Catalog
download and install the nvidia graphics driver
download CUDA
download CUDnn
install CUDA
installation (recommended default installation)
set the environment variable After installing
the test environment cuda
install cuDNN
installation Anaconda
installation pytorch
yolov5 cloning project and install
on labelimg markup tools:
**
Download and install nvidia graphics driver
**After
installation, cmd to run nvidia-smi,
if it does not work, add C:\Program Files\NVIDIA Corporation\NVSMI to the path of the environment variable. Then reopen the cmd window.
Download CUDA
https://developer.nvidia.com/cuda-downloads?target_os=Windows&target_arch=x86_64&target_version=10&target_type=exelocal
download cuda_10.2.89_441.22_win10.exe
(If your download is only tens of B in size, it may be a problem with the download. Try again after changing the network.)
Download CUDnn
https://developer.nvidia.com/cudnn
download and get cudnn-10.2-windows10-x64-v7.6.5.32.zip
(If your download is only tens of B in size, it may be a problem with the download. Try again after changing the network.)
Install CUDA
Installation (default installation is recommended)
Set environment variables after installation After
installation, you can see that there are two more environment variables, CUDA_PATH and CUDA_PATH_V10_2, in the system.
At this time we need to add some environment variables:
CUDA_SDK_PATH = C:\ProgramData\NVIDIA Corporation\CUDA Samples\v10.2
CUDA_LIB_PATH=%CUDA_PATH%\lib\x64
CUDA_BIN_PATH = %CUDA_PATH%\bin
CUDA_SDK_BIN_PATH=%CUDA_SDK_PATH%\bin\ win64
CUDA_SDK_LIB_PATH = %CUDA_SDK_PATH%\common\lib\x64
C:\ProgramData\NVIDIACorporation\CUDA Samples\v10.2 is the default installation location, if it is not the default installation, you need to make corresponding modifications
Add at the end of the system variable Path:
%CUDA_LIB_PATH%;%CUDA_BIN_PATH%;%CUDA_SDK_LIB_PATH%;%CUDA_SDK_BIN_PATH%;
Add the following 5 more items (default installation path):
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\lib\x64
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\include
C :\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2\extras\CUPTI\lib64
C:\ProgramData\NVIDIA Corporation\CUDA\Samples\v10.2\bin\win64
C:\ProgramData\NVIDIA Corporation\CUDA\ Samples\v10.2\common\lib\x64
Test the cuda environment
Open CMD to execute:
nvcc -V
Install cuDNN
Copy the cudnn file.
For cudnn, unzip the compressed package directly, and then copy and paste the files in bin, include, and lib to the cuda folder
C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.2
Install Anaconda
- Download the installation package
Anaconda Download the Windows version: https://www.anaconda.com/products/individual - Then install anaconda
- Add Aanaconda domestic mirror configuration
Tsinghua TUNA provides a mirror of the Anaconda warehouse, run the following command
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/
conda config --set show_channel_urls yes
Install pytorch
Note: You need to install pytorch 1.6 or above to create a virtual environment. The environment name can be determined by yourself. Here I use pytorch1.7 as the environment name:
conda create -n pytorch1.7 python=3.8 After
creating it, activate the pytorch1.7 environment:
conda activate The pytorch1.7
installation command is on the official website
Copy it to get the commands needed for installation
yolov5 project clone and install
Install Git software (https://git-scm.com/downloads)
clone the project to the local (such as d:)
git clone https://github.com/ultralytics /yolov5.git
Install required libraries
Use Tsinghua mirroring source:
execution path at yolov5
pip install -i https://pypi.tuna.tsinghua.edu.cn/simple -r requirements.txt
Note: simple and no less, is not http https
About labelimg labeling tool:
https://github.com/tzutalin/labelImg
Just operate on it