Clarification on the VLM project

[mishra-18]

Hi,
I'm Shubh, I'm working on the benchmarking suite for Vision-LLM project. The proposal-making period is about to end in a week and I just wanted to make sure from the mentors @junwenwu and @helena-intel that I'm getting everything right!

The accuracy checker for the models is for checking the performance of their respective tasks and the performance benchmark scripts for both unimodal and joint are for running these accuracy checkers on their respective models. Does this benchmark in addition to the accuracy checkers also involve checking other performance metrics like memory consumption, time, etc? Also, all these metrics and scripts are to be built from scratch or to be added to the pre-existing files in openvino.genai/llm_bench/python/who_what_benchmark?

[mishra-18]

I request the associated mentors @junwenwu and @helena-intel to review my proposal before the application period ends. I've sent them to your respective emails!