Jiaxin Shan
Jiaxin Shan
 Seems the 2nd build for gid patch result in larger image.. But we only install `wget` and download the whl. whl is just 9MiB
@yyzxw sorry for late response. `Dockerfile.kvcache` is not the right dockerfile. that's the image to sync kv cache information to redis. Here, we focus more on the infinistore image itself....
@kerthcet great point! the AIBrix repo used to have the ruleset on main branch and release branch. It was not successfully transferred to current repo automatically. - I will create...
VKE team already have some tools, we should review and evaluate that work.
the model parameter like context length and parallelism could be different which brings additional challenges to get apple-to-apple comparison result
We should leverage the deepseek-33b case to perfect the solution here. @kr11 Let's have a short discussion tomorrow on the next steps. VKE will publish their tools and we probably...
In v0.1.0, we should focus on using, polishing, improving existing tools build by VKE.
parameter tuning or profiling would be advance features, we plan to work on in v0.2.0.
This is the auto-tuning or profiling related stories. We also come up ideas like LLMPilot, v0.3.0 is too tight for this story and it can be postponed to v0.4.0
related paper work: https://arxiv.org/html/2502.13965v1