Heterogeneous computing for Deep Learning: deploying generative models via Intel OneAPI

The main goal of the project is to test Intel"s oneAPI as a framework to speed up the 3DGAN training processes across multiple hardware architectures. The work consists in analyzing the behavior of 3DGAN under Intel oneAPI by using the Intel AI Analytics Toolkit. This toolkit provides Python frameworks and tools to accelerate end-to-end data science and analytics pipelines on Intel architectures. It uses oneAPI libraries for low-level compute optimizations, thus, maximizing performance from preprocessing through machine learning. As a result, oneAPI made a difference, the model converged faster