Computing Colloquium Student Presentation featuring Dave Johnson

Building Scalable Deep Learning Workflows with PyTorch

Presenter: Dave Johnson
Computing PhD Student, Data Science Emphasis

Location: In person in CCP 259 or register to attend via Zoom

Abstract: I will discuss building deep learning (DL) workflows from scratch in PyTorch, with an eye towards scaling their training to high-performance computing clusters. The discussion will be motivated by disparate DL projects; accordingly, it will focus on how to implement desirable workflow features and write modular PyTorch code in order to minimize changes across projects and machines. An exchange of ideas is encouraged!