Towards a Standardized Representation for Deep Learning Collective Algorithms
HOTI - Hot Interconnects Symposium via YouTube
Overview
Learn about standardizing deep learning collective algorithms in this 19-minute technical conference presentation from the Hot Interconnects Symposium. Explore research findings presented by a team of experts including Jinsun Yoo, William Won, Meghan Cowan, Nan Jiang, Benjamin Klenk, Srinivas Sridharan and Tushar Krishna as part of Technical Paper Session C focused on Collectives for Deep Learning. Gain insights into the latest developments in creating unified representations for collective algorithms used in deep learning applications, chaired by Amanda Bienz from the University of New Mexico.
Syllabus
Day 2 09:10: Towards a Standardized Representation for Deep Learning Collective Algorithms
Taught by
HOTI - Hot Interconnects Symposium