[Feature Request] Publish results

[JackWolfard]

Hi, thanks for creating this infrastructure, I was curious if publishing the results of running this against different releases of SGLang, vLLM, TGI, etc. on different HW for different models and use cases is in the roadmap?

[2016bgeyer]

Yes, please do this!

Having published results for some common configuations, even if not in a leaderboard fashion, would be super helpful

[Hugoch]

Hey @JackWolfard , we internally started to do that for evaluation of text-generation-inference, maybe we could publish those. What would be the most relevant models for you rn?

[JackWolfard]

Hey @JackWolfard , we internally started to do that for evaluation of text-generation-inference, maybe we could publish those. What would be the most relevant models for you rn?

I would be interested in the language models from MLPerf Inference: Datacenter.