Remove bigdl-nano and add ipex into inference-cpu image (#9452)

* remove bigdl-nano and add ipex into inference-cpu image

* remove bigdl-nano in docker

* remove bigdl-nano in docker

docker/llm/inference/cpu/docker/Dockerfile

@@ -25,7 +25,6 @@ RUN env DEBIAN_FRONTEND=noninteractive apt-get update && \
     pip install --upgrade requests argparse urllib3 && \
     pip3 install --no-cache-dir --upgrade torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu && \
     pip install --pre --upgrade bigdl-llm[all] && \
-    pip install --pre --upgrade bigdl-nano && \
 # Download bigdl-llm-tutorial
     cd /llm && \
     pip install --upgrade jupyterlab && \

docker/llm/serving/cpu/docker/entrypoint.sh

@@ -171,8 +171,6 @@ else
   else
     # Logic for non-controller(worker) mode
     worker_address="http://$worker_host:$worker_port"
-    # Apply optimizations from bigdl-nano
-    source bigdl-nano-init -t
     # First check if user have set OMP_NUM_THREADS by themselves
     if [[ -n "${omp_num_threads}" ]]; then
       echo "Setting OMP_NUM_THREADS to its original value: $omp_num_threads"