Steps to build a deep learning environment GPU

1 understand GPU

gpu Parallel through a large number of threads , Make it faster to deal with deep learning than cpu Much faster , Specifically gpu Concept please click gpu.

2 View graphics card

Check whether your graphics card supports cuda and cudnn Accelerate deep learning .Intel The beginning is the integrated graphics card ,AMD、NVIDIA The beginning is an independent graphics card .

Yes NVIDIA（ Ying Wei Da ） A separate graphics card .

3 install CUDA and cuDNN

1. Check the graphics card support CUDA Highest version
2. Download and install graphics card support CUDA
3. Configure environment variables
4. Download and CUDA Corresponding cuDNN
5. install cuDNN（ Plug in installation ）
6. verification CUDA And cuDNN install

4 install anconda

1. Download and install anconda
2. Switch image source
3. establish pytorch A virtual environment
4. Get into pytorch Official website , Download and install the corresponding version pytorch

5 install pycharm

Import pytorch Environmental Science