mx num seqs expose to benchmark_throughput script

[seungrokj]

mx num seqs expose to benchmark_throughput script

@shajrawi @dllehr-amd @JArnoldAMD

[shajrawi]

Can you please make the changes in benchmarks/profiling version of the *py benchmark files? This is our own copy we can change in rocm/vllm.

Changing the benchmarks/ folder scripts would require going through upstream first to avoid merge conflicts.

[seungrokj]

Can you please make the changes in benchmarks/profiling version of the *py benchmark files? This is our own copy we can change in rocm/vllm.

Changing the benchmarks/ folder scripts would require going through upstream first to avoid merge conflicts.

@shajrawi
Sure will do! Thank you for the guidance!