diff --git a/.buildkite/scripts/hardware_ci/run-xpu-test.sh b/.buildkite/scripts/hardware_ci/run-xpu-test.sh index afed058529..ab8c2fd5d6 100755 --- a/.buildkite/scripts/hardware_ci/run-xpu-test.sh +++ b/.buildkite/scripts/hardware_ci/run-xpu-test.sh @@ -52,5 +52,6 @@ time timeout -k 30 30m docker run \ pip install tblib==3.1.0 cd /workspace/vllm-omni pytest -v -s -m "core_model and xpu and B60" - pytest -v -s -m "advanced_model and xpu and B60" -k "not omni_expansion" + export VLLM_XPU_USE_SAMPLER_KERNEL=0 # NOTE: Remove this after vLLM v0.21.1 is merged. Fixes Qwen2-5 omni-expansion tests. + pytest -v -s -m "advanced_model and xpu and B60" '