Export quantized/co-trained MTP weights instead of copying BF16 - #2174

Export quantized/co-trained MTP weights instead of copying BF16#2174

yeyu-nvidia merged 5 commits into

NVIDIA:mainNVIDIA/Model-Optimizer:mainfrom

yeyu-nvidia:yeyu/mtp-quant-exportyeyu-nvidia/Model-Optimizer:yeyu/mtp-quant-exportCopy head branch name to clipboard