Since AI recognition is a process that consumes a large amount of engineering, it will not be accelerated if CPU is used for AI recognition, and the effect is not ideal, so we are considering whether GPU can be used to accelerate AI recognition.
First check if the computer has a GPU. Open task Manager, and click Performance, the bottom of the GPU information, if the computer has a GPU, it means that the computer supports GPU.
The next step is to view the GPU information on the computer. To do this, right click on the desktop to find the NVIDIA control panel and click on the screen below:
Once in the NVIDIA control panel, click system info in the lower left corner. After entering this system information, we will see the relevant graphics card information. The graphics card details are on the right.
Once the graphics card details are found, we can install the corresponding Python Torch to speed up the identification using the GPU. You can’t use the GPU unless you install Nvidia’s CUDA Toolkit to use GPU acceleration.
Here are the version compatibilities of CUDA Toolkit:
