diff --git a/.github/scripts/build-linux-release-artifacts.sh b/.github/scripts/build-linux-release-artifacts.sh
index 41b7d4a6d5..ef1113f8d4 100644
--- a/.github/scripts/build-linux-release-artifacts.sh
+++ b/.github/scripts/build-linux-release-artifacts.sh
@@ -6,13 +6,20 @@ export CXX=g++
 CURRENT_DIR=`pwd`
 
 if [[ "${PYTHON_VERSION}" == "3.8" ]]; then
+  PY_SINGLE_BUILD_CODE=cp38
   PY_BUILD_CODE=cp38-cp38
 elif [[ "${PYTHON_VERSION}" == "3.9" ]]; then
+  PY_SINGLE_BUILD_CODE=cp39
   PY_BUILD_CODE=cp39-cp39
 elif [[ "${PYTHON_VERSION}" == "3.10" ]]; then
+  PY_SINGLE_BUILD_CODE=cp310
   PY_BUILD_CODE=cp310-cp310
 elif [[ "${PYTHON_VERSION}" == "3.11" ]]; then
-  PY_BUILD_CODE=cp38-cp38
+  PY_SINGLE_BUILD_CODE=cp311
+  PY_BUILD_CODE=cp311-cp311
+elif [[ "${PYTHON_VERSION}" == "3.12" ]]; then
+  PY_SINGLE_BUILD_CODE=cp312
+  PY_BUILD_CODE=cp312-cp312
 else
    echo "python version: ${PYTHON_VERSION} is not supported"
    exit
@@ -33,7 +40,9 @@ wget -q https://developer.nvidia.com/downloads/compute/machine-learning/tensorrt
 && tar -xvf TensorRT-10.0.1.6.Linux.x86_64-gnu.cuda-12.4.tar \
 && rm TensorRT-10.0.1.6.Linux.x86_64-gnu.cuda-12.4.tar
 
-TENSERRT_DIR=${CURRENT_DIR}/TensorRT-10.0.1.6
+TENSERRT_DIR=${CURRENT_DIR}/TensorRT-10.0.1.6/
+
+python -m pip install ${TENSERRT_DIR}/python/tensorrt-${TRT_VERSION}-${PY_SINGLE_BUILD_CODE}-none-linux_x86_64.whl
 
 SITE_PKG_DIR=`python -c  'import sysconfig; print(sysconfig.get_paths()["purelib"])'`
 
diff --git a/.github/workflows/linux-release-artifacts.yml b/.github/workflows/linux-release-artifacts.yml
index de7b869d4a..378fe2c87c 100644
--- a/.github/workflows/linux-release-artifacts.yml
+++ b/.github/workflows/linux-release-artifacts.yml
@@ -77,7 +77,9 @@ jobs:
     container:
       image: ${{ matrix.container_image }}
       options: ${{ matrix.gpu_arch_type == 'cuda' && '--gpus all' || ' ' }}
+    # TO DO: ask Naren, I am not able to skip the release build if it is not cuda12.1 on the job level
     # if: ${{ env.CU_VERSION == '12.1' }}
+
     # If a build is taking longer than 60 minutes on these runners we need
     # to have a conversation
     timeout-minutes: 60
@@ -127,12 +129,12 @@ jobs:
         with:
           name: ${{ env.ARTIFACT_NAME }}
           path: /opt/torch-tensorrt-builds/
-      # - name: Install torch and torch-tensorrt
-      #   if: ${{ inputs.pre-script != '' }}
-      #   uses: ./test-infra/.github/actions/run-script-with-cache
-      #   with:
-      #     repository: ${{ inputs.repository  }}
-      #     script: .github/scripts/install-torch-tensorrt.sh
+      - name: Install torch and torch-tensorrt
+        if: ${{ inputs.pre-script != '' }}
+        uses: ./test-infra/.github/actions/run-script-with-cache
+        with:
+          repository: ${{ inputs.repository  }}
+          script: .github/scripts/install-torch-tensorrt.sh
       - name: Pack script
         continue-on-error: ${{ inputs.continue-on-error }}
         working-directory: ${{ inputs.repository }}
@@ -142,7 +144,6 @@ jobs:
           set -euxo pipefail
           # shellcheck disable=SC2086
           source "${BUILD_ENV_FILE}"
-          echo "lan added"
           {
             echo "${SCRIPT}";
           } > "user_script"
diff --git a/py/ci/soname_excludes.params b/py/ci/soname_excludes.params
index c413172647..7a16c5bf3d 100644
--- a/py/ci/soname_excludes.params
+++ b/py/ci/soname_excludes.params
@@ -42,4 +42,6 @@
 --exclude libnvinfer_plugin.so.10
 --exclude libnvinfer-ea523553.so.10.0.1
 --exclude libcudart-9335f6a2.so.12.1.105
---exclude libnvinfer_plugin-24528888.so.10.0.1
\ No newline at end of file
+--exclude libnvinfer_plugin-24528888.so.10.0.1
+
+--exclude libcudart-d0da41ae.so.11.8.89
\ No newline at end of file