Yihua Cheng
I quickly went through this. Since it's quite a large PR, can we maybe split this into 2 PRs? One is a pure controller interface, and the other is the...
Another thing is I think we should talk about how to deal with the `hot_cache` since it will probably be replaced by a local CPU backend (#514 )
Closing this issue because we have already done it via #918
Thanks @qyy2003 !! I have a question about the format of the configuration: In the configuration: ``` local_device: "disk_url://http://localhost:4322" ``` This is not a standard format of a URI. Usually,...
I think this is neither LMCache-core issue nor vLLM issue. Should be an example-related issue. @YaoJiayi I believe you have met this before?
@pramodk Btw, we do have a docker file at https://github.com/LMCache/LMCache/blob/dev/docker/Dockerfile
@maobaolong The table is great! Thanks! Just want to add: there will be a nightly docker build pipeline merged soon: #756
I also think we should put the something like a compatibility matrix. The documentation website is also a good place. Any ideas @YaoJiayi @KuntaiDu @Siddhant-Ray
@pramodk The current latest manually-built docker image is `lmcache/vllm-openai:2025-05-27-v1`. We plan to update the documentation site about the docker image tags once the nightly docker image build pipeline is online....
@maobaolong Please also take a look since you are most familiar with the plugin systems.