Efficient Deep Neural Network Training with a Novel 5.3-bit Block Floating Point Data Format

M. H. Sadi, C. Sudarshan, S. Nassif, N. Wehn
January 2025

Abstract

See publication for more information.

Type
Conference paper
Publication
Efficient Deep Neural Network Training with a Novel 5.3-bit Block Floating Point Data Format

Citation: Efficient Deep Neural Network Training with a Novel 5.3-bit Block Floating Point Data Format. M. H. Sadi, C. Sudarshan, S. Nassif, N. Wehn. Accepted for Publication, IEEE Transactions on Circuits and Systems for Artificial Intelligence. https://ieeexplore.ieee.org/document/10878478