Model techniques map
Techniquesinference & servinginference quantization

specific method · filed under inference & serving

Per-tensor FP8 quantization

Uses static max-calibrated per-tensor FP8 scales for shared-expert and Mamba linear-layer GEMMs.

source
1
model
1
lab adopt it
1
strongest
core

How sources treat it

One count per evidence span, weakest treatment to strongest.

core 1

Documented in

Evidence

1 span quoted from the sources, strongest treatment first.

FP8 per-tensor GEMMs for shared experts and Mamba linear layers; static max-calibrated per-tensor scales.

corepost trainingin Nemotron 3 UltraNVIDIA

Filed alongside

Other methods under inference & serving :: inference quantization.