Efficient DNN Training Using Vectorized Block-Scaled GeMMs with Adaptive Block Shapes

IFIP advances in information and communication technology 2026

View at publisher

Authors: Nitish Satya Murthy, Nathan Laubeuf, Debjyoti Bhattacharjee, Francky Catthoor, Marian Verhelst

DOI: 10.1007/978-3-032-09575-6_3

Cited by: 0

References: 13

Fields of Study & Subjects

Tensor decomposition and applications Stochastic Gradient Optimization Techniques Numerical Methods and Algorithms

References 13