diff --git a/tools/ci_build/github/azure-pipelines/templates/py-linux-gpu-cuda12.yml b/tools/ci_build/github/azure-pipelines/templates/py-linux-gpu-cuda12.yml
index f68847afff379..bb79356169f1d 100644
--- a/tools/ci_build/github/azure-pipelines/templates/py-linux-gpu-cuda12.yml
+++ b/tools/ci_build/github/azure-pipelines/templates/py-linux-gpu-cuda12.yml
@@ -53,8 +53,8 @@ jobs:
       inputs:
         targetType: filePath
         filePath: tools/ci_build/github/linux/run_python_dockerbuild.sh
-        # please check ONNXRUNTIME_CUDA_VERSION in tools/ci_build/github/linux/build_linux_arm64_python_package.sh
-        arguments: -i onnxruntimecuda118xtrt86build${{ parameters.arch }} -d "GPU" -c ${{ parameters.cmake_build_type }} -x "${{ parameters.extra_build_arg }}"
+        # please check ONNXRUNTIME_CUDA_VERSION in tools/ci_build/github/linux/build_linux_python_package.sh
+        arguments: -i onnxruntimecuda118xtrt86build${{ parameters.arch }} -d "GPU" -c ${{ parameters.cmake_build_type }} -x "${{ parameters.extra_build_arg }} -u "12.2"
 
     - task: PublishBuildArtifacts@1
       displayName: 'Publish Artifact: ONNXRuntime python wheel'
diff --git a/tools/ci_build/github/azure-pipelines/templates/py-linux-gpu.yml b/tools/ci_build/github/azure-pipelines/templates/py-linux-gpu.yml
index 0f988039339af..412a02e628e2f 100644
--- a/tools/ci_build/github/azure-pipelines/templates/py-linux-gpu.yml
+++ b/tools/ci_build/github/azure-pipelines/templates/py-linux-gpu.yml
@@ -64,8 +64,8 @@ jobs:
       inputs:
         targetType: filePath
         filePath: tools/ci_build/github/linux/run_python_dockerbuild.sh
-        # please check ONNXRUNTIME_CUDA_VERSION in tools/ci_build/github/linux/build_linux_arm64_python_package.sh
-        arguments: -i onnxruntimecuda118xtrt86build${{ parameters.arch }} -d "GPU" -c ${{ parameters.cmake_build_type }} -x "${{ parameters.extra_build_arg }}"
+        # please check ONNXRUNTIME_CUDA_VERSION in tools/ci_build/github/linux/build_linux_python_package.sh
+        arguments: -i onnxruntimecuda118xtrt86build${{ parameters.arch }} -d "GPU" -c ${{ parameters.cmake_build_type }} -x "${{ parameters.extra_build_arg } -u "12.2"
 
     - task: PublishBuildArtifacts@1
       displayName: 'Publish Artifact: ONNXRuntime python wheel'
diff --git a/tools/ci_build/github/linux/build_linux_arm64_python_package.sh b/tools/ci_build/github/linux/build_linux_python_package.sh
similarity index 86%
rename from tools/ci_build/github/linux/build_linux_arm64_python_package.sh
rename to tools/ci_build/github/linux/build_linux_python_package.sh
index 516f320cd64c4..e3d5acb30a8f3 100755
--- a/tools/ci_build/github/linux/build_linux_arm64_python_package.sh
+++ b/tools/ci_build/github/linux/build_linux_python_package.sh
@@ -10,7 +10,7 @@ EXTRA_ARG=""
 # Put 3.8 at the last because Ubuntu 20.04 use python 3.8 and we will upload the intermediate build files of this 
 # config to Azure DevOps Artifacts and download them to a Ubuntu 20.04 machine to run the tests.
 PYTHON_EXES=("/opt/python/cp39-cp39/bin/python3.9" "/opt/python/cp310-cp310/bin/python3.10" "/opt/python/cp311-cp311/bin/python3.11" "/opt/python/cp38-cp38/bin/python3.8")
-while getopts "d:p:x:c:" parameter_Option
+while getopts "d:p:x:c:u:" parameter_Option
 do case "${parameter_Option}"
 in
 #GPU or CPU.
@@ -59,10 +59,9 @@ if [ "$ARCH" == "x86_64" ]; then
     BUILD_ARGS+=("--enable_onnx_tests")
 fi
 
-if [ "$BUILD_DEVICE" == "GPU" ]; then
+if [ "$BUILD_DEVICE" == "GPU" ] && [ "$CUDA_VERSION" != ""]; then
     #Enable CUDA and TRT EPs.
-    ONNXRUNTIME_CUDA_VERSION="11.8"
-    BUILD_ARGS+=("--nvcc_threads=1" "--use_cuda" "--use_tensorrt" "--cuda_version=$ONNXRUNTIME_CUDA_VERSION" "--tensorrt_home=/usr" "--cuda_home=/usr/local/cuda-$ONNXRUNTIME_CUDA_VERSION" "--cudnn_home=/usr/local/cuda-$ONNXRUNTIME_CUDA_VERSION" "--cmake_extra_defines" "CMAKE_CUDA_ARCHITECTURES=52;60;61;70;75;80")
+    BUILD_ARGS+=("--nvcc_threads=1" "--use_cuda" "--use_tensorrt" "--cuda_version=$CUDA_VERSION" "--tensorrt_home=/usr" "--cuda_home=/usr/local/cuda-$CUDA_VERSION" "--cudnn_home=/usr/local/cuda-$CUDA_VERSION" "--cmake_extra_defines" "CMAKE_CUDA_ARCHITECTURES=52;60;61;70;75;80")
 fi
 
 export CFLAGS
diff --git a/tools/ci_build/github/linux/build_linux_python_package_cuda12.sh b/tools/ci_build/github/linux/build_linux_python_package_cuda12.sh
deleted file mode 100755
index a04e72ed1a348..0000000000000
--- a/tools/ci_build/github/linux/build_linux_python_package_cuda12.sh
+++ /dev/null
@@ -1,78 +0,0 @@
-#!/bin/bash
-set -e -x
-
-# This script invokes build.py
-
-mkdir -p /build/dist
-
-EXTRA_ARG=""
-
-# Put 3.8 at the last because Ubuntu 20.04 use python 3.8 and we will upload the intermediate build files of this 
-# config to Azure DevOps Artifacts and download them to a Ubuntu 20.04 machine to run the tests.
-PYTHON_EXES=("/opt/python/cp39-cp39/bin/python3.9" "/opt/python/cp310-cp310/bin/python3.10" "/opt/python/cp311-cp311/bin/python3.11" "/opt/python/cp38-cp38/bin/python3.8")
-while getopts "d:p:x:c:" parameter_Option
-do case "${parameter_Option}"
-in
-#GPU or CPU.
-d) BUILD_DEVICE=${OPTARG};;
-p) PYTHON_EXES=(${OPTARG});;
-x) EXTRA_ARG=(${OPTARG});;
-c) BUILD_CONFIG=${OPTARG};;
-esac
-done
-
-BUILD_ARGS=("--build_dir" "/build" "--config" "$BUILD_CONFIG" "--update" "--build" "--skip_submodule_sync" "--parallel" "--build_wheel")
-
-if [ "$BUILD_CONFIG" == "Debug" ]; then
-    CFLAGS="-ggdb3"
-    CXXFLAGS="-ggdb3"
-else
-    CFLAGS="-Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -fstack-protector-strong -O3 -pipe -Wl,--strip-all"
-    CXXFLAGS="-Wp,-D_FORTIFY_SOURCE=2 -Wp,-D_GLIBCXX_ASSERTIONS -fstack-protector-strong -O3 -pipe -Wl,--strip-all"
-    BUILD_ARGS+=("--enable_lto")
-fi
-
-# Depending on how the compiler has been configured when it was built, sometimes "gcc -dumpversion" shows the full version.
-GCC_VERSION=$(gcc -dumpversion | cut -d . -f 1)
-#-fstack-clash-protection prevents attacks based on an overlapping heap and stack.
-if [ "$GCC_VERSION" -ge 8 ]; then
-    CFLAGS="$CFLAGS -fstack-clash-protection"
-    CXXFLAGS="$CXXFLAGS -fstack-clash-protection"
-fi
-
-ARCH=$(uname -m)
-
-if [ "$ARCH" == "x86_64" ] && [ "$GCC_VERSION" -ge 9 ]; then
-    CFLAGS="$CFLAGS -fcf-protection"
-    CXXFLAGS="$CXXFLAGS -fcf-protection"
-fi
-
-echo "EXTRA_ARG:"
-echo $EXTRA_ARG
-
-if [ "$EXTRA_ARG" != "" ]; then
-    BUILD_ARGS+=("$EXTRA_ARG")
-fi
-
-if [ "$ARCH" == "x86_64" ]; then
-    #ARM build machines do not have the test data yet.
-    BUILD_ARGS+=("--enable_onnx_tests")
-fi
-
-if [ "$BUILD_DEVICE" == "GPU" ]; then
-    #Enable CUDA and TRT EPs.
-    ONNXRUNTIME_CUDA_VERSION="12.2"
-    BUILD_ARGS+=("--nvcc_threads=1" "--use_cuda" "--use_tensorrt" "--cuda_version=$ONNXRUNTIME_CUDA_VERSION" "--tensorrt_home=/usr" "--cuda_home=/usr/local/cuda-$ONNXRUNTIME_CUDA_VERSION" "--cudnn_home=/usr/local/cuda-$ONNXRUNTIME_CUDA_VERSION" "--cmake_extra_defines" "CMAKE_CUDA_ARCHITECTURES=52;60;61;70;75;80")
-fi
-
-export CFLAGS
-export CXXFLAGS
-for PYTHON_EXE in "${PYTHON_EXES[@]}"
-do
-  rm -rf /build/$BUILD_CONFIG
-  ${PYTHON_EXE} /onnxruntime_src/tools/ci_build/build.py "${BUILD_ARGS[@]}"
-
-  cp /build/$BUILD_CONFIG/dist/*.whl /build/dist
-done
-
-which ccache && ccache -sv && ccache -z
diff --git a/tools/ci_build/github/linux/run_python_dockerbuild.sh b/tools/ci_build/github/linux/run_python_dockerbuild.sh
index 0f36550819c4e..c24c390c24afc 100755
--- a/tools/ci_build/github/linux/run_python_dockerbuild.sh
+++ b/tools/ci_build/github/linux/run_python_dockerbuild.sh
@@ -2,13 +2,14 @@
 set -e -x
 BUILD_CONFIG="Release"
 
-while getopts "i:d:x:c:" parameter_Option
+while getopts "i:d:x:c:u:" parameter_Option
 do case "${parameter_Option}"
 in
 i) DOCKER_IMAGE=${OPTARG};;
 d) DEVICE=${OPTARG};;
 x) BUILD_EXTR_PAR=${OPTARG};;
 c) BUILD_CONFIG=${OPTARG};;
+u) CUDA_VERSION=${OPTARG};;
 esac
 done
 
@@ -23,7 +24,7 @@ docker run --rm \
     -e NIGHTLY_BUILD \
     -e BUILD_BUILDNUMBER \
     $ADDITIONAL_DOCKER_PARAMETER \
-    $DOCKER_IMAGE tools/ci_build/github/linux/build_linux_python_package_cuda12.sh -d $DEVICE -c $BUILD_CONFIG -x $BUILD_EXTR_PAR
+    $DOCKER_IMAGE tools/ci_build/github/linux/build_linux_arm64_python_package.sh -d $DEVICE -c $BUILD_CONFIG -x $BUILD_EXTR_PAR -u $CUDA_VERSION
 
 sudo rm -rf $BUILD_BINARIESDIRECTORY/$BUILD_CONFIG/onnxruntime $BUILD_BINARIESDIRECTORY/$BUILD_CONFIG/pybind11 \
     $BUILD_BINARIESDIRECTORY/$BUILD_CONFIG/models $BUILD_BINARIESDIRECTORY/$BUILD_CONFIG/_deps \