fix PR review comments

Signed-off-by: Srikanth Ramakrishna <srikanth.ramakrishna@intel.com>

Author: sramakintel

pytorch/Dockerfile

@@ -184,16 +184,16 @@ RUN apt-get update && \
 
 RUN rm -rf /etc/apt/sources.list.d/intel-gpu-jammy.list /etc/apt/sources.list.d/oneAPI.list
 
-FROM ipex-xpu-base AS ipex-xpu-base-req
+ENV LD_LIBRARY_PATH=/opt/intel/oneapi/redist/lib:$LD_LIBRARY_PATH
+
+FROM ipex-xpu-base AS ipex-xpu-base-wheels
 
 WORKDIR /
 COPY xpu-requirements.txt .
 
 RUN python -m pip install --no-cache-dir -r xpu-requirements.txt && \
     rm -rf xpu-requirements.txt
 
-ENV LD_LIBRARY_PATH=/opt/intel/oneapi/redist/lib:$LD_LIBRARY_PATH
-
 FROM ipex-xpu-base AS ipex-xpu-jupyter
 
 WORKDIR /jupyter
@@ -258,13 +258,15 @@ COPY serving/torchserve-requirements.txt .
 COPY requirements.txt .
 
 RUN python -m pip install --no-cache-dir -r requirements.txt && \
-    python -m pip install --no-cache-dir -r torchserve-requirements.txt
+    python -m pip install --no-cache-dir -r torchserve-requirements.txt && \
+    rm -rf requirements.txt torchserve-requirements.txt
 
 FROM torchserve-base AS compile-xpu
 
 COPY serving/torchserve-xpu-requirements.txt .
 
-RUN python -m pip install --no-cache-dir -r torchserve-xpu-requirements.txt
+RUN python -m pip install --no-cache-dir -r torchserve-xpu-requirements.txt && \
+    rm -rf torchserve-xpu-requirements.txt
 
 FROM torchserve-base AS torchserve-cpu
 

pytorch/docker-compose.yaml

@@ -122,7 +122,7 @@ services:
         org.opencontainers.base.name: "intel/python:3.10-core"
         org.opencontainers.image.title: "Intel® Extension for PyTorch XPU Base Image"
         org.opencontainers.image.version: ${IPEX_VERSION:-2.1.40}-xpu-${PACKAGE_OPTION:-pip}-base
-      target: ipex-xpu-base-req
+      target: ipex-xpu-base-wheels
     command: >
       python -c "import torch;print(torch.device('xpu'));import
       intel_extension_for_pytorch as
@@ -190,6 +190,7 @@ services:
       labels:
         dependency.apt.numactl: true
         dependency.apt.openjdk-17-jdk: true
+        dependency.apt.xpu-smi: ${XPU_SMI_VERSION:-1.2.38}
         dependency.idp: false
         dependency.python.pip: serving/torchserve-xpu-requirements.txt
         docs: serving

pytorch/serving/README.md

@@ -12,7 +12,7 @@ The [Torchserve Model Archiver](https://github.com/pytorch/serve/blob/master/mod
 
 Follow the instructions found in the link above depending on whether you are intending to archive a model or a workflow. Use the provided container rather than installing the archiver with the example command below:
 
-#### For CPU
+#### Create a Model Archive for CPU device
 
 ```bash
 curl -O https://download.pytorch.org/models/squeezenet1_1-b8a52dc0.pth
@@ -29,7 +29,7 @@ docker run --rm -it \
            --export-path /home/model-server
 ```
 
-### For XPU
+### Create a Model Archive for XPU device
 
 Use a squeezenet model [optimized](./model-store/ipex_squeezenet.py) for XPU using Intel® Extension for PyTorch*.
 
@@ -52,7 +52,7 @@ docker run --rm -it \
 
 Test Torchserve with the new archived model. The example below is for the squeezenet model.
 
-#### For CPU
+#### Run Torchserve for CPU device
 
 ```bash
 # Assuming that the above pre-archived model is in the current working directory
@@ -63,8 +63,7 @@ docker run -d --rm --name server \
           intel/intel-optimized-pytorch:2.4.0-serving-cpu
 ```
 
-#### For XPU
-
+#### Run Torchserve for XPU device
 ```bash
 # Assuming that the above pre-archived model is in the current working directory
 docker run -d --rm --name server \

pytorch/serving/config-xpu.properties

@@ -1,4 +1,3 @@
-
 inference_address=http://0.0.0.0:8080
 management_address=http://0.0.0.0:8081
 metrics_address=http://0.0.0.0:8082

pytorch/serving/tests.yaml

@@ -12,7 +12,7 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-ipex-serving-model-archive:
+ipex-serving-cpu-model-archive:
   img: ${REGISTRY}/${REPO}:b-${GITHUB_RUN_NUMBER:-0}-${BASE_IMAGE_NAME:-ubuntu}-${BASE_IMAGE_TAG:-22.04}-py${PYTHON_VERSION:-3.10}-torchserve-cpu
   cmd: /home/model-server/model-archive/mar-test.sh cpu
   entrypoint: /bin/bash
@@ -35,7 +35,7 @@ ipex-serving-xpu-model-archive:
     dst: /home/model-server/model-store
   user: root
   workdir: /home/model-server/model-archive
-ipex-serving-workflow-archive:
+ipex-serving-cpu-workflow-archive:
   img: ${REGISTRY}/${REPO}:b-${GITHUB_RUN_NUMBER:-0}-${BASE_IMAGE_NAME:-ubuntu}-${BASE_IMAGE_TAG:-22.04}-py${PYTHON_VERSION:-3.10}-torchserve-cpu
   cmd: /home/model-server/model-archive/war-test.sh
   entrypoint: /bin/bash
@@ -46,7 +46,7 @@ ipex-serving-workflow-archive:
     dst: /home/model-server/wf-store
   user: root
   workdir: /home/model-server/model-archive
-ipex-serving-rest-workflow:
+ipex-serving-cpu-rest-workflow:
   img: ${REGISTRY}/${REPO}:b-${GITHUB_RUN_NUMBER:-0}-${BASE_IMAGE_NAME:-ubuntu}-${BASE_IMAGE_TAG:-22.04}-py${PYTHON_VERSION:-3.10}-torchserve-cpu
   cmd: bash /home/model-server/wf-store/rest-test.sh
   serving: True
@@ -72,7 +72,7 @@ ipex-serving-xpu-rest-workflow:
     ENABLE_TORCH_PROFILER: 'true'
   shm_size: 1g
   workdir: /home/model-server/wf-store
-ipex-serving-rest-inference:
+ipex-serving-cpu-rest-inference:
   img: ${REGISTRY}/${REPO}:b-${GITHUB_RUN_NUMBER:-0}-${BASE_IMAGE_NAME:-ubuntu}-${BASE_IMAGE_TAG:-22.04}-py${PYTHON_VERSION:-3.10}-torchserve-cpu
   cmd: bash /home/model-server/model-store/rest-test.sh
   serving: True
@@ -85,7 +85,7 @@ ipex-serving-rest-inference:
     ENABLE_TORCH_PROFILER: 'true'
   shm_size: 1g
   workdir: /home/model-server/model-store
-ipex-serving-grpc-inference:
+ipex-serving-cpu-grpc-inference:
   img: ${REGISTRY}/${REPO}:b-${GITHUB_RUN_NUMBER:-0}-${BASE_IMAGE_NAME:-ubuntu}-${BASE_IMAGE_TAG:-22.04}-py${PYTHON_VERSION:-3.10}-torchserve-cpu
   cmd: bash /home/model-server/model-store/grpc-test.sh
   serving: True