection and instance segmentation toolkit based on pad-dlepaddle

Field of Study

Journal Information

Name

Volume

Venue Information

Name

Type

URL

Alternate Names