🗜🗜Edge#226: DeepSpeed Compression, a new library for extreme compression of deep learning models

It combines compression and system optimization techniques for building smaller and more efficient deep learning architectures

On Thursdays, we dive deep into one of the freshest research papers or technology frameworks that is worth your attention. Our goal is to keep you up to date with new developments in AI to complement the concepts we debate in other editions of our newsletter.

💥 What’s New in AI: Microsoft’s Open Sourced a New Library for Extreme Compression of Deep Learning Models

Large neural networks have been dominating the deep learning space for the last few years. While the performance of large deep learning architectures is certainly impressive, its operational requirements remain prohibited for most organizations. Not surprisingly, there has been a lot of effort in areas like model compression that can help reduce the size and inference computation of deep learning models. Similarly, there has also been a resurgence of system optimization techniques that can improve the inference of models without sacrificing their size. The combination of model compression and system optimization is quite powerful in order to enable more efficient deep learning architectures. Recently,