Community support and discussions about the Intel® Distribution of OpenVINO™ toolkit, OpenCV, and all things computer vision-related on Intel® platforms.
From the tutorials they only seem to be using the Intel® Neural Compute Stick 2 on pre-trained models to determine the Neural Networks outcome only. But can this hardware be used to train the weights of new models? Or do you have to buy a seperate GPU to do this?
the NCS 2 focuses on deploying networks for inference. Therefore, it is not adapted for training, I would suggest you look first at training on CPU and eventually to GPU.
