Today, Artificial Intelligence (AI) is a hot topic for researchers and developers in academia and industry. In this two day course, we consider AI to be comprised of two major fields namely Machine Learning (ML) and Deep Learning(DL). Neither are new, but their recent revival and attention was made possible by technological progress. The two-day training course will highlight this progress and get you started with what you need to know for state of the art ML/DL development. We separate the course into two days, with the first covering the Intel architectures related to AI, and NVIDIA on the second day, respectively. For each architecture, hands-on Tensorflow exercises show you what DL is about and how to use the different hardware back-ends efficiently. After the training is finished, optimization of the trained DL models and their deployment to the target application/architecture is discussed. Depending on the architecture, it is necessary to know about different toolkits such as OpenVINO or TensorRT.
On day 1 we also give you an introduction to ML with scikit-learn and what you need to know to run it efficiently on the CPU (i.e. Intel Architecture). We also give you the chance to offload the trained DL model to the Intel Movidius Compute Stick (inference engine) and run it on your own. The training will close at the end of day 2 with a hands-on with IT4Innovations' new AI clusters (NVIDIA DGX-2 and Barbora). Both the hands-on DL/ML exercises and access to IT4Inovations' clusters give you the skills to start your own ML/DL projects more easily. Exercises use Python and Jupyter notebooks for convenience.
The course is for anyone interested in Machine/Deep Learning (ML/DL) training, optimization and deployment on the latest Intel (i.e. CPUs, iGPUs, FPGAs, NPUs) and NVIDIA architectures (i.e. GPUs). To avoid redundancies, the course is balanced by giving a general introduction on day 1 and highlighting differences between the architectures on day 2. Participants will be given working examples for DL and ML that they can use as a starting point for their own projects. Furthermore, the training includes a hands-on to get started with AI on IT4Innovation's clusters NVIDIA DGX-2 and Barbora.
Georg Zitzlsberger is a research specialist for Machine and Deep Learning at IT4Innovations. He recently received his certification from NVIDIA as a University Ambassador of the NVIDIA Deep Learning Institute (DLI) program. This certification allows him to offer NVIDIA DLI courses to academic users of IT4Innovations' HPC services. In addition, in collaboration with Bayncore, he is a trainer for the Intel HPC and AI workshops and conferences carried out across Europe. He has been contributing to these events, which are held for audiences from industry and academia, for over two years.
Stephen Blair-Chappell is currently Technical Director at Bayncore and was previously a Technical Consulting Engineer at Intel for 18 years. At Bayncore he currently leads the team of consultants providing HPC and AI training on Intel Architecture.
The first day of the course is kindly provided by Bayncore. This work was also supported by The Ministry of Education, Youth and Sports from the Large Infrastructures for Research, Experimental Development and Innovations project “IT4Innovations National Supercomputing Center – LM2015070” and by the PRACE-6IP project - the European Union’s Horizon 2020 research and innovation programme under grant agreement No 823767.