From the course: NVIDIA Certified Associate AI Infrastructure and Operations (NCA-AIIO) Cert Prep

Unlock this course with a free trial

Join today to access over 25,300 courses taught by industry experts.

The NVIDIA differentiator

The NVIDIA differentiator - NVIDIA Tutorial

From the course: NVIDIA Certified Associate AI Infrastructure and Operations (NCA-AIIO) Cert Prep

Start my 1-month free trial Buy for my team

The NVIDIA differentiator

“

So, if these frameworks exist, what is NVIDIA differentiating here? What is the NVIDIA differentiator? Let me explain that. So, let's say you have written your Python code, right? Whatever model you are building, whatever use case you have, you have written your Python code for that. And now, that would be provided to the framework engine which you have used behind the scene like PyTorch or TensorFlow. What these frameworks will do, they will automatically use optimized library so your model trains much faster on the GPUs. What these frameworks will do behind the scene, they may leverage something called Optimization Layer like CUDNN. It is your CUDA Distributed Neural Network. What the CUDNN is, it is a collection of highly optimized low-level GPU routine that provides you pre-built natural network operation for things like CNN and RNN. These are different methods and this would allow you to ensure that you are using optimal libraries for your model training. through this cuDNN…

Contents