Skip to content

[PTQ] Store FP32 global scaling factors (absmax or scale_inv) for all quantized activations and weights. - #3296

Open
cspades wants to merge 5 commits into
NVIDIA:mainfrom
cspades:cye/scale_factor_buffering
Open

[PTQ] Store FP32 global scaling factors (absmax or scale_inv) for all quantized activations and weights.#3296
cspades wants to merge 5 commits into
NVIDIA:mainfrom
cspades:cye/scale_factor_buffering

Commits

Commits on Jul 31, 2026