Keras yolov3环境搭建

客官°小女子只卖身不卖艺 2023-03-04 09:05 59阅读 0赞

## 训练教程在我这篇博文：[Yolov3训练教程Keras][Yolov3_Keras] ##

**目录**

1、下载并安装Anaconda

2、给Anaconda换源

3、创建yolov3虚拟环境

4、切换环境

5、安装keras

6、安装tensorflow

7、安装Pillow

8、安装matplotlib

9、安装 ffmpeg

10、安装opencv

11、下载yolov3工程及权重文件

12、安装pycharm

13、pycharm配置yolov3工程

14、运行Yolov3

--------------------

# 1、下载并安装环境及工程文件 #

所需IDE：Anaconda3-5.2.0-Windows-x86\_64\_2.exe，也就是Python3.6

所需文件：yolov3-keras工程、权重文件

需要的朋友可以扫码关注公众号，输入yolov3获取所有文件。

![20201003141543451.png][]

# 2、给Anaconda换源 #

用记事本打开本地文件C:\\Users\\用户名\\.condarc把里面的配置信息，新建一个txt文档拷贝进去备份

![20200802230905324.png][]

![20200802230833397.png][]

把.condarc里面的内容替换为，保存退出。

channels:
      - defaults
    show_channel_urls: true
    channel_alias: https://mirrors.tuna.tsinghua.edu.cn/anaconda
    default_channels:
      - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main
      - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free
      - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/r
      - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/pro
      - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/msys2
    custom_channels:
      conda-forge: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
      msys2: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
      bioconda: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
      menpo: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
      pytorch: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
      simpleitk: https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud
    ssl_verify: true

# 3、创建yolov3虚拟环境 #

在开始菜单选择 Anaconda Prompt，右键以管理员身份运行

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70][]

输入：

conda create -n forYolov3 pip python=3.6

![2020080223001722.png][]

回车后会出现：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 1][]

输入y回车即可，等待环境创建完毕。

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 2][]

出现这个表示环境创建完毕了。

# 4、切换环境 #

接着上步输入：conda activate forYolov3  并回车

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 3][]

出现下面这个表示切换成功

![20200802231352358.png][]

# 5、安装keras #

输入命令：

pip install -i https://pypi.tuna.tsinghua.edu.cn/simple keras==2.1.5

并回车，稍微等待即可，速度很快，出现红色箭头所示表示安装成功。

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 4][]

# 6、安装tensorflow #

输入命令：

pip install -i https://pypi.tuna.tsinghua.edu.cn/simple tensorflow==1.12.0

回车等待…………

如下所示，安装成功

![20200803003404363.png][]

# 7、安装Pillow #

输入：

pip install -i https://pypi.tuna.tsinghua.edu.cn/simple Pillow

回车，等待片刻，出现下图所示安装成功：

![20200802232843760.png][]

# 8、安装matplotlib #

输入，回车，等待片刻：

pip install -i https://pypi.tuna.tsinghua.edu.cn/simple matplotlib

下图所示安装成功：

# ![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 5][] 9、安装 ffmpeg #

输入命令：

pip install -i https://pypi.tuna.tsinghua.edu.cn/simple ffmpeg

安装成功：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 6][]

# 10、安装opencv #

输入：

pip install -i https://pypi.tuna.tsinghua.edu.cn/simple opencv-contrib-python

安装成功：

![20200802233515982.png][]

至此依赖库安装好了

# 11、下载yolov3工程及权重文件 #

yolov3+权重+标注工具，已在公众号整理好，需要的话请在文章开头关注获取即可。

解压出来，把权重文件移动到工程文件下：

# ![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 7][]12、安装pycharm #

（1）下载安装包：

http://www.jetbrains.com/pycharm/download/#section=windows

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 8][]

(2)安装：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 9][]

一路next即可，我默认安装在C盘，可自行选择。

到这里勾选上添加环境变量:

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 10][]

等待安装完成：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 11][]

安装完成：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 12][]

# 13、pycharm配置yolov3工程 #

先用pycharm打开yolov3工程：点击File选择Open，浏览打开本地下载好的yolov3工程，如图：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 13][]

然后单击左上角的File选择Setting，配置python解释器

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 14][]

选择Project Interpreter

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 15][]

选择右上角按钮

![20200807101128546.png][]

单击add![20200807101208928.png][]

选择Anaconda安装目录，选择envs下创建的环境的python.exe

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 16][]

到此配置完毕。

# 14、运行Yolov3 #

（1）把权重文件转换成h5文件

点击左下方的 Terminal (和IDLE一个功能，就是命令行)

输入并回车等待：

python convert.py yolov3.cfg yolov3.weights model_data/yolo.h5

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQ1NTA0MTE5_size_16_color_FFFFFF_t_70][]

等待……

↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓↓

PS：如果提示缺少numpy（没有的话请跳过本段），如图：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 17][]

请在下方输入，回车：

pip install -i https://pypi.tuna.tsinghua.edu.cn/simple numpy

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 18][]

然后再输入，上述权重转换命令。

**↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑↑**

**无numpy报错直接看下面：**

转换进行中……

最后会显示有句提示：Saved Keras model to model\_data/yolo.h5

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 19][]

转换完成，生成h5文件：

![20200807103620810.png][]

（2）运行yolov3

在工程目录下新建一个py程序，名字自拟如testYolov3，粘贴下面的代码：

from yolo3.model import yolo_body
    from keras.layers import Input
    from yolo import YOLO
    from PIL import Image
    import os
    os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
    
    
    yolo = YOLO()
    
    while True:
        img = 'img/0.jpg'
        try:
            image = Image.open(img)
        except:
            print('Open Error!')
            continue
        else:
            rel_image = yolo.detect_image(image)
            rel_image.show()
    yolo.close_session()

再在工程目录下新建img文件夹，并放入一张照片，命名为0.jpg

如下图：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 20][]

运行 testYolov3程序：

![watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 21][]

至此yolov3配置完成

[Yolov3_Keras]: https://blog.csdn.net/weixin_41967600/article/details/115267092
[20201003141543451.png]: /images/20230208/40e02830b1b0404f8777b90d1b0f3ea1.png
[20200802230905324.png]: /images/20230208/5581a37e05ba45f0a9d1f07751061406.png
[20200802230833397.png]: /images/20230208/dc0778320b374ed281a79760e43be39d.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70]: /images/20230208/8ac642c505a5444488d345c02683d50e.png
[2020080223001722.png]: /images/20230208/c831279cd3424a74b4806c96f96bf9f9.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 1]: /images/20230208/28941cd633a844a4b343f04a89f712c0.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 2]: /images/20230208/8075c09585bb46f9b26dcdd50c303f0b.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 3]: /images/20230208/5e326a43bf0a459ab94aa0561097df5f.png
[20200802231352358.png]: /images/20230208/a38233e3a3de4d3eb3732559c1cba453.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 4]: /images/20230208/8933a463d00a4901a4af9652f4e9b258.png
[20200803003404363.png]: /images/20230208/9bc7f0228a3041e499e525bae51f2037.png
[20200802232843760.png]: /images/20230208/6de63581b9114e36b4c5a512f94c241e.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 5]: /images/20230208/3a093ee3dbdc4f37b68017ecfad89e3e.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 6]: /images/20230208/2778819e283d4fc7aa6b7c2f303ecc96.png
[20200802233515982.png]: /images/20230208/e7b47f269d7b4219ab4e6052016375ce.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 7]: /images/20230208/871c9ffce3424f8eaa5fbf89827d9dcb.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 8]: /images/20230208/8b3eb0b77f5e40dca57c0661cd3a739e.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 9]: /images/20230208/92e6e3fe0e6342178f8a8d17d8a95e26.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 10]: /images/20230208/c8ccfee9ed7443efa6d3c9f302841b32.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 11]: /images/20230208/1055bc8fe666408aa0ee99cec77a6139.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 12]: /images/20230208/3bffcb3bd529432bba0e567c97aa5a01.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 13]: /images/20230208/e5996f97d6b741f9803c97b10db4f445.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 14]: /images/20230208/09b3d07ad08248cc97089d29b58fd591.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 15]: /images/20230208/54e15bc6613a4d4993d3854869489dc6.png
[20200807101128546.png]: /images/20230208/6be5886972f447ceab26aab960f98327.png
[20200807101208928.png]: /images/20230208/a49ada50189142e3ad99c450a48c23c2.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 16]: /images/20230208/b547d847a74c4c78956464b8bff680b5.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3FxXzQ1NTA0MTE5_size_16_color_FFFFFF_t_70]: /images/20230208/7496bd578b9147ee97e9d1c12ba7da61.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 17]: /images/20230208/40f539ee0c53446eb8ab7a10ae40239a.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 18]: /images/20230208/17755ec66e8649a4be8e896aff90bc8d.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 19]: /images/20230208/f5e5d035e84d4d7285ce9d1b23765b47.png
[20200807103620810.png]: /images/20230208/51096e21fb124241bd90cbb2d5c55f6a.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 20]: /images/20230208/9df80eb39f774490973accbdd2688ce8.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L3dlaXhpbl80MTk2NzYwMA_size_16_color_FFFFFF_t_70 21]: /images/20230208/23e193eb602f4908abd658731ecdf7e5.png