diff --git a/.github/workflows/workflow_finetune.yml b/.github/workflows/workflow_finetune.yml
index 181f46df4..5a8e32720 100644
--- a/.github/workflows/workflow_finetune.yml
+++ b/.github/workflows/workflow_finetune.yml
@@ -11,10 +11,10 @@ on:
         default: '10.1.2.13:5000/llmray-build'
       http_proxy:
         type: string
-        default: 'http://proxy-chain.intel.com:911'
+        default: 'http://10.24.221.149:911'
       https_proxy:
         type: string
-        default: 'http://proxy-chain.intel.com:911'
+        default: 'http://10.24.221.149:911'
       runner_config_path:
         type: string
         default: '/home/ci/llm-ray-actions-runner'
diff --git a/.github/workflows/workflow_inference.yml b/.github/workflows/workflow_inference.yml
index 8b3ac1d56..03269a4a4 100644
--- a/.github/workflows/workflow_inference.yml
+++ b/.github/workflows/workflow_inference.yml
@@ -11,10 +11,10 @@ on:
         default: '10.1.2.13:5000/llmray-build'
       http_proxy:
         type: string
-        default: 'http://proxy-chain.intel.com:911'
+        default: 'http://10.24.221.149:911'
       https_proxy:
         type: string
-        default: 'http://proxy-chain.intel.com:911'
+        default: 'http://10.24.221.149:911'
       runner_config_path:
         type: string
         default: '/home/ci/llm-ray-actions-runner'
diff --git a/dev/docker/Dockerfile.vllm b/dev/docker/Dockerfile.vllm
index 3b4bd00c8..e4eb63d06 100644
--- a/dev/docker/Dockerfile.vllm
+++ b/dev/docker/Dockerfile.vllm
@@ -38,6 +38,3 @@ RUN --mount=type=cache,target=/root/.cache/pip pip install -e .[cpu] -f https://
 RUN --mount=type=cache,target=/root/.cache/pip \
     source /opt/conda/bin/activate base && ./install-vllm-cpu.sh
 
-# TODO: workaround, remove this when fixed in vllm-cpu upstream
-RUN --mount=type=cache,target=/root/.cache/pip \
-    pip install xformers