Hardware's to deploy machine learning codes

Hello there! I want to know that there are how many hardwares that are used to deploy machine Learning codes. Tell me the names of the hardware's and the methods to deploy codes in the hardwares. Thanks

Answer 1

[Walter Roberson]

There is a long list of options; see https://www.mathworks.com/hardware-support/home.html

You have the difficulty that you have not distinguished between the training phase and the phase of running using the already-trained information. Deploying the training phase is more difficult and tends to be rather inefficient unless the target platform has an nvidia GPU. Deploying a pre-trained network can be done to a wide variety of hardware, limited mostly by memory and required execution speed.

For example it might be practical to train a modest facial recognition network on a known set of pictures, all authorized people (and possibly also some people specifically prohibitted), and deploy that set of image into an arduino -- you might be able to handle (for example) 10 people. But the deployed version would not have any capacity to learn new people, and probably would not be able to handle hundreds of people.. for hundreds of people you might have to deploy to something like an Raspberry Pi 3b+ (more memory, faster)