Pramod Kumbhar

thank you for quick reply, @KuntaiDu! A related question: If I want to build a **latest & compatible versions** of lmcache and vllm, which commits/tags I use for source build?...

> Btw, we do have a docker file at https://github.com/LMCache/LMCache/blob/dev/docker/Dockerfile Thanks for pointing that! I see that it uses vllm latest tag i.e. `v0.9.0.1`. Does it mean if I take...

> Just want to add: there will be a nightly docker build pipeline merged soon: https://github.com/LMCache/LMCache/pull/756 Just would like to confirm - as part of thus building workflow, we are...

> In the new docker images, vLLM and LMCache and all the other dependencies are installed under /opt/venv thanks, I missed that change! Have been using `--entrypoint /bin/bash` for offline...

We are creating venv here: https://github.com/LMCache/LMCache/blob/2eea5ae50503ffd287fc9b94144a09d246c93709/docker/Dockerfile#L33 So we just do: ```console $ docker run ... --entrypoint /bin/bash lmcache/vllm-openai:2025-05-27-v1 root@x008:/workspace# source /opt/venv/bin/activate (venv) root@x008:/workspace# ```

what do you mean by "as the runtime lmcache module"? Once environment is activated, we have vllm also installed and available in the same environment: ```console # vllm --help INFO...

thanks, @hickeyma! Certainly! The building workflow is very helpful, thanks for all the work! I was curious about is the following: > Just would like to confirm - as part...

From our internal discussion, we said this is good to go! @nrnhines: if you see any corner cases, let us know!

Hello All, Is there a target date for releasing aarch64 wheels via official pypi releases? As far as [I see](https://pypi.org/project/vllm/0.9.1/#files), we only provide `x86_64` wheels today. Thanks!