Exploring the latest developments in DL on mobile devices

With the complexities of DL and AI combining with mobile applications, software and hardware optimizations are being constantly made to run the models efficiently on devices. Let's look at some of them.

Google's MobileNet

Google's MobileNet was launched in 2017. It is a set of mobile-first computer vision models based on TensorFlow, very carefully optimized to run efficiently within a restrictive mobile environment. It acts as a bridge between accuracy in complex neural network structures and performance constraints on mobile runtimes. Since the models have the capability to run locally on the device itself, MobileNet has the advantages of security,...