As deep learning becoming prevalent in adoption of practical AI, it is important to lower barrier to entry for DL adoption for the developer community. This can be done by 1) retaining and reusing existing and popular libraries as much as possible and employing them in development of DL applications, and 2) Automating model development as much as possible though usage open source AutoML tools.
Deep Numpy enhances Numpy by adding GPU support and parallel processing to it, while aspiring to remain 100% numpy compatible. It enables easy-to-use and easy-to-extend AutoML with a focus on deep learning, and making AutoML deploy in real-world applications.
In this talk, we focus on the use of Deep Numpy and AutoGluon for rapid development of DL models.