Install yolov3 (Anaconda)

Environmental preparation

1. anaconda（ But there is no ）
2. yolov3-master

win10 install yolov3 package

1. First of all to enter yolov3 package , Input cmd Enter command line mode
2. adopt anaconda Create a new environment

conda create -n yolov3 python=3.7

3. Installation required environment

pip install -r requirements.txt -i https://pypi.mirrors.ustc.edu.cn/simple/

because pip Download things from abroad , So add the domestic image source behind

Environmental requirements ：

• python >= 3.7
• pytorch >= 1.1
• numpy
• tqdm
• opencv-python
  Among them, only attention is needed pytorch Installation ：
  To https://pytorch.org/ According to the operating system ,python edition ,cuda Select the command such as version .

without vs Your compiler may not be installed successfully , So download vs tool
vs tool install ：
link ：https://pan.baidu.com/s/1UPwgRKHWr-uW5Jc56Ucl7Q
Extraction code ：khw2

Download it vs tool Then repeat step 2, You can install it yolov3 The environment of

colab install yolov3

1. land google Cloud disk Go to the link : Google cloud disk .

2. I'll have a good yolov3 Packages uploaded to Google cloud disk
   

3. Open after uploading colab.colab

4. Connect colab And Google cloud
   

5. Click on the 2 Then a piece of code will pop up in the middle , Then click on the left run Button to run
   

6. Then a link and a box to fill in the password will pop up , First, click the link to authorize the permission of Google cloud disk , Then there will be a password string for us to copy , Back to colab Paste the password string in the box , press return .

After the link is successful, it can be officially in colab install yolov3 Environment .

7. First, change the operation type to GPU( Very important , Change it every time you restart . This is a colab One of the reasons why it runs faster than our computer )
   
8. New code snippet , Enter the following code and run

!cd /content/drive/MyDrive/yolov3-master && pip install -r requirements.txt

The previous path is written according to personal conditions

then colab The environment in Google cloud has been configured , But every time you restart colab All should be repeated 5 Step on . Then we can colab Train your training set .

9. To solve Colab The problem of dropping the line without operation for a long time
   You will find out , When we are training our models , If you don't operate for a long time Colab Then he will automatically disconnect , This is a headache , We can't just stare at it and run . Mo panic , You can solve it through the following simple operations ：
   First, in the Colab The interface opens the console （ Shortcut key Ctrl+Shift+I）, Then copy the following code （ The function of this code is to set every 60000ms That is to say 1min Automatic click Colab Of “ Connect ” operation , In this way, the connection will not be automatically disconnected due to long-time misoperation , Of course, this is just an example , There's not only one way , You can also adjust your interval ）：

function ClickConnect(){
    
    console.log("Working");
    document.querySelector("colab-toolbar-button#connect").click()
}
setInterval(ClickConnect, 60000)