single mention · inference & serving
Sensitivity-driven mixed-precision quantization
Sensitivity-driven mixed-precision quantization assigns different quantization formats to different layers via a constrained search rather than fixed rules.1
How this definition was written
Summarised by glm-5.3-flash from 1 source; every sentence is checked against the quote it cites.
replace hand-authored layer rules with a sensitivity- constrained mixed-precision search
— NVIDIA Nemotron 3.5 Super and Lightning Technical Report
Also called Sensitivity-Driven Mixed Precision.
Filed under inference & serving › inference quantization