I tried to apply the CAM algorithm to YOLOv4 and it worked.

Overview

YOLOV4:You Only Look Once目标检测模型在pytorch当中的实现


2021年2月7日更新:
加入letterbox_image的选项,关闭letterbox_image后网络的map得到大幅度提升。

目录

  1. 性能情况 Performance
  2. 实现的内容 Achievement
  3. 所需环境 Environment
  4. 注意事项 Attention
  5. 小技巧的设置 TricksSet
  6. 文件下载 Download
  7. 预测步骤 How2predict
  8. 训练步骤 How2train
  9. 参考资料 Reference

性能情况

训练数据集 权值文件名称 测试数据集 输入图片大小 mAP 0.5:0.95 mAP 0.5
VOC07+12+COCO yolo4_voc_weights.pth VOC-Test07 416x416 - 89.0
COCO-Train2017 yolo4_weights.pth COCO-Val2017 416x416 46.1 70.2

实现的内容

  • 主干特征提取网络:DarkNet53 => CSPDarkNet53
  • 特征金字塔:SPP,PAN
  • 训练用到的小技巧:Mosaic数据增强、Label Smoothing平滑、CIOU、学习率余弦退火衰减
  • 激活函数:使用Mish激活函数
  • ……balabla

所需环境

torch==1.2.0

注意事项

代码中的yolo4_weights.pth是基于608x608的图片训练的,但是由于显存原因。我将代码中的图片大小修改成了416x416。有需要的可以修改回来。 代码中的默认anchors是基于608x608的图片的。
注意不要使用中文标签,文件夹中不要有空格!
在训练前需要务必在model_data下新建一个txt文档,文档中输入需要分的类,在train.py中将classes_path指向该文件

小技巧的设置

在train.py文件下:
1、mosaic参数可用于控制是否实现Mosaic数据增强。
2、Cosine_scheduler可用于控制是否使用学习率余弦退火衰减。
3、label_smoothing可用于控制是否Label Smoothing平滑。

文件下载

训练所需的yolo4_weights.pth可在百度网盘中下载。
链接: https://pan.baidu.com/s/1WlDNPtGO1pwQbqwKx1gRZA 提取码: p4sc
yolo4_weights.pth是coco数据集的权重。
yolo4_voc_weights.pth是voc数据集的权重。

预测步骤

a、使用预训练权重

  1. 下载完库后解压,在百度网盘下载yolo4_weights.pth或者yolo4_voc_weights.pth,放入model_data,运行predict.py,输入
img/street.jpg
  1. 利用video.py可进行摄像头检测。

b、使用自己训练的权重

  1. 按照训练步骤训练。
  2. 在yolo.py文件里面,在如下部分修改model_path和classes_path使其对应训练好的文件;model_path对应logs文件夹下面的权值文件,classes_path是model_path对应分的类
_defaults = {
    "model_path": 'model_data/yolo4_weights.pth',
    "anchors_path": 'model_data/yolo_anchors.txt',
    "classes_path": 'model_data/coco_classes.txt',
    "model_image_size" : (416, 416, 3),
    "confidence": 0.5,
    "cuda": True
}
  1. 运行predict.py,输入
img/street.jpg
  1. 利用video.py可进行摄像头检测。

训练步骤

  1. 本文使用VOC格式进行训练。
  2. 训练前将标签文件放在VOCdevkit文件夹下的VOC2007文件夹下的Annotation中。
  3. 训练前将图片文件放在VOCdevkit文件夹下的VOC2007文件夹下的JPEGImages中。
  4. 在训练前利用voc2yolo4.py文件生成对应的txt。
  5. 再运行根目录下的voc_annotation.py,运行前需要将classes改成你自己的classes。注意不要使用中文标签,文件夹中不要有空格!
classes = ["aeroplane", "bicycle", "bird", "boat", "bottle", "bus", "car", "cat", "chair", "cow", "diningtable", "dog", "horse", "motorbike", "person", "pottedplant", "sheep", "sofa", "train", "tvmonitor"]
  1. 此时会生成对应的2007_train.txt,每一行对应其图片位置及其真实框的位置
  2. 在训练前需要务必在model_data下新建一个txt文档,文档中输入需要分的类,在train.py中将classes_path指向该文件,示例如下:
classes_path = 'model_data/new_classes.txt'    

model_data/new_classes.txt文件内容为:

cat
dog
...
  1. 运行train.py即可开始训练。

mAP目标检测精度计算更新

更新了get_gt_txt.py、get_dr_txt.py和get_map.py文件。
get_map文件克隆自https://github.com/Cartucho/mAP
具体mAP计算过程可参考:https://www.bilibili.com/video/BV1zE411u7Vw

Reference

https://github.com/qqwweee/keras-yolo3/
https://github.com/Cartucho/mAP
https://github.com/Ma-Dan/keras-yolo4

The above is original readme.md

My work

I tried to train the YOLOv4 to detect the helmet and it's color. In order to know whether it learned well, I visualized the output of the YOLO-Head.

origin.jpg detection.jpg
origin.jpg detection.jpg
head0 head1 head2
head0layer0score.jpg head1layer0score.jpg head2layer0score.jpg
head0layer1class.jpg head1layer1class.jpg head2layer1class.jpg
head0layer2class_score.jpg head1layer2class_score.jpg head2layer2class_score.jpg

The above are shown the visualization of the "yellow", we can easily see the hot area focus on the yellow helmets. So I think this can help us to train the model to a certain extent.

Official Pytorch implementation of Online Continual Learning on Class Incremental Blurry Task Configuration with Anytime Inference (ICLR 2022)

The Official Implementation of CLIB (Continual Learning for i-Blurry) Online Continual Learning on Class Incremental Blurry Task Configuration with An

NAVER AI 34 Oct 26, 2022
Eff video representation - Efficient video representation through neural fields

Neural Residual Flow Fields for Efficient Video Representations 1. Download MPI

41 Jan 06, 2023
TensorFlow implementation of "TokenLearner: What Can 8 Learned Tokens Do for Images and Videos?"

TokenLearner: What Can 8 Learned Tokens Do for Images and Videos? Source: Improving Vision Transformer Efficiency and Accuracy by Learning to Tokenize

Aritra Roy Gosthipaty 23 Dec 24, 2022
Dynamic View Synthesis from Dynamic Monocular Video

Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfer This repository contains code to compute depth from a

Intelligent Systems Lab Org 2.3k Jan 01, 2023
ELSED: Enhanced Line SEgment Drawing

ELSED: Enhanced Line SEgment Drawing This repository contains the source code of ELSED: Enhanced Line SEgment Drawing the fastest line segment detecto

Iago Suárez 125 Dec 31, 2022
x-transformers-paddle 2.x version

x-transformers-paddle x-transformers-paddle 2.x version paddle 2.x版本 https://github.com/lucidrains/x-transformers 。 requirements paddlepaddle-gpu==2.2

yujun 7 Dec 08, 2022
[NeurIPS'20] Self-supervised Co-Training for Video Representation Learning. Tengda Han, Weidi Xie, Andrew Zisserman.

CoCLR: Self-supervised Co-Training for Video Representation Learning This repository contains the implementation of: InfoNCE (MoCo on videos) UberNCE

Tengda Han 271 Jan 02, 2023
Seeing All the Angles: Learning Multiview Manipulation Policies for Contact-Rich Tasks from Demonstrations

Seeing All the Angles: Learning Multiview Manipulation Policies for Contact-Rich Tasks from Demonstrations Trevor Ablett, Daniel (Yifan) Zhai, Jonatha

STARS Laboratory 3 Feb 01, 2022
Unofficial PyTorch implementation of MobileViT.

MobileViT Overview This is a PyTorch implementation of MobileViT specified in "MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Tr

Chin-Hsuan Wu 348 Dec 23, 2022
TAP: Text-Aware Pre-training for Text-VQA and Text-Caption, CVPR 2021 (Oral)

TAP: Text-Aware Pre-training TAP: Text-Aware Pre-training for Text-VQA and Text-Caption by Zhengyuan Yang, Yijuan Lu, Jianfeng Wang, Xi Yin, Dinei Flo

Microsoft 61 Nov 14, 2022
Satellite labelling tool for manual labelling of storm top features such as overshooting tops, above-anvil plumes, cold U/Vs, rings etc.

Satellite labelling tool About this app A tool for manual labelling of storm top features such as overshooting tops, above-anvil plumes, cold U/Vs, ri

Czech Hydrometeorological Institute - Satellite Department 10 Sep 14, 2022
Python package for missing-data imputation with deep learning

MIDASpy Overview MIDASpy is a Python package for multiply imputing missing data using deep learning methods. The MIDASpy algorithm offers significant

MIDASverse 77 Dec 03, 2022
A minimal solution to hand motion capture from a single color camera at over 100fps. Easy to use, plug to run.

Minimal Hand A minimal solution to hand motion capture from a single color camera at over 100fps. Easy to use, plug to run. This project provides the

Yuxiao Zhou 824 Jan 07, 2023
Custom implementation of Corrleation Module

Pytorch Correlation module this is a custom C++/Cuda implementation of Correlation module, used e.g. in FlowNetC This tutorial was used as a basis for

Clément Pinard 361 Dec 12, 2022
Official Pytorch implementation of "DivCo: Diverse Conditional Image Synthesis via Contrastive Generative Adversarial Network" (CVPR'21)

DivCo: Diverse Conditional Image Synthesis via Contrastive Generative Adversarial Network Pytorch implementation for our DivCo. We propose a simple ye

64 Nov 22, 2022
Flybirds - BDD-driven natural language automated testing framework, present by Trip Flight

Flybird | English Version 行为驱动开发(Behavior-driven development,缩写BDD),是一种软件过程的思想或者

Ctrip, Inc. 706 Dec 30, 2022
Explaining in Style: Training a GAN to explain a classifier in StyleSpace

Explaining in Style: Official TensorFlow Colab Explaining in Style: Training a GAN to explain a classifier in StyleSpace Oran Lang, Yossi Gandelsman,

Google 197 Nov 08, 2022
The Fundamental Clustering Problems Suite (FCPS) summaries 54 state-of-the-art clustering algorithms, common cluster challenges and estimations of the number of clusters as well as the testing for cluster tendency.

FCPS Fundamental Clustering Problems Suite The package provides over sixty state-of-the-art clustering algorithms for unsupervised machine learning pu

9 Nov 27, 2022
Background-Click Supervision for Temporal Action Localization

Background-Click Supervision for Temporal Action Localization This repository is the official implementation of BackTAL. In this work, we study the te

LeYang 221 Oct 09, 2022
Auto-updating data to assist in investment to NEPSE

Symbol Ratios Summary Sector LTP Undervalued Bonus % MEGA Strong Commercial Banks 368 5 10 JBBL Strong Development Banks 568 5 10 SIFC Strong Finance

Amit Chaudhary 16 Nov 01, 2022