-
Notifications
You must be signed in to change notification settings - Fork 2
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Added huggingfaceserver
rock and tests
#103
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Missing this from upstream https://github.com/kserve/kserve/blob/v0.13.0/python/huggingface_server.Dockerfile#L35
Upstream uses nvidia cuda base image so we need to provide also those to the image https://github.com/kserve/kserve/blob/v0.13.0/python/huggingface_server.Dockerfile#L1
@misohu Added missing vllm pip package and cuda-12-1 runtime. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There is a problem with the rock when trying to run it with docker ... container is stuck in restarts (that is expected). But when I bash into container and check logs
docker exec -ti d7b09f4eb8c5 bash
pebble logs
I am getting
_daemon_@d7b09f4eb8c5:/$ pebble logs
2025-01-27T10:59:37.671Z [huggingfaceserver] : No module named huggingfaceserver
2025-01-27T10:59:38.223Z [huggingfaceserver] : No module named huggingfaceserver
2025-01-27T10:59:39.332Z [huggingfaceserver] : No module named huggingfaceserver
2025-01-27T10:59:41.514Z [huggingfaceserver] : No module named huggingfaceserver
2025-01-27T10:59:45.927Z [huggingfaceserver] : No module named huggingfaceserver
2025-01-27T10:59:53.972Z [huggingfaceserver] : No module named huggingfaceserver
@misohu Issue with "No module named huggingfaceserver" fixed by adding |
I was able to test this rock with kserrve operator by running the Inferenceservice for google's bert with hugging face serving runtime. The test is now part of this PR canonical/kserve-operators#298. Test was running also this rock. |
We agreed to merge this rock based on manual tests. |
Description
This is re-open request for #91.
Logs from tests
juju status
after all tests are passed:Logs of passed integration tests
tox -vve integration -- --model kubeflow --keep-models
(only last included due to PR message maximum length):