feat(megatron-bridge): adopt TEGroupedMLP per-expert quantizers (#1550) for W4A16 NVFP4 four_over_six - #2072

feat(megatron-bridge): adopt TEGroupedMLP per-expert quantizers (#1550) for W4A16 NVFP4 four_over_six#2072

yueshen2016 wants to merge 18 commits into

mainNVIDIA/Model-Optimizer:mainfrom

qad/te-per-expert-1550NVIDIA/Model-Optimizer:qad/te-per-expert-1550Copy head branch name to clipboard