Windows中配置yolov5所需环境

刺骨的言语ヽ痛彻心扉 2023-10-11 15:23 53阅读 0赞

### 配置yolov5所需环境 ###

##### 一、安装anaconda与pycharm #####

首先需要下载anaconda与pycharm，这两个工具均可从官网直接下载。  
anaconda：[https://www.anaconda.com/products/individual\#macos][https_www.anaconda.com_products_individual_macos]  
pycharm：[https://www.jetbrains.com/pycharm/download/\#section=windows][https_www.jetbrains.com_pycharm_download_section_windows]  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70]  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70 1]  
选择适合自己电脑的版本下载即可，网上很多anaconda和pycharm下载安装教程，这里就不详细介绍了。

##### 二、安装CUDA #####

首先在NVIDIA控制面板中查看自己电脑的cuda版本，我的版本是cuda11.0。  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70 2]  
在官网下载对应的版本，所有版本：[https://developer.nvidia.com/cuda-toolkit-archive][https_developer.nvidia.com_cuda-toolkit-archive]  
下载好后进行安装  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70 3]  
安装完成后可在cmd中用nvcc --version检查是否成功安装：  
![在这里插入图片描述][20210310194505818.png]

##### 三、安装pytorch #####

在[pytorch官网][pytorch]查看自己适合的环境，![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70 4]  
之后在anaconda prompt中先创建yolov5虚拟环境，输入：conda create -n yolov5 python=3.8  
![在这里插入图片描述][20210310195803113.png]  
进入环境：conda activate yolov5  
![在这里插入图片描述][20210310195933836.png]  
安装pytorch，输入：conda install pytorch torchvision torchaudio cudatoolkit=11.0  
![在这里插入图片描述][20210310200317844.png]  
若下载过程太慢，可以使用清华源镜像，输入：  
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/  
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/  
![在这里插入图片描述][20210310200538238.png]  
安装完成后检查是否成功安装，输入：

> python  
> import torch  
> torch.cuda.is\_available()

![在这里插入图片描述][20210310200734710.png]

##### 四、配置yolov5所需环境 #####

首先下载yolov5源码，[https://github.com/ultralytics/yolov5][https_github.com_ultralytics_yolov5]  
按照requirement.txt文件逐个进行安装  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70 5]  
依次输入：

> pip install Cython  
> pip install numpy  
> pip install opencv-python  
> pip install matplotlib  
> pip install pillow  
> pip install tensorboard  
> pip install PyYAML  
> pip install torchvision  
> pip install scipy  
> pip install tqdm

![在这里插入图片描述][20210310201858739.png]  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70 6]  
全部安装好之后运行detect.py进行测试  
![在这里插入图片描述][20210310202253389.png]  
可以自己找图片放在data/images进行测试,测试结果如下：  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70 7]  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70 8]  
到此yolov5所需环境就配置成功啦~

[https_www.anaconda.com_products_individual_macos]: https://www.anaconda.com/products/individual#macos
[https_www.jetbrains.com_pycharm_download_section_windows]: https://www.jetbrains.com/pycharm/download/#section=windows
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70]: https://img-blog.csdnimg.cn/20210310173426916.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70 1]: https://img-blog.csdnimg.cn/20210310173744437.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70 2]: https://img-blog.csdnimg.cn/202103101936541.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg==,size_16,color_FFFFFF,t_70
[https_developer.nvidia.com_cuda-toolkit-archive]: https://developer.nvidia.com/cuda-toolkit-archive
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70 3]: https://img-blog.csdnimg.cn/20210310194308830.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg==,size_16,color_FFFFFF,t_70
[20210310194505818.png]: https://img-blog.csdnimg.cn/20210310194505818.png
[pytorch]: https://pytorch.org/
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70 4]: https://img-blog.csdnimg.cn/2021031019525185.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg==,size_16,color_FFFFFF,t_70
[20210310195803113.png]: https://img-blog.csdnimg.cn/20210310195803113.png
[20210310195933836.png]: https://img-blog.csdnimg.cn/20210310195933836.png
[20210310200317844.png]: https://img-blog.csdnimg.cn/20210310200317844.png
[20210310200538238.png]: https://img-blog.csdnimg.cn/20210310200538238.png
[20210310200734710.png]: https://img-blog.csdnimg.cn/20210310200734710.png
[https_github.com_ultralytics_yolov5]: https://github.com/ultralytics/yolov5
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70 5]: https://img-blog.csdnimg.cn/20210310201555482.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg==,size_16,color_FFFFFF,t_70
[20210310201858739.png]: https://img-blog.csdnimg.cn/20210310201858739.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70 6]: https://img-blog.csdnimg.cn/20210310201957432.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg==,size_16,color_FFFFFF,t_70
[20210310202253389.png]: https://img-blog.csdnimg.cn/20210310202253389.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70 7]: https://img-blog.csdnimg.cn/20210310202529607.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg==,size_16,color_FFFFFF,t_70
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg_size_16_color_FFFFFF_t_70 8]: https://img-blog.csdnimg.cn/20210310202600405.png?x-oss-process=image/watermark,type_ZmFuZ3poZW5naGVpdGk,shadow_10,text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L1d4eTk3MTEyMg==,size_16,color_FFFFFF,t_70