This blog post describes theoretical methods to reduce model size. Size reduction for deep learning models is an active field of research. Those methods are truly performant, but the specific type of machine learning models used involves extremely deep and complex architectures (Simonyan et al. (2014), He et al. (2015), Szegedy et al. (2016)). How can we simply transform a deep model into a lighter one without decreasing drastically its performances ? Moreover, does it exist specialized architectures to build light models while achieving state-of-the-art performances ? Note that researchers test their algorithms using different datasets. Thus the cited accuracies cannot be directly compared per se.
This blog post is available on Medium.