Helm: Add Nvidia GPU support for ChatQnA #225
Conversation
|
I don't think we have test bed for Nvidia in CI infrastructure. So it is ok that no tests cover this feature. |
PeterYang12
force-pushed
the
chatqa-gpu-support-helm-charts
branch
from
9c92d5a to
8b04a6d
Compare
Done. |
Got it. Thank you. |
PeterYang12
force-pushed
the
chatqa-gpu-support-helm-charts
branch
from
8b04a6d to
542d94d
Compare
PeterYang12
force-pushed
the
chatqa-gpu-support-helm-charts
branch
from
542d94d to
00c2743
Compare
There was a problem hiding this comment.
Even though there is an CI error for ChatQnA, but we're 100% sure this error is not related to this PR. We'll tackle this kind of ephemeral failure in other places. We've seen this kind of ephemeral failures several times, especially when there is quite some tests from different repo on the same xeon node.
|
How about to re-trigger the failed test task? It should be okay if there is no any nv test resource in CI. |
Do you know how to re-trigger the tests? Should I update My PR? |
you can re-trigger it by re-run the job, click the details and there is a "re-run jobs" button |
I don't have the access to do this. @daisy-ycguo Can you help me? |
I've re-run it |
TY. |
|
Our CI infrastructure doesn't support Nvidia test. nv-values is not covered by CI tests and will not cause any CI test failures. I'm fine to merge this PR if @PeterYang12 has tested OK locally. @PeterYang12 please rebase this PR, then we can merge this one. |
1. Add Helm-charts support 2. Add manifests support Signed-off-by: PeterYang12 <[email protected]>
daisy-ycguo
force-pushed
the
chatqa-gpu-support-helm-charts
branch
from
35be7a1 to
71daaf1
Compare
Automatically popluate the default endpoint into no_proxy if proxy is set. If users specify the endpoints with their own value, it won't be populated into no_proxy, because we can't assume the user specified endpoints should be accessed through proxy. Fix issue opea-project#225. Signed-off-by: Lianhao Lu <[email protected]>
Automatically popluate the default endpoint into no_proxy if proxy is set. If users specify the endpoints with their own value, it won't be populated into no_proxy, because we can't assume the user specified endpoints should be accessed through proxy. Fix issue opea-project#225. Signed-off-by: Lianhao Lu <[email protected]>
Description
Add Nvidia GPU support for ChatQnA in helm-charts.
Issues
N/A
Type of change
List the type of change like below. Please delete options that are not relevant.
Dependencies
N/A
Tests
Describe the tests that you ran to verify your changes.