llama3.1-8b support and calibration

[arekay-nv]

Ensure that llama3.1-8b executes end-to-end and reports number statistically similar to legacy loadgen as submitted in MLPerf inference v5.1 for all scenarios.
Add documentation to enable reproducing the results.

[tianmu-li]

Just to clarify, does this mean offline/server performance/accuracy and compliance?

[arekay-nv]

Just to clarify, does this mean offline/server performance/accuracy and compliance?

Yes, complete coverage for the model. If focusing on a specific scenario, you can create a sub-issue under this issue and work on it.

[arekay-nv]

@attafosu - can this be closed?

[attafosu]

@attafosu - can this be closed?

Yes. Resolved by #53 and #100