基于Paddlepaddle复现yolov5，支持PaddleDetection接口

Related tags

Deep LearningPaddleDetection-Yolov5
Overview

PaddleDetection yolov5

https://github.com/Sharpiless/PaddleDetection-Yolov5

简介

PaddleDetection飞桨目标检测开发套件，旨在帮助开发者更快更好地完成检测模型的组建、训练、优化及部署等全开发流程。

PaddleDetection模块化地实现了多种主流目标检测算法，提供了丰富的数据增强策略、网络模块组件（如骨干网络）、损失函数等，并集成了模型压缩和跨平台高性能部署能力。

经过长时间产业实践打磨，PaddleDetection已拥有顺畅、卓越的使用体验，被工业质检、遥感图像检测、无人巡检、新零售、互联网、科研等十多个行业的开发者广泛应用。

Yolov5：

YOLOV4出现之后不久，YOLOv5横空出世。YOLOv5在YOLOv4算法的基础上做了进一步的改进，检测性能得到进一步的提升。虽然YOLOv5算法并没有与YOLOv4算法进行性能比较与分析，但是YOLOv5在COCO数据集上面的测试效果还是挺不错的。大家对YOLOv5算法的创新性半信半疑，有的人对其持肯定态度，有的人对其持否定态度。在我看来，YOLOv5检测算法中还是存在很多可以学习的地方，虽然这些改进思路看来比较简单或者创新点不足，但是它们确定可以提升检测算法的性能。其实工业界往往更喜欢使用这些方法，而不是利用一个超级复杂的算法来获得较高的检测精度。本文将对YOLOv5检测算法进行复现。

下载预训练模型：

https://drive.google.com/file/d/16tREOOJzKgOLw31bSiUNz0iBdqoRzq1i/view?usp=sharing

训练Yolov5：

python tools/train.py -c configs/yolov5/yolov5s_CSPdarknet_roadsign.yml

实验结果：

0.9087 mAP on roadsign dataset.

01

01

关注我的公众号：

感兴趣的同学关注我的公众号——可达鸭的深度学习教程：

在这里插入图片描述

联系作者：

B站：https://space.bilibili.com/470550823

CSDN：https://blog.csdn.net/weixin_44936889

AI Studio：https://aistudio.baidu.com/aistudio/personalcenter/thirdview/67156

Github：https://github.com/Sharpiless

%cd work/
/home/aistudio/work

!unzip PPDet-yolov5v2.zip -d ./
!python tools/train.py -c configs/yolov5/yolov5s_CSPdarknet_roadsign.yml --eval
/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/tensor/creation.py:125: DeprecationWarning: `np.object` is a deprecated alias for the builtin `object`. To silence this warning, use `object` by itself. Doing this will not modify any behavior and is safe. 
Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations
  if data.dtype == np.object:
[07/15 10:17:41] ppdet.utils.download WARNING: Config annotation dataset/roadsign_voc/train.txt is not a file, dataset config is not valid
[07/15 10:17:41] ppdet.utils.download INFO: Dataset /home/aistudio/work/dataset/roadsign_voc is not valid for reason above, try searching /home/aistudio/.cache/paddle/dataset or downloading dataset...
[07/15 10:17:41] ppdet.utils.download INFO: Found /home/aistudio/.cache/paddle/dataset/roadsign_voc/annotations
[07/15 10:17:41] ppdet.utils.download INFO: Found /home/aistudio/.cache/paddle/dataset/roadsign_voc/images
[07/15 10:17:41] reader WARNING: Shared memory size is less than 1G, disable shared_memory in DataLoader
[07/15 10:17:42] ppdet.utils.checkpoint INFO: Finish loading model weights: output.pdparams
[07/15 10:17:51] ppdet.engine INFO: Epoch: [0] [ 0/87] learning_rate: 0.000033 loss_xy: 0.752040 loss_wh: 0.698217 loss_iou: 2.634957 loss_obj: 11.301561 loss_cls: 1.041652 loss: 16.428429 eta: 8:28:32 batch_cost: 8.7679 data_cost: 0.9061 ips: 0.9124 images/s
[07/15 10:19:42] ppdet.engine INFO: Epoch: [0] [20/87] learning_rate: 0.000047 loss_xy: 0.529626 loss_wh: 0.569290 loss_iou: 2.436198 loss_obj: 8.576855 loss_cls: 1.023474 loss: 13.317031 eta: 5:29:28 batch_cost: 5.5608 data_cost: 0.0002 ips: 1.4386 images/s
[07/15 10:21:42] ppdet.engine INFO: Epoch: [0] [40/87] learning_rate: 0.000060 loss_xy: 0.500230 loss_wh: 0.502719 loss_iou: 2.226187 loss_obj: 4.208471 loss_cls: 0.890207 loss: 8.235611 eta: 5:35:40 batch_cost: 6.0032 data_cost: 0.0003 ips: 1.3326 images/s
[07/15 10:23:23] ppdet.engine INFO: Epoch: [0] [60/87] learning_rate: 0.000073 loss_xy: 0.519860 loss_wh: 0.599364 loss_iou: 2.455585 loss_obj: 3.626266 loss_cls: 1.031202 loss: 8.345335 eta: 5:18:38 batch_cost: 5.0474 data_cost: 0.0003 ips: 1.5850 images/s
[07/15 10:25:13] ppdet.engine INFO: Epoch: [0] [80/87] learning_rate: 0.000087 loss_xy: 0.568008 loss_wh: 0.618775 loss_iou: 2.583227 loss_obj: 3.632595 loss_cls: 0.863238 loss: 7.575019 eta: 5:15:29 batch_cost: 5.4984 data_cost: 0.0002 ips: 1.4550 images/s
[07/15 10:25:47] ppdet.utils.checkpoint INFO: Save checkpoint: output/yolov5s_CSPdarknet_roadsign
[07/15 10:25:47] ppdet.utils.download WARNING: Config annotation dataset/roadsign_voc/valid.txt is not a file, dataset config is not valid
[07/15 10:25:47] ppdet.utils.download INFO: Dataset /home/aistudio/work/dataset/roadsign_voc is not valid for reason above, try searching /home/aistudio/.cache/paddle/dataset or downloading dataset...
[07/15 10:25:47] ppdet.utils.download INFO: Found /home/aistudio/.cache/paddle/dataset/roadsign_voc/annotations
[07/15 10:25:47] ppdet.utils.download INFO: Found /home/aistudio/.cache/paddle/dataset/roadsign_voc/images
[07/15 10:25:48] ppdet.engine INFO: Eval iter: 0
[07/15 10:26:09] ppdet.engine INFO: Eval iter: 100
[07/15 10:26:25] ppdet.metrics.metrics INFO: Accumulating evaluatation results...
[07/15 10:26:25] ppdet.metrics.metrics INFO: mAP(0.50, integral) = 85.84%
[07/15 10:26:25] ppdet.engine INFO: Total sample number: 176, averge FPS: 4.751870228058035
[07/15 10:26:25] ppdet.engine INFO: Best test bbox ap is 0.858.
[07/15 10:26:25] ppdet.utils.checkpoint INFO: Save checkpoint: output/yolov5s_CSPdarknet_roadsign
[07/15 10:26:35] ppdet.engine INFO: Epoch: [1] [ 0/87] learning_rate: 0.000091 loss_xy: 0.567437 loss_wh: 0.623783 loss_iou: 2.511684 loss_obj: 3.314124 loss_cls: 0.949793 loss: 7.338743 eta: 5:16:15 batch_cost: 6.2481 data_cost: 0.0003 ips: 1.2804 images/s
[07/15 10:28:39] ppdet.engine INFO: Epoch: [1] [20/87] learning_rate: 0.000100 loss_xy: 0.583728 loss_wh: 0.708465 loss_iou: 2.704193 loss_obj: 3.461134 loss_cls: 1.127932 loss: 9.057523 eta: 5:20:59 batch_cost: 6.2270 data_cost: 0.0003 ips: 1.2847 images/s
[07/15 10:30:28] ppdet.engine INFO: Epoch: [1] [40/87] learning_rate: 0.000100 loss_xy: 0.576615 loss_wh: 0.655194 loss_iou: 2.566234 loss_obj: 2.921384 loss_cls: 1.010778 loss: 7.844104 eta: 5:16:43 batch_cost: 5.4392 data_cost: 0.0003 ips: 1.4708 images/s
[07/15 10:32:34] ppdet.engine INFO: Epoch: [1] [60/87] learning_rate: 0.000100 loss_xy: 0.583071 loss_wh: 0.726098 loss_iou: 2.730413 loss_obj: 3.053501 loss_cls: 0.991524 loss: 8.496977 eta: 5:19:40 batch_cost: 6.3128 data_cost: 0.0003 ips: 1.2673 images/s
[07/15 10:34:31] ppdet.engine INFO: Epoch: [1] [80/87] learning_rate: 0.000100 loss_xy: 0.606061 loss_wh: 0.652358 loss_iou: 2.841094 loss_obj: 3.237591 loss_cls: 1.084277 loss: 8.605825 eta: 5:18:16 batch_cost: 5.8318 data_cost: 0.0003 ips: 1.3718 images/s
[07/15 10:34:59] ppdet.utils.checkpoint INFO: Save checkpoint: output/yolov5s_CSPdarknet_roadsign
[07/15 10:35:00] ppdet.engine INFO: Eval iter: 0
[07/15 10:35:19] ppdet.engine INFO: Eval iter: 100
[07/15 10:35:33] ppdet.metrics.metrics INFO: Accumulating evaluatation results...
[07/15 10:35:33] ppdet.metrics.metrics INFO: mAP(0.50, integral) = 85.30%
[07/15 10:35:33] ppdet.engine INFO: Total sample number: 176, averge FPS: 5.151774310709877
[07/15 10:35:33] ppdet.engine INFO: Best test bbox ap is 0.858.
[07/15 10:35:46] ppdet.engine INFO: Epoch: [2] [ 0/87] learning_rate: 0.000100 loss_xy: 0.537015 loss_wh: 0.587401 loss_iou: 2.352699 loss_obj: 3.121367 loss_cls: 1.012583 loss: 7.857001 eta: 5:17:11 batch_cost: 5.8271 data_cost: 0.0003 ips: 1.3729 images/s
^C

!rm -rf output/
!zip -r code.zip ./*
Owner
BIT可达鸭
GitHub Repository
Codebase for the solution that won first place and was awarded the most human-like agent in the 2021 NeurIPS Competition MineRL BASALT Challenge.

KAIROS MineRL BASALT Codebase for the solution that won first place and was awarded the most human-like agent in the 2021 NeurIPS Competition MineRL B

Vinicius G. Goecks 37 Oct 30, 2022
Empower Sequence Labeling with Task-Aware Language Model

LM-LSTM-CRF Check Our New NER Toolkit 🚀 🚀 🚀 Inference: LightNER: inference w. models pre-trained / trained w. any following tools, efficiently. Tra

Liyuan Liu 838 Jan 05, 2023
Detectron2 is FAIR's next-generation platform for object detection and segmentation.

Detectron2 is Facebook AI Research's next generation software system that implements state-of-the-art object detection algorithms. It is a ground-up r

Facebook Research 23.3k Jan 08, 2023
Code base for the paper "Scalable One-Pass Optimisation of High-Dimensional Weight-Update Hyperparameters by Implicit Differentiation"

This repository contains code for the paper Scalable One-Pass Optimisation of High-Dimensional Weight-Update Hyperparameters by Implicit Differentiati

8 Aug 28, 2022
Implementation for our ICCV 2021 paper: Dual-Camera Super-Resolution with Aligned Attention Modules

DCSR: Dual Camera Super-Resolution Implementation for our ICCV 2021 oral paper: Dual-Camera Super-Resolution with Aligned Attention Modules paper | pr

Tengfei Wang 110 Dec 20, 2022
A PyTorch Implementation of SphereFace.

SphereFace A PyTorch Implementation of SphereFace. The code can be trained on CASIA-Webface and the best accuracy on LFW is 99.22%. SphereFace: Deep H

carwin 685 Dec 09, 2022
Code for "Learning Canonical Representations for Scene Graph to Image Generation", Herzig & Bar et al., ECCV2020

Learning Canonical Representations for Scene Graph to Image Generation (ECCV 2020) Roei Herzig*, Amir Bar*, Huijuan Xu, Gal Chechik, Trevor Darrell, A

roei_herzig 24 Jul 07, 2022
A repository that finds a person who looks like you by using face recognition technology.

Find Your Twin Hello everyone, I've always wondered how casting agencies do the casting for a scene where a certain actor is young or old for a movie

Cengizhan Yurdakul 3 Jan 29, 2022
FCOS: Fully Convolutional One-Stage Object Detection (ICCV'19)

FCOS: Fully Convolutional One-Stage Object Detection This project hosts the code for implementing the FCOS algorithm for object detection, as presente

Tian Zhi 3.1k Jan 05, 2023
The mini-MusicNet dataset

mini-MusicNet A music-domain dataset for multi-label classification Music transcription is sequence-to-sequence prediction problem: given an audio per

John Thickstun 4 Nov 09, 2022
A community run, 5-day PyTorch Deep Learning Bootcamp

Deep Learning Winter School, November 2107. Tel Aviv Deep Learning Bootcamp : http://deep-ml.com. About Tel-Aviv Deep Learning Bootcamp is an intensiv

Shlomo Kashani. 1.3k Sep 04, 2021
ATOMIC 2020: On Symbolic and Neural Commonsense Knowledge Graphs

(Comet-) ATOMIC 2020: On Symbolic and Neural Commonsense Knowledge Graphs Paper Jena D. Hwang, Chandra Bhagavatula, Ronan Le Bras, Jeff Da, Keisuke Sa

AI2 152 Dec 27, 2022
PyTorch implementation of paper: HPNet: Deep Primitive Segmentation Using Hybrid Representations.

HPNet This repository contains the PyTorch implementation of paper: HPNet: Deep Primitive Segmentation Using Hybrid Representations. Installation The

Siming Yan 42 Dec 07, 2022
An index of recommendation algorithms that are based on Graph Neural Networks.

An index of recommendation algorithms that are based on Graph Neural Networks.

FIB LAB, Tsinghua University 564 Jan 07, 2023
Hashformers is a framework for hashtag segmentation with transformers.

Hashtag segmentation is the task of automatically inserting the missing spaces between the words in a hashtag. Hashformers applies Transformer models

Ruan Chaves 41 Nov 09, 2022
Semantic Scholar's Author Disambiguation Algorithm & Evaluation Suite

S2AND This repository provides access to the S2AND dataset and S2AND reference model described in the paper S2AND: A Benchmark and Evaluation System f

AI2 54 Nov 28, 2022
Conformer: Local Features Coupling Global Representations for Visual Recognition

Conformer: Local Features Coupling Global Representations for Visual Recognition (arxiv) This repository is built upon DeiT and timm Usage First, inst

Zhiliang Peng 378 Jan 08, 2023
Official implementation of the Neurips 2021 paper Searching Parameterized AP Loss for Object Detection.

Parameterized AP Loss By Chenxin Tao, Zizhang Li, Xizhou Zhu, Gao Huang, Yong Liu, Jifeng Dai This is the official implementation of the Neurips 2021

46 Jul 06, 2022
This repository contains code, network definitions and pre-trained models for working on remote sensing images using deep learning

Deep learning for Earth Observation This repository contains code, network definitions and pre-trained models for working on remote sensing images usi

Nicolas Audebert 447 Jan 05, 2023
For the paper entitled ''A Case Study and Qualitative Analysis of Simple Cross-Lingual Opinion Mining''

Summary This is the source code for the paper "A Case Study and Qualitative Analysis of Simple Cross-Lingual Opinion Mining", which was accepted as fu

1 Nov 10, 2021