How to use NVIDIA GPUs for Ozeki AI

This page provides information on how to use your NVIDIA GPU in Ozeki AI studio to execute AI models.

Step 1. Install NVIDIA Drivers

You need to ensure that you have the correct NVIDIA drivers installed for y our GPU. These drivers are essential to enabling CUDA (NVIDIA’s parallel computing platform) to work with your hardware.

  • Download and install the latest drivers for your GPU from the NVIDIA website.
  • After installing, you can verify your installation by running nvidia-smi in a terminal or command prompt to check the installed drivers and CUDA version.

Nvidia driver downloads:

https://www.nvidia.com/Download/index.aspx

Step 2. Install CUDA Toolkit

CUDA is a parallel computing platform and API model from NVIDIA that allows you to use your NVIDIA GPU for general-purpose computing. Even though Ozeki AI studio may use CUDA internally, you need to have the correct version of the CUDA toolkit installed that matches your NVIDIA driver.

  • Download and install the CUDA toolkit from the CUDA Toolkit Download page.
  • During installation, it will automatically install the required components such as:
    • CUDA libraries (cuBLAS, cuDNN, etc.)
    • nvcc (NVIDIA's compiler for CUDA)
    • The driver (if required)

After installation, verify that CUDA is properly installed by running:

nvcc --version

or

nvidia-smi

Nvidia CUDA Toolkit downloads:

https://developer.nvidia.com/cuda-downloads

Optional steps

The following steps are optional. They are only needed if you would like to do AI development on your PC. If you simply want to use AI chat you can stop here.

Step 3. Install cuDNN (CUDA Deep Neural Network Library)

cuDNN is a GPU-accelerated library for deep neural networks, and many machine learning libraries (such as those used in Ozeki AI studio models) often depend on cuDNN for better performance.

  • You can download cuDNN from the NVIDIA cuDNN page.
  • Extract and copy the cuDNN files into the appropriate CUDA directories, typically under:
    • C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\vX.X\lib\x64\
    • C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\vX.X\include\

Make sure to match the cuDNN version with the installed version of the CUDA Toolkit.

Step 4. Set Environment Variables

After installing CUDA and cuDNN, ensure that your environment variables point to the correct paths. This step is necessary for Windows systems:

  • Add the CUDA bin directory to your PATH environment variable: 
    C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\vX.X\bin
  • Also add cuDNN library paths if not done automatically.

Step 5. Verify the Installation

Once you have installed all the necessary components, you can check the installation using a CUDA-enabled application, or through nvidia-smi to ensure the GPU is recognized and CUDA is available.

Summary of Installations:

  1. NVIDIA Drivers
  2. CUDA Toolkit
  3. cuDNN Library
  4. Set Environment Variables (CUDA & cuDNN paths)

With these components installed, your system should be ready to run your AI models in Ozeki AI Studio with CUDA acceleration on your NVIDIA GPU.

More information