Update Run Results and benchmark results

[TheKanter]

Hi, can we update run results and benchmark results to be clear

• [x] that accuracy is the primary metric,

• [x] time to train and throughput are optional. A

• [x] additionally, per discussion, we probably want to simplify # of runs and scoring given accuracy is the goal.

• https://github.com/mlcommons/science/blob/main/policy.adoc

• https://github.com/mlcommons/science/blob/main/submit.adoc

[nv-rborkar]

Thanks @laszewsk for taking care of this. I will review & comment, latest by Monday.

[nv-rborkar]

Thank you @laszewsk , Juri, Gregor. Policy & Submission documents look good to me, thanks for the clarifications.