Common CM interface and automation to reproduce MLPerf inference v3.1

[gfursin]

Run MLPerf inference benchmarks via CM (CM workflows automatically adapt to different OS and hardware):

cmr "generate-run-cmds inference" --implementation=nvidia --model=bert-99

Prepare official submission for Edge category:

cmr "generate-run-cmds inference _submission _full" --implementation=nvidia --model=bert-99

Prepare official submission for DataCenter category:

cmr "generate-run-cmds inference _submission _full" --implementation=nvidia --model=bert-99 --category=datacenter --division=closed

The MLCommons taskforce on automation and reproducibility is gradually adding CM support for all implementations and models:

{model}	Ref implementation	Nvidia	Intel	Qualcomm
bert-99