testing vllm cpu support in ci

Signed-off-by: Sivanantham Chinnaiyan <[email protected]>
sivanantha321 · Jun 26, 2024 · 4f08630 · 4f08630
1 parent a33ab02
commit 4f08630
Show file tree

Hide file tree

Showing 4 changed files with 5 additions and 4 deletions.
diff --git a/.github/workflows/e2e-test.yml b/.github/workflows/e2e-test.yml
@@ -860,7 +860,7 @@ jobs:
   test-llm:
     runs-on: ubuntu-latest
     needs:
-      [ kserve-image-build, predictor-runtime-build]
+      [ kserve-image-build]
     steps:
       - name: Checkout source
         uses: actions/checkout@v4

diff --git a/.github/workflows/huggingface-cpu.yml b/.github/workflows/huggingface-cpu.yml
@@ -52,10 +52,10 @@ jobs:
   # See also https://docs.docker.com/docker-hub/builds/
   push:
     # Ensure test job passes before pushing image.
-#    needs: test
+    needs: test
 
     runs-on: ubuntu-latest
-#    if: github.event_name == 'push'
+    if: github.event_name == 'push'
 
     steps:
       - name: Checkout source

diff --git a/test/e2e/predictor/test_huggingface.py b/test/e2e/predictor/test_huggingface.py
@@ -289,6 +289,7 @@ def test_vllm_openai_chat_completions():
             model_format=V1beta1ModelFormat(
                 name="huggingface",
             ),
+            image="sivanantha/huggingfaceserver-vllm-cpu@sha256:c4cffce3377c0fcc68e477058c14b61b8e98ad2f41f825302febf0d5eabf0d3d",
             args=[
                 "--model_id",
                 "facebook/opt-125m",

diff --git a/test/scripts/gh-actions/run-e2e-tests.sh b/test/scripts/gh-actions/run-e2e-tests.sh
@@ -30,5 +30,5 @@ fi
 PARALLELISM="${2:-1}"
 source python/kserve/.venv/bin/activate
 pushd test/e2e >/dev/null
-  pytest -m "$1" --ignore=qpext --log-level=INFO -o log_cli=true -n $PARALLELISM --dist worksteal
+  pytest -m "local" --ignore=qpext --log-level=INFO -o log_cli=true -n $PARALLELISM --dist worksteal
 popd