YOLOv5白皮书-第Y1周：调用官方权重进行检测

前言

YOLO诞生之前，目标检测研究领域主要以Faster-RCNN 为首的two-stage算法为主。two-stage算法通过卷积神经网络提取图片特征，再用RPN网络预测目标位置，最后再通过分类和回归来完成对候选目标位置和类别的判定，检测速度较慢但精度相对较高。而one-stage算法直接通过主干网络计算候选目标的类别和位置信息，其相较于two-stage 算法精度较低，但是网络运行的速度更快，内存占用更小，非常适合便携式设备。

YOLO系列是目前最热门的目标检测算法，那就拿它“开刀”了。YOLO目前已经更新到了YOLOv5,由于YOLOv5太新了，目前TensorFlow2版本还未出来(网上存在一些tf版本的，但是大概率不靠谱，就不踩坑了)，那就用PyTorch吧，反正PyTorch迟早也是要学的嘛。
PyTorch与TensorFlow的区别如下:

TensorFlow：倾向于工程人员，落地比较好。

PyTorch: 比较倾向于研究人员，前沿算法用得比较多。

YOLOv5又分为Yolov5s、Yolov5m、Yolov5I、Yolov5x四个版本，我这里以Yolov5s为例。

YOLOv5是什么？

YOLOv5 模型是 Ultralytics 公司于 2020 年 6 月 9 日公开发布的。它包括骨干网络、颈部和头部和输出.

（1）输入端：Mosaic数据增强、自适应锚框计算、自适应图片缩放

（2）Backbone：Focus结构，CSP结构

（3）Neck：FPN+PAN结构

（4）Prediction：GIOU_Loss

使用步骤

源码下载

可以去github官网下载https://github.com/ultralytics/yolov5

Download zip

环境安装

本机在服务器Linux环境下安装。

cd yolov5-master/yolov5-master

/home/deep learning/Y1/yolov5-master/yolov5-master

ls

benchmarks.py    detect.py   [0m[01;34mmodels[0m/           setup.cfg       val.py
[01;34mclassify[0m/        export.py   README.md         train.py
CONTRIBUTING.md  hubconf.py  requirements.txt  tutorial.ipynb
[01;34mdata[0m/            LICENSE     [01;34msegment[0m/          [01;34mutils[0m/

!pip install -r requirements.txt  # install

运行代码

需要的环境
Python>=3.6.0.
PyTorch> =1.7

以同样的方法打开cmd，输入命令: python path/to/detect.py --source path/to/img.jpg --weights yolov5s.pt --img 640，注意需要将图片路径替换成你的，

例如: python detect.py --source data\images\zidane.jpg --weights yolov5s.pt --img 640。这里我们使用的是官方已经帮我们训练好的模型。

在这个过程中，你中途可能会遇到这两个问题，不用担心，这是网络问题多试几次就好了。

TimeoutError: [winError 10060]由于连接方在一段时间后没有正确答复或连接的主机没有反应，连接尝试失败。

TypeError: unlink( ) got an unexpected keyword argument 'missing_ok ’

如果你还是一直这样，将yolov5s.pt(提取码: bum6)文件放到项目的主目录下就ok了，上面的两个错误都是由于下载这个文件、下载失败缺少这个文件导致的。
命令运行成功后，你将会得到如下的反馈:
啊

import torch

# Model
model = torch.hub.load('ultralytics/yolov5', 'yolov5s')  # or yolov5n - yolov5x6, custom

# Images
img = 'https://ultralytics.com/images/zidane.jpg'  # or file, Path, PIL, OpenCV, numpy, list

# Inference
results = model(img)

# Results
results.print()  # or .show(), .save(), .crop(), .pandas(), etc.

Using cache found in /home/liangjie/.cache/torch/hub/ultralytics_yolov5_master
[31m[1mrequirements:[0m YOLOv5 requirement "requests>=2.23.0" not found, attempting AutoUpdate...
Requirement already satisfied: requests>=2.23.0 in /home/liangjie/anaconda3/envs/newcdo/lib/python3.9/site-packages (2.28.1)
Requirement already satisfied: certifi>=2017.4.17 in /home/liangjie/anaconda3/envs/newcdo/lib/python3.9/site-packages (from requests>=2.23.0) (2022.6.15)
Requirement already satisfied: urllib3<1.27,>=1.21.1 in /home/liangjie/anaconda3/envs/newcdo/lib/python3.9/site-packages (from requests>=2.23.0) (1.24.3)
Requirement already satisfied: charset-normalizer<3,>=2 in /home/liangjie/anaconda3/envs/newcdo/lib/python3.9/site-packages (from requests>=2.23.0) (2.1.0)
Requirement already satisfied: idna<4,>=2.5 in /home/liangjie/anaconda3/envs/newcdo/lib/python3.9/site-packages (from requests>=2.23.0) (2.8)

[31m[1mrequirements:[0m 1 package updated per /home/liangjie/.cache/torch/hub/ultralytics_yolov5_master/requirements.txt
[31m[1mrequirements:[0m ⚠️ [1mRestart runtime or rerun command for updates to take effect[0m

YOLOv5  2022-12-2 Python-3.9.12 torch-1.13.0+cu117 CPU

Downloading https://github.com/ultralytics/yolov5/releases/download/v7.0/yolov5s.pt to yolov5s.pt...
ERROR: 
Re-attempting https://github.com/ultralytics/yolov5/releases/download/v7.0/yolov5s.pt to yolov5s.pt...
######################################################################## 100.0%

Fusing layers... 
YOLOv5s summary: 213 layers, 7225885 parameters, 0 gradients
Adding AutoShape... 
image 1/1: 720x1280 2 persons, 2 ties
Speed: 1066.8ms pre-process, 77.1ms inference, 33.3ms NMS per image at shape (1, 3, 384, 640)

results.show()

视频检测

!python detect.py --source MyVideo_2.mp4

Traceback (most recent call last):
  File "/home/liangjie/test/Modelwhale/deep learning/Y1/yolov5-master/yolov5-master/detect.py", line 43, in 
    from models.common import DetectMultiBackend
  File "/home/liangjie/test/Modelwhale/deep learning/Y1/yolov5-master/yolov5-master/models/common.py", line 20, in 
    import pandas as pd
  File "/home/liangjie/anaconda3/envs/newcdo/lib/python3.9/site-packages/pandas/__init__.py", line 48, in 
    from pandas.core.api import (
  File "/home/liangjie/anaconda3/envs/newcdo/lib/python3.9/site-packages/pandas/core/api.py", line 48, in 
    from pandas.core.groupby import (
  File "/home/liangjie/anaconda3/envs/newcdo/lib/python3.9/site-packages/pandas/core/groupby/__init__.py", line 1, in 
    from pandas.core.groupby.generic import (
  File "/home/liangjie/anaconda3/envs/newcdo/lib/python3.9/site-packages/pandas/core/groupby/generic.py", line 70, in 
    from pandas.core.frame import DataFrame
  File "/home/liangjie/anaconda3/envs/newcdo/lib/python3.9/site-packages/pandas/core/frame.py", line 157, in 
    from pandas.core.generic import NDFrame
  File "/home/liangjie/anaconda3/envs/newcdo/lib/python3.9/site-packages/pandas/core/generic.py", line 152, in 
    from pandas.core.window import (
  File "/home/liangjie/anaconda3/envs/newcdo/lib/python3.9/site-packages/pandas/core/window/__init__.py", line 1, in 
    from pandas.core.window.ewm import (  # noqa:F401
  File "/home/liangjie/anaconda3/envs/newcdo/lib/python3.9/site-packages/pandas/core/window/ewm.py", line 12, in 
    import pandas._libs.window.aggregations as window_aggregations
ImportError: /lib64/libstdc++.so.6: version `GLIBCXX_3.4.29' not found (required by /home/liangjie/anaconda3/envs/newcdo/lib/python3.9/site-packages/pandas/_libs/window/aggregations.cpython-39-x86_64-linux-gnu.so)

这个问题是GLIBCXX_3.4.29’ not found ，确实没有3.4.29,需要更新到3.4.29z在运行