Unleashing the potential of Nvidia Rapids: Accelerating ML Algorithms with cuML libraryCheck Reference
Intro to Rapids
Nvidia Rapids is an open-source suite of software libraries that provide high-performance data science and machine learning tools for GPUs. cuML is a library within Rapids that provides GPU-accelerated implementations of common machine learning algorithms, such as linear regression, logistic regression, and support vector machines.
What talk will comprise of!
The talk will be a presentation-cum-colab based session where I'll be training one machine learning algorithm using Nvidia Rapids cuML library alongwith the usage of cuDF while telling how it overpowers pandas in terms of functioning & performance and will also train it using Scikit-learn i.e. in CPU in order to tell the audience about why Nvidia Rapids offers such high-performance. This FOSS project, launched and maintained by Nvidia has a lot of varied use cases which will be discussed in my presentation alongwith the pros & cons of having GPU accelerated data science and how can we contribute to such an important FOSS project as well.