This is a tutorial session on tensor comprehensions. Facebook AI Research (FAIR) released a new C++ library called tensor comprehensions that aims to bridge the gap between researchers who think regarding mathematical equations and engineers focusing on the practical needs of running large-scale models on various hardware backend. The main differentiating feature of Tensor Comprehensions is that it represents a unique take on Just-In-Time compilation to produce the high-performance codes that the machine learning community needs, automatically and on-demand. The aim of this tutorial is two-fold: (1) New HPC users will be able to appreciate the importance of HPC in machine learning (2) New HPC users will be able to appreciate the sophisticated tools that enable to them produce automatic code in relatively simple and straightforward manner.
