The x86 container image seems to be missing about llm-starter-pack HOT 8 CLOSED

Thank you, @brandonp42! We will take a look at this next week after KubeCon + CloudNativeCon NA wraps

from llm-starter-pack.

Similarly in kind/bot-deployment.yaml
https://hub.docker.com/r/pritianka/bot/tags
This is arm64 only.

from llm-starter-pack.

I managed to get the LLM to deploy on amd64 using the stock ollama/ollama image, but the final bit -- the bot-deployment.yaml -- references an arm-only image. Perhaps posting the Dockerfile for the bot would be sufficient to allow full deployment on non-arm systems.

from llm-starter-pack.

Warning Failed 25m (x6 over 27m) kubelet Error: ImagePullBackOff
Normal BackOff 2m28s (x107 over 27m) kubelet Back-off pulling image "onlydole/ollama-mistral:latest"

from llm-starter-pack.

Same issue. Will try an ARM EC2

from llm-starter-pack.

If the source and Dockerfile for the 2 images is available somewhere, I'd be happy to test build on intel/amd and make a PR. I'm dying to try this...

from llm-starter-pack.

Hi everyone, sorry for the delay!

We've gone ahead an automated building the container here. That means a fresh one with each PR! It's getting late so I haven't had a chance to try so give it a whirl!

docker pull ghcr.io/cncf/llm-starter-pack:latest

If any of you have the expertise to make the github action better it would be appreciated!

It'd be great to have a full set of artifacthub labels, convenience tags, and some more placeholder documentation. So that we can just have one kit to make anyone's demos autobuilding. Feedback appreciated!

https://github.com/cncf/llm-starter-pack/pkgs/container/llm-starter-pack

from llm-starter-pack.

We have added these new dynamic builds (thank you again @castrojo for working on these), and any PRs/Issues/improvements are welcome.

from llm-starter-pack.