The Luxembourg NVidia AI Technology Center (NVAITC) organizes with the support of the UL HPC team a Data Science Webinar
When? 26 March – 2 pm CET
Who? Dr. Christian Hundt, NVIDIA AI Technology Center
What?
Efficient pipeline design is crucial in data science. When composing complex end-to-end workflows, you may choose from a wide variety of building blocks, each of them specialized for a dedicated task. For example, you may use an efficient CSV reader to load time series from text files to later pre-process them in a feature extraction stage before training a classifier. Conversion between data formats is an error-prone and performance degrading task. Let us change that!
This webinar discusses zero-copy interoperability across the following GPU-accelerated data science frameworks: TensorFlow, PyTorch, MXNet, cuDF, CuPy, Numba and Jax. It also details how to interact with two of the most commonly used CPU-based data science frameworks: Pandas and Numpy.
Registration:
Write an informal mail to chundt@nvidia.com to request access to the webinar. You will then receive an invitation via mail.