Global ETD Search

Comparing the Performance of Small Word-Size Floating-Point Numerics to Fixed-Point Numerics in Neural Networks

Vennapusa, Lakshmi GrishmaGeorgia Institute of Technology2025

… it suffers from limited dynamic range and quantization inflexibility. This thesis introduces an alternative approach—Adaptive Precision Training (APT)—which leverages reduced-precision floating-point formats (FP8, FP12, FP16) for dynamic, layer-wise quantization during training. APT …

gatech Repository record for Comparing the Performance of Small Word-Size Floating-Point Numerics to Fixed-Point Numerics in Neural Networks (opens in a new tab)