diff --git a/.github/workflows/cpp-graph-test.yml b/.github/workflows/cpp-graph-test.yml
index fe60897cecf..a63e2b787c5 100644
--- a/.github/workflows/cpp-graph-test.yml
+++ b/.github/workflows/cpp-graph-test.yml
@@ -37,6 +37,7 @@ jobs:
         uses: actions/checkout@v3
         with:
           submodules: "recursive"
+          fetch-tags: true
 
       - name: Env build
         run: |
diff --git a/.github/workflows/deploy-test.yml b/.github/workflows/deploy-test.yml
index 758cfa0dc4b..318b1fda042 100644
--- a/.github/workflows/deploy-test.yml
+++ b/.github/workflows/deploy-test.yml
@@ -65,6 +65,7 @@ jobs:
         uses: actions/checkout@v3
         with:
           submodules: "recursive"
+          fetch-tags: true
       # We need this because GitHub needs to clone the branch to pipeline
       - name: Docker Build
         run: |
diff --git a/.github/workflows/llm-test.yml b/.github/workflows/llm-test.yml
index 2ca234c4760..450fa455305 100644
--- a/.github/workflows/llm-test.yml
+++ b/.github/workflows/llm-test.yml
@@ -41,6 +41,7 @@ jobs:
         uses: actions/checkout@v3
         with:
           submodules: "recursive"
+          fetch-tags: true
 
       - name: Env build
         run: |
diff --git a/.github/workflows/optimize-test.yml b/.github/workflows/optimize-test.yml
index 8fe27a3f31e..8811dc91039 100644
--- a/.github/workflows/optimize-test.yml
+++ b/.github/workflows/optimize-test.yml
@@ -62,6 +62,7 @@ jobs:
         uses: actions/checkout@v3
         with:
           submodules: "recursive"
+          fetch-tags: true
     # We need this because GitHub needs to clone the branch to pipeline
       - name: Docker Build
         run: |
diff --git a/.github/workflows/script/SparseLibCI/run_sparse_lib.sh b/.github/workflows/script/SparseLibCI/run_sparse_lib.sh
index 899216ffb7a..d6001a1a458 100644
--- a/.github/workflows/script/SparseLibCI/run_sparse_lib.sh
+++ b/.github/workflows/script/SparseLibCI/run_sparse_lib.sh
@@ -1,6 +1,4 @@
 #!/bin/bash
-export COVERAGE_RCFILE="/intel-extension-for-transformers/.github/workflows/script/unitTest/coverage/.coveragerc"
-
 output_log_dir="/intel-extension-for-transformers/benchmark_log"
 WORKING_DIR="/intel-extension-for-transformers/intel_extension_for_transformers/llm/runtime/deprecated"
 mkdir ${output_log_dir}
diff --git a/.github/workflows/script/formatScan/pylint.sh b/.github/workflows/script/formatScan/pylint.sh
index 9dd77b50431..1aa9e9bec3b 100644
--- a/.github/workflows/script/formatScan/pylint.sh
+++ b/.github/workflows/script/formatScan/pylint.sh
@@ -12,7 +12,6 @@ $BOLD_YELLOW && echo "---------------- run python setup.py sdist bdist_wheel ---
 export PYTHONPATH=`pwd`
 #$BOLD_YELLOW && echo "---------------- pip install binary -------------" && $RESET
 #pip install dist/intel_extension_for_transformers*.whl
-
 pip list
 
 cd /intel-extension-for-transformers/intel_extension_for_transformers/neural_chat/
diff --git a/.github/workflows/script/unitTest/coverage/.coveragerc b/.github/workflows/script/unitTest/coverage/.coveragerc
deleted file mode 100644
index c7c4fc7a42b..00000000000
--- a/.github/workflows/script/unitTest/coverage/.coveragerc
+++ /dev/null
@@ -1,8 +0,0 @@
-[run]
-branch = True
-
-[report]
-omit =
- */**/fake*yaml
-exclude_lines =
- pragma: no cover
diff --git a/.github/workflows/script/unitTest/coverage/.engine-coveragerc b/.github/workflows/script/unitTest/coverage/.engine-coveragerc
new file mode 100644
index 00000000000..a60803774a9
--- /dev/null
+++ b/.github/workflows/script/unitTest/coverage/.engine-coveragerc
@@ -0,0 +1,14 @@
+[run]
+branch = True
+
+[report]
+omit =
+ */**/fake*yaml
+ */intel_extension_for_transformers/neural_chat/**
+ */intel_extension_for_transformers/transformers/**
+ */intel_extension_for_transformers/llm/evaluation/**
+ */intel_extension_for_transformers/llm/finetuning/**
+ */intel_extension_for_transformers/llm/inference/**
+ */intel_extension_for_transformers/llm/quantization/**
+exclude_lines =
+ pragma: no cover
diff --git a/.github/workflows/script/unitTest/coverage/.neural-chat-coveragerc b/.github/workflows/script/unitTest/coverage/.neural-chat-coveragerc
new file mode 100644
index 00000000000..30050b5d48b
--- /dev/null
+++ b/.github/workflows/script/unitTest/coverage/.neural-chat-coveragerc
@@ -0,0 +1,24 @@
+[run]
+branch = True
+
+[report]
+omit =
+ */**/fake*yaml
+ */**/fake.py
+ */intel_extension_for_transformers/llm/amp/**
+ */intel_extension_for_transformers/llm/evaluation/**
+ */intel_extension_for_transformers/llm/quantization/**
+ */intel_extension_for_transformers/llm/library/**
+ */intel_extension_for_transformers/llm/operator/**
+ */intel_extension_for_transformers/llm/runtime/**
+ */intel_extension_for_transformers/transformers/**
+exclude_lines =
+ pragma: no cover
+ raise NotImplementedError
+ raise TypeError
+ if self.device == "gpu":
+ if device == "gpu":
+ except ImportError:
+ except Exception as e:
+ onnx_version < ONNX18_VERSION
+ onnx_version >= ONNX18_VERSION
diff --git a/.github/workflows/script/unitTest/coverage/calc_coverage.sh b/.github/workflows/script/unitTest/coverage/calc_coverage.sh
index dea595b3d2a..08e3a3bece1 100644
--- a/.github/workflows/script/unitTest/coverage/calc_coverage.sh
+++ b/.github/workflows/script/unitTest/coverage/calc_coverage.sh
@@ -87,18 +87,22 @@ function check_coverage_status() {
             $BOLD_RED && echo "Unit Test failed with ${item} coverage decrease ${decrease}%" && $RESET
         done
         $BOLD_RED && echo "compare coverage to give detail info" && $RESET
-        bash -x compare_coverage.sh ${coverage_compare} ${coverage_log_pr}/coverage.log ${coverage_log_base}/coverage.log "FAILED" ${coverage_PR_lines_rate} ${coverage_base_lines_rate} ${coverage_PR_branches_rate} ${coverage_base_branches_rate}
+        bash compare_coverage.sh ${coverage_compare} ${coverage_log_pr}/coverage.log ${coverage_log_base}/coverage.log "FAILED" ${coverage_PR_lines_rate} ${coverage_base_lines_rate} ${coverage_PR_branches_rate} ${coverage_base_branches_rate}
         exit 1
     else
         $BOLD_GREEN && echo "Unit Test success with coverage lines: ${coverage_PR_lines_rate}%, branches: ${coverage_PR_branches_rate}%" && $RESET
         $BOLD_GREEN && echo "compare coverage to give detail info" && $RESET
-        bash -x compare_coverage.sh ${coverage_compare} ${coverage_log_pr}/coverage.log ${coverage_log_base}/coverage.log "SUCCESS" ${coverage_PR_lines_rate} ${coverage_base_lines_rate} ${coverage_PR_branches_rate} ${coverage_base_branches_rate}
+        bash compare_coverage.sh ${coverage_compare} ${coverage_log_pr}/coverage.log ${coverage_log_base}/coverage.log "SUCCESS" ${coverage_PR_lines_rate} ${coverage_base_lines_rate} ${coverage_PR_branches_rate} ${coverage_base_branches_rate}
     fi
 }
 
 function main() {
     compare_coverage
     check_coverage_status
+    $BOLD_BLUE && echo "PR lines coverage: $lines_PR_covered/$lines_PR_valid ($coverage_PR_lines_rate%)" && $RESET
+    $BOLD_BLUE && echo "PR branches coverage: $branches_PR_covered/$branches_PR_valid ($coverage_PR_branches_rate%)" && $RESET
+    $BOLD_BLUE && echo "BASE lines coverage: $lines_base_covered/$lines_base_valid ($coverage_base_lines_rate%)" && $RESET
+    $BOLD_BLUE && echo "BASE branches coverage: $branches_base_covered/$branches_base_valid ($coverage_base_branches_rate%)" && $RESET
 }
 
 main
diff --git a/.github/workflows/script/unitTest/run_unit_test_engine.sh b/.github/workflows/script/unitTest/run_unit_test_engine.sh
index 0ee166c565a..543df411c19 100644
--- a/.github/workflows/script/unitTest/run_unit_test_engine.sh
+++ b/.github/workflows/script/unitTest/run_unit_test_engine.sh
@@ -1,6 +1,6 @@
 #!/bin/bash
 source /intel-extension-for-transformers/.github/workflows/script/change_color.sh
-export COVERAGE_RCFILE="/intel-extension-for-transformers/.github/workflows/script/unitTest/coverage/.coveragerc"
+export COVERAGE_RCFILE="/intel-extension-for-transformers/.github/workflows/script/unitTest/coverage/.engine-coveragerc"
 LOG_DIR=/log_dir
 mkdir -p ${LOG_DIR}
 WORKING_DIR="/intel-extension-for-transformers/intel_extension_for_transformers/llm/runtime/deprecated"
@@ -86,7 +86,13 @@ function gtest() {
 
 function main() {
     bash /intel-extension-for-transformers/.github/workflows/script/unitTest/env_setup.sh
-    echo "test on ${test_name}"
+    cd ${WORKING_DIR}/test/pytest || exit 1
+    if [ -f "requirements.txt" ]; then
+        python -m pip install --default-timeout=100 -r requirements.txt
+        pip list
+    else
+        echo "Not found requirements.txt file."
+    fi
     if [[ $test_name == "PR-test" ]]; then
         pytest "${LOG_DIR}/coverage_pr"
         gtest
diff --git a/.github/workflows/script/unitTest/run_unit_test_neuralchat.sh b/.github/workflows/script/unitTest/run_unit_test_neuralchat.sh
new file mode 100644
index 00000000000..a75be157a09
--- /dev/null
+++ b/.github/workflows/script/unitTest/run_unit_test_neuralchat.sh
@@ -0,0 +1,71 @@
+#!/bin/bash
+source /intel-extension-for-transformers/.github/workflows/script/change_color.sh
+export COVERAGE_RCFILE="/intel-extension-for-transformers/.github/workflows/script/unitTest/coverage/.neural-chat-coveragerc"
+LOG_DIR=/log_dir
+mkdir -p ${LOG_DIR}
+WORKING_DIR="/intel-extension-for-transformers/intel_extension_for_transformers/neural_chat/tests"
+# get parameters
+PATTERN='[-a-zA-Z0-9_]*='
+PERF_STABLE_CHECK=true
+
+for i in "$@"; do
+    case $i in
+        --test_name=*)
+            test_name=`echo $i | sed "s/${PATTERN}//"`;;
+        *)
+            echo "Parameter $i not recognized."; exit 1;;
+    esac
+done
+
+function pytest() {
+    local coverage_log_dir=$1
+    mkdir -p ${coverage_log_dir}
+
+    cd ${WORKING_DIR} || exit 1
+    JOB_NAME=unit_test
+    ut_log_name=${LOG_DIR}/${JOB_NAME}.log
+    export GLOG_minloglevel=2
+
+    itrex_path=$(python -c 'import intel_extension_for_transformers; import os; print(os.path.dirname(intel_extension_for_transformers.__file__))')
+    find . -name "test*.py" | sed 's,\.\/,coverage run --source='"${itrex_path}"' --append ,g' | sed 's/$/ --verbose/' >run.sh
+    coverage erase
+
+    # run UT
+    $BOLD_YELLOW && echo "cat run.sh..." && $RESET
+    cat run.sh | tee ${ut_log_name}
+    $BOLD_YELLOW && echo "------UT start-------" && $RESET
+    bash run.sh 2>&1 | tee -a ${ut_log_name}
+    $BOLD_YELLOW && echo "------UT end -------" && $RESET
+
+    # run coverage report
+    coverage report -m --rcfile=${COVERAGE_RCFILE} | tee ${coverage_log_dir}/coverage.log
+    coverage html -d ${coverage_log_dir}/htmlcov --rcfile=${COVERAGE_RCFILE}
+    coverage xml -o ${coverage_log_dir}/coverage.xml --rcfile=${COVERAGE_RCFILE}
+
+    # check UT status
+    if [ $(grep -c "FAILED" ${ut_log_name}) != 0 ] || [ $(grep -c "OK" ${ut_log_name}) == 0 ]; then
+        $BOLD_RED && echo "Find errors in UT test, please check the output..." && $RESET
+        exit 1
+    fi
+    $BOLD_GREEN && echo "UT finished successfully! " && $RESET
+}
+
+function main() {
+    bash /intel-extension-for-transformers/.github/workflows/script/unitTest/env_setup.sh
+    wget http://nz2.archive.ubuntu.com/ubuntu/pool/main/o/openssl/libssl1.1_1.1.1f-1ubuntu2.19_amd64.deb
+    dpkg -i libssl1.1_1.1.1f-1ubuntu2.19_amd64.deb
+    cd ${WORKING_DIR} || exit 1
+    if [ -f "requirements.txt" ]; then
+        python -m pip install --default-timeout=100 -r requirements.txt
+        pip list
+    else
+        echo "Not found requirements.txt file."
+    fi
+    echo "test on ${test_name}"
+    if [[ $test_name == "PR-test" ]]; then
+        pytest "${LOG_DIR}/coverage_pr"
+    elif [[ $test_name == "baseline" ]]; then
+        pytest "${LOG_DIR}/coverage_base"  
+    fi
+}
+main
diff --git a/.github/workflows/unit-test-engine.yml b/.github/workflows/unit-test-engine.yml
index 3d75c65142a..f21efb4ed5c 100644
--- a/.github/workflows/unit-test-engine.yml
+++ b/.github/workflows/unit-test-engine.yml
@@ -56,6 +56,7 @@ jobs:
         with:
           submodules: "recursive"
           ref: ${{ matrix.test_branch }}
+          fetch-tags: true
 
       - name: Docker Build
         run: |
diff --git a/.github/workflows/unit-test-kernel.yml b/.github/workflows/unit-test-kernel.yml
index 5839731bf34..32a88194214 100644
--- a/.github/workflows/unit-test-kernel.yml
+++ b/.github/workflows/unit-test-kernel.yml
@@ -47,6 +47,7 @@ jobs:
         uses: actions/checkout@v3
         with:
           submodules: "recursive"
+          fetch-tags: true
 
       - name: Docker Build
         run: |
diff --git a/.github/workflows/unit-test-neuralchat.yml b/.github/workflows/unit-test-neuralchat.yml
new file mode 100644
index 00000000000..540031e1d41
--- /dev/null
+++ b/.github/workflows/unit-test-neuralchat.yml
@@ -0,0 +1,141 @@
+name: NeuralChat Unit Test
+
+on:
+  pull_request:
+    branches: [main]
+    paths:
+      - intel_extension_for_transformers/neural_chat/tests/**
+      - .github/workflows/unit-test-neuralchat.yml
+      - .github/workflows/script/unitTest/**
+  workflow_dispatch:
+
+# If there is a new commit, the previous jobs will be canceled
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: true
+
+env:
+  DOCKER_CONFIG_NAME: "commonDockerConfig"
+  REPO_NAME: "intel-extension-for-transformers"
+  REPO_TAG: "py38"
+  DOCKER_FILE_NAME: "devel"
+  CONTAINER_NAME: "utTest"
+  EXTRA_CONTAINER_NAME: "modelTest"
+
+jobs:
+  unit-test:
+    runs-on: [self-hosted, Linux, X64, itrex-node]
+    strategy:
+      matrix:
+        include:
+          - test_branch: ${{ github.ref }}
+            test_name: "PR-test"
+          - test_branch: "main"
+            test_name: "baseline"
+    steps:
+      - name: Docker Clean Up
+        run: |
+          docker ps -a
+          if [[ $(docker ps -a | grep -i '${{ env.CONTAINER_NAME }}'$) ]]; then
+              docker start ${{ env.CONTAINER_NAME }}
+              echo "remove left files through container ..."
+              docker exec ${{ env.CONTAINER_NAME }} bash -c "ls -a /intel-extension-for-transformers && rm -fr /intel-extension-for-transformers/* && rm -fr /intel-extension-for-transformers/.* || true"
+          fi
+          if [[ $(docker ps -a | grep -i '${{ env.EXTRA_CONTAINER_NAME }}'$) ]]; then
+              docker start ${{ env.EXTRA_CONTAINER_NAME }}
+              echo "remove left files through container ..."
+              docker exec ${{ env.EXTRA_CONTAINER_NAME }} bash -c "ls -a /intel-extension-for-transformers && rm -fr /intel-extension-for-transformers/* && rm -fr /intel-extension-for-transformers/.* || true"
+          fi
+
+      - name: Checkout out Repo
+        uses: actions/checkout@v3
+        with:
+          submodules: "recursive"
+          ref: ${{ matrix.test_branch }}
+          fetch-tags: true
+
+      - name: Docker Build
+        run: |
+          docker build -f ${{ github.workspace }}/.github/workflows/docker/${{ env.DOCKER_FILE_NAME }}.dockerfile -t ${{ env.REPO_NAME }}:${{ env.REPO_TAG }} .
+
+      - name: Docker Run
+        run: |
+          if [[ $(docker ps -a | grep -i '${{ env.CONTAINER_NAME }}'$) ]]; then
+            docker stop ${{ env.CONTAINER_NAME }}
+            docker rm -vf ${{ env.CONTAINER_NAME }} || true
+          fi
+          docker run -dit --disable-content-trust --privileged --name=${{ env.CONTAINER_NAME }} -v /dev/shm:/dev/shm \
+          -v ${{ github.workspace }}:/intel-extension-for-transformers \
+          ${{ env.REPO_NAME }}:${{ env.REPO_TAG }}
+
+      - name: Env build
+        run: |
+          docker exec ${{ env.CONTAINER_NAME }} \
+          bash /intel-extension-for-transformers/.github/workflows/script/prepare_env.sh
+
+      - name: Binary build
+        run: |
+          docker exec ${{ env.CONTAINER_NAME }} \
+          bash -c "cd /intel-extension-for-transformers/.github/workflows/script \
+          && bash install_binary.sh \
+          && pip install intel_extension_for_pytorch wget sentencepiece \
+          && pip install transformers>=4.32.0 \
+          && pip install peft torch num2words speechbrain paddlepaddle \
+          && pip install paddlespeech==1.4.1 shortuuid gptcache evaluate \
+          && pip install fschat pydub python-multipart PyPDF2  langchain \
+          && pip install python-docx scikit-learn farm-haystack librosa beautifulsoup4 \
+          && pip install InstructorEmbedding chromadb pydantic fastapi starlette \
+          && pip install yacs uvicorn optimum optimum[habana] \
+          && pip install sentence_transformers unstructured markdown rouge_score \
+          && pip install --upgrade accelerate"
+
+      - name: Run UT
+        run: |
+          docker exec ${{ env.CONTAINER_NAME }} \
+          bash -c "cd /intel-extension-for-transformers/.github/workflows/script/unitTest \
+          && bash run_unit_test_neuralchat.sh --test_name=${{ matrix.test_name }}"
+
+      - name: Collect log
+        if: ${{ !cancelled() }}
+        run: |
+          docker exec ${{ env.CONTAINER_NAME }} \
+          bash -c "cd /intel-extension-for-transformers && \
+                 mv /log_dir . "
+      
+      - name: Publish pipeline artifact
+        uses: actions/upload-artifact@v3
+        if: ${{ !cancelled() }}
+        with:
+          name: "UnitTest${{ matrix.test_name }}"
+          path: ${{ github.workspace }}/log_dir
+           
+           
+  Genreate-Report:
+    runs-on: itrex-node-spell
+    needs: [unit-test]
+    steps:
+      - name: Checkout out Repo
+        uses: actions/checkout@v3
+        with:
+          submodules: "recursive"
+        
+      - name: Download UT PR Log
+        uses: actions/download-artifact@v3
+        with:
+          path: ${{ github.workspace }}/log_dir
+           
+      - name: Display structure of downloaded files
+        run: cd ${{ github.workspace }}/log_dir && ls -R
+        
+      - name: Calculte coverage
+        run: |
+          cd ${{ github.workspace }}/.github/workflows/script/unitTest/coverage
+          /usr/bin/bash calc_coverage.sh ${{ github.workspace }}/log_dir
+
+      - name: Publish pipeline artifact
+        uses: actions/upload-artifact@v3
+        if: ${{ !cancelled() }}
+        with:
+          name: Neural Chat Unit Test
+          path: ${{ github.workspace }}/log_dir
+          retention-days: 5
\ No newline at end of file
diff --git a/.github/workflows/unit-test-optimize.yml b/.github/workflows/unit-test-optimize.yml
index 860c887fbc7..620f249bb7c 100644
--- a/.github/workflows/unit-test-optimize.yml
+++ b/.github/workflows/unit-test-optimize.yml
@@ -56,6 +56,7 @@ jobs:
         with:
           submodules: "recursive"
           ref: ${{ matrix.test_branch }}
+          fetch-tags: true
 
       - name: Docker Build
         run: |
@@ -131,7 +132,7 @@ jobs:
         uses: actions/upload-artifact@v3
         if: ${{ !cancelled() }}
         with:
-          name: Engine Unit Test
+          name: Optimize Unit Test
           path: ${{ github.workspace }}/log_dir
           retention-days: 5
 
diff --git a/intel_extension_for_transformers/__init__.py b/intel_extension_for_transformers/__init__.py
index 5603667ff8f..e1bfd064448 100644
--- a/intel_extension_for_transformers/__init__.py
+++ b/intel_extension_for_transformers/__init__.py
@@ -18,4 +18,4 @@
 try:
     from ._version import __version__  # load _version file generated by setuptools_scm 
 except ModuleNotFoundError:
-    __version__ = "0.0.0"
\ No newline at end of file
+    __version__ = "1.1"
\ No newline at end of file
diff --git a/intel_extension_for_transformers/llm/finetuning/finetuning.py b/intel_extension_for_transformers/llm/finetuning/finetuning.py
index de58b0ff965..7cadf1b01cc 100644
--- a/intel_extension_for_transformers/llm/finetuning/finetuning.py
+++ b/intel_extension_for_transformers/llm/finetuning/finetuning.py
@@ -78,12 +78,12 @@ def __init__(self, finetuning_config: BaseFinetuningConfig):
             else:
                 finetuning_config.finetune_args.device = "cpu"
         if finetuning_config.finetune_args.device == "cpu":
-            finetuning_config.training_args.no_cuda = True
             Arguments = type(finetuning_config.training_args)
             training_args = {
                 k: getattr(finetuning_config.training_args, k) \
                     for k in Arguments.__dataclass_fields__.keys() if Arguments.__dataclass_fields__[k].init
             }
+            training_args["no_cuda"] = True
             self.training_args = Arguments(**training_args)
 
         os.environ["WANDB_DISABLED"] = "true"
diff --git a/intel_extension_for_transformers/neural_chat/assets/audio/pat.wav b/intel_extension_for_transformers/neural_chat/assets/audio/sample.wav
similarity index 100%
rename from intel_extension_for_transformers/neural_chat/assets/audio/pat.wav
rename to intel_extension_for_transformers/neural_chat/assets/audio/sample.wav
diff --git a/intel_extension_for_transformers/neural_chat/assets/docs/sample.xlsx b/intel_extension_for_transformers/neural_chat/assets/docs/sample.xlsx
index 0b1c43c9254..71abab41430 100644
Binary files a/intel_extension_for_transformers/neural_chat/assets/docs/sample.xlsx and b/intel_extension_for_transformers/neural_chat/assets/docs/sample.xlsx differ
diff --git a/intel_extension_for_transformers/neural_chat/cli/cli_commands.py b/intel_extension_for_transformers/neural_chat/cli/cli_commands.py
index 2b6edc9af04..794f388bd35 100644
--- a/intel_extension_for_transformers/neural_chat/cli/cli_commands.py
+++ b/intel_extension_for_transformers/neural_chat/cli/cli_commands.py
@@ -336,4 +336,4 @@ def __call__(self):
     command_register(
         name='neuralchat.{}'.format(com),
         description=info[0],
-        cls='neural_chat.cli.cli_commands.{}'.format(info[1]))
+        cls='intel_extension_for_transformers.neural_chat.cli.cli_commands.{}'.format(info[1]))
diff --git a/intel_extension_for_transformers/neural_chat/config.py b/intel_extension_for_transformers/neural_chat/config.py
index 56223dce62f..bd3a587109c 100644
--- a/intel_extension_for_transformers/neural_chat/config.py
+++ b/intel_extension_for_transformers/neural_chat/config.py
@@ -22,14 +22,6 @@
 from transformers.utils.versions import require_version
 from dataclasses import dataclass
 
-from .pipeline.plugins.audio.asr import AudioSpeechRecognition
-from .pipeline.plugins.audio.asr_chinese import ChineseAudioSpeechRecognition
-from .pipeline.plugins.audio.tts import TextToSpeech
-from .pipeline.plugins.audio.tts_chinese import ChineseTextToSpeech
-from .pipeline.plugins.retrieval.indexing import DocumentIndexing
-from .pipeline.plugins.retrieval import SparseBM25Retriever, ChromaRetriever
-from .pipeline.plugins.retrieval.detector import IntentDetector
-from .pipeline.plugins.security import SafetyChecker
 from .plugins import plugins
 
 from enum import Enum, auto
diff --git a/intel_extension_for_transformers/neural_chat/docs/notebooks/chatbot_on_intel_cpu.ipynb b/intel_extension_for_transformers/neural_chat/docs/notebooks/chatbot_on_intel_cpu.ipynb
index 873619cffe3..ca6c5478c15 100644
--- a/intel_extension_for_transformers/neural_chat/docs/notebooks/chatbot_on_intel_cpu.ipynb
+++ b/intel_extension_for_transformers/neural_chat/docs/notebooks/chatbot_on_intel_cpu.ipynb
@@ -120,7 +120,7 @@
     "from intel_extension_for_transformers.neural_chat import build_chatbot\n",
     "config = PipelineConfig(audio_input=True, audio_output=True)\n",
     "chatbot = build_chatbot(config)\n",
-    "result = chatbot.predict(query=\"./assets/audio/pat.wav\")"
+    "result = chatbot.predict(query=\"./assets/audio/sample.wav\")"
    ]
   },
   {
@@ -302,7 +302,7 @@
     "from neural_chat import VoiceChatClientExecutor\n",
     "executor = VoiceChatClientExecutor()\n",
     "result = executor(\n",
-    "    audio_input_path='./assets/audio/pat.wav',\n",
+    "    audio_input_path='./assets/audio/sample.wav',\n",
     "    audio_output_path='./results.wav',\n",
     "    server_ip=\"127.0.0.1\", # master server ip\n",
     "    port=8000 # master server entry point \n",
@@ -318,7 +318,7 @@
     "import IPython\n",
     "# Play input audio\n",
     "print(\"     Play Input Audio ......\")\n",
-    "IPython.display.display(IPython.display.Audio(\"./assets/audio/pat.wav\"))\n",
+    "IPython.display.display(IPython.display.Audio(\"./assets/audio/sample.wav\"))\n",
     "# Play output audio\n",
     "print(\"     Play Output Audio ......\")\n",
     "IPython.display.display(IPython.display.Audio(\"./assets/audio/welcome.wav\"))\n"
diff --git a/intel_extension_for_transformers/neural_chat/docs/notebooks/chatbot_on_intel_habana_hpu.ipynb b/intel_extension_for_transformers/neural_chat/docs/notebooks/chatbot_on_intel_habana_hpu.ipynb
index 2110f4dd834..efd2f5b5dc3 100644
--- a/intel_extension_for_transformers/neural_chat/docs/notebooks/chatbot_on_intel_habana_hpu.ipynb
+++ b/intel_extension_for_transformers/neural_chat/docs/notebooks/chatbot_on_intel_habana_hpu.ipynb
@@ -141,7 +141,7 @@
     "from intel_extension_for_transformers.neural_chat import build_chatbot\n",
     "config = PipelineConfig(audio_input=True, audio_output=True)\n",
     "chatbot = build_chatbot(config)\n",
-    "result = chatbot.predict(query=\"./assets/audio/pat.wav\")"
+    "result = chatbot.predict(query=\"./assets/audio/sample.wav\")"
    ]
   },
   {
@@ -323,7 +323,7 @@
     "from neural_chat import VoiceChatClientExecutor\n",
     "executor = VoiceChatClientExecutor()\n",
     "result = executor(\n",
-    "    audio_input_path='./assets/audio/pat.wav',\n",
+    "    audio_input_path='./assets/audio/sample.wav',\n",
     "    audio_output_path='./results.wav',\n",
     "    server_ip=\"127.0.0.1\", # master server ip\n",
     "    port=8000 # master server entry point \n",
@@ -339,7 +339,7 @@
     "import IPython\n",
     "# Play input audio\n",
     "print(\"     Play Input Audio ......\")\n",
-    "IPython.display.display(IPython.display.Audio(\"./assets/audio/pat.wav\"))\n",
+    "IPython.display.display(IPython.display.Audio(\"./assets/audio/sample.wav\"))\n",
     "# Play output audio\n",
     "print(\"     Play Output Audio ......\")\n",
     "IPython.display.display(IPython.display.Audio(\"./assets/audio/welcome.wav\"))\n"
diff --git a/intel_extension_for_transformers/neural_chat/docs/notebooks/chatbot_on_nv_gpu.ipynb b/intel_extension_for_transformers/neural_chat/docs/notebooks/chatbot_on_nv_gpu.ipynb
index a3c15037738..c31e0d367cb 100644
--- a/intel_extension_for_transformers/neural_chat/docs/notebooks/chatbot_on_nv_gpu.ipynb
+++ b/intel_extension_for_transformers/neural_chat/docs/notebooks/chatbot_on_nv_gpu.ipynb
@@ -120,7 +120,7 @@
     "from intel_extension_for_transformers.neural_chat import build_chatbot\n",
     "config = PipelineConfig(audio_input=True, audio_output=True)\n",
     "chatbot = build_chatbot(config)\n",
-    "result = chatbot.predict(query=\"./assets/audio/pat.wav\")"
+    "result = chatbot.predict(query=\"./assets/audio/sample.wav\")"
    ]
   },
   {
@@ -331,7 +331,7 @@
     "from neural_chat import VoiceChatClientExecutor\n",
     "executor = VoiceChatClientExecutor()\n",
     "result = executor(\n",
-    "    audio_input_path='./assets/audio/pat.wav',\n",
+    "    audio_input_path='./assets/audio/sample.wav',\n",
     "    audio_output_path='./results.wav',\n",
     "    server_ip=\"127.0.0.1\", # master server ip\n",
     "    port=8000 # master server entry point \n",
@@ -347,7 +347,7 @@
     "import IPython\n",
     "# Play input audio\n",
     "print(\"     Play Input Audio ......\")\n",
-    "IPython.display.display(IPython.display.Audio(\"./assets/audio/pat.wav\"))\n",
+    "IPython.display.display(IPython.display.Audio(\"./assets/audio/sample.wav\"))\n",
     "# Play output audio\n",
     "print(\"     Play Output Audio ......\")\n",
     "IPython.display.display(IPython.display.Audio(\"./assets/audio/welcome.wav\"))\n"
diff --git a/intel_extension_for_transformers/neural_chat/models/mpt_model.py b/intel_extension_for_transformers/neural_chat/models/mpt_model.py
index c7b204f8700..1afd55a08ed 100644
--- a/intel_extension_for_transformers/neural_chat/models/mpt_model.py
+++ b/intel_extension_for_transformers/neural_chat/models/mpt_model.py
@@ -37,7 +37,7 @@ def match(self, model_path: str):
         Returns:
             bool: True if the model_path matches, False otherwise.
         """
-        return "mpt" in model_path.lower()
+        return "mpt" in model_path.lower() or "opt" in model_path.lower()
 
     def get_default_conv_template(self, model_path: str) -> Conversation:
         """
diff --git a/intel_extension_for_transformers/neural_chat/pipeline/plugins/audio/tts.py b/intel_extension_for_transformers/neural_chat/pipeline/plugins/audio/tts.py
index 8c77dcb12d4..fb0c0cec4f3 100644
--- a/intel_extension_for_transformers/neural_chat/pipeline/plugins/audio/tts.py
+++ b/intel_extension_for_transformers/neural_chat/pipeline/plugins/audio/tts.py
@@ -38,7 +38,8 @@ class TextToSpeech():
     2) Finetuned voice (Fine-tuned offline model of specific person, such as Pat's voice + corresponding embedding)
     3) Customized voice (Original model + User's customized input voice embedding)
     """
-    def __init__(self, output_audio_path="./response.wav", voice="default", stream_mode=False, device="cpu"):
+    def __init__(self, output_audio_path="./response.wav", voice="default", stream_mode=False, device="cpu", 
+      asset_path="/intel-extension-for-transformers/intel_extension_for_transformers/neural_chat/assets"):
         """Make sure your export LD_PRELOAD=<path to libiomp5.so and libtcmalloc> beforehand."""
         # default setting
         self.original_model = SpeechT5ForTextToSpeech.from_pretrained("microsoft/speecht5_tts")
@@ -56,8 +57,13 @@ def __init__(self, output_audio_path="./response.wav", voice="default", stream_m
         self.vocoder = SpeechT5HifiGan.from_pretrained("microsoft/speecht5_hifigan")
         self.vocoder.eval()
         script_dir = os.path.dirname(os.path.abspath(__file__))
-        default_speaker_embedding_path = os.path.join(script_dir,
-                                                      '../../../assets/speaker_embeddings/spk_embed_default.pt')
+        if os.path.exists(os.path.join(script_dir, '../../../assets/speaker_embeddings/spk_embed_default.pt')):
+            default_speaker_embedding_path = os.path.join(script_dir,
+                                                '../../../assets/speaker_embeddings/spk_embed_default.pt')
+        elif os.path.exists(os.path.join(asset_path, 'speaker_embeddings/spk_embed_default.pt')):
+            default_speaker_embedding_path = os.path.join(asset_path, 'speaker_embeddings/spk_embed_default.pt')
+        else:
+            print("Warning! Need to prepare speaker_embeddings")
         # load the default speaker embedding
         self.default_speaker_embedding = torch.load(default_speaker_embedding_path)
 
@@ -70,6 +76,8 @@ def __init__(self, output_audio_path="./response.wav", voice="default", stream_m
         pat_speaker_embedding_path = os.path.join(script_dir, '../../../assets/speaker_embeddings/spk_embed_pat.pt')
         if os.path.exists(pat_speaker_embedding_path):
             self.pat_speaker_embeddings = torch.load(pat_speaker_embedding_path)
+        elif os.path.exists(os.path.join(asset_path, 'speaker_embeddings/spk_embed_pat.pt')):
+            self.pat_speaker_embeddings = torch.load(os.path.join(asset_path, 'speaker_embeddings/spk_embed_pat.pt'))
 
         self.cpu_pool = None
         if not torch.cuda.is_available():
@@ -96,12 +104,16 @@ def create_speaker_embedding(self, driven_audio_path):
             speaker_embeddings = speaker_embeddings[0] # [1,512]
         return speaker_embeddings.cpu()
 
-    def _lookup_voice_embedding(self, voice):
+    def _lookup_voice_embedding(self, voice, 
+      asset_path="/intel-extension-for-transformers/intel_extension_for_transformers/neural_chat/assets"):
         script_dir = os.path.dirname(os.path.abspath(__file__))
         if os.path.exists(f"speaker_embeddings/spk_embed_{voice}.pt") == False:
             print("No customized speaker embedding is found! Use the default one")
-            default_speaker_embedding_path = os.path.join(script_dir,
-                                        '../../../assets/speaker_embeddings/spk_embed_default.pt')
+            if os.path.exists(os.path.join(script_dir, '../../../assets/speaker_embeddings/spk_embed_default.pt')):
+                default_speaker_embedding_path = os.path.join(script_dir,
+                                                    '../../../assets/speaker_embeddings/spk_embed_default.pt')
+            elif os.path.exists(os.path.join(asset_path, 'speaker_embeddings/spk_embed_default.pt')):
+                default_speaker_embedding_path = (asset_path, 'speaker_embeddings/spk_embed_default.pt')
             return default_speaker_embedding_path
         else:
             specific_speaker_embedding_path = os.path.join(script_dir,
diff --git a/intel_extension_for_transformers/neural_chat/pipeline/plugins/retrieval/retrieval_agent.py b/intel_extension_for_transformers/neural_chat/pipeline/plugins/retrieval/retrieval_agent.py
index c877ecb160a..3eed6105ac4 100644
--- a/intel_extension_for_transformers/neural_chat/pipeline/plugins/retrieval/retrieval_agent.py
+++ b/intel_extension_for_transformers/neural_chat/pipeline/plugins/retrieval/retrieval_agent.py
@@ -18,9 +18,9 @@
 import os
 import torch
 import transformers
-from intel_extension_for_transformers.neural_chat.pipeline.plugins.retrieval import Retriever
-from intel_extension_for_transformers.neural_chat.pipeline.plugins.retrieval.detector import IntentDetector
-from intel_extension_for_transformers.neural_chat.pipeline.plugins.retrieval.indexing import DocumentIndexing
+from .retrieval_base import Retriever
+from .detector import IntentDetector
+from .indexing import DocumentIndexing
 from intel_extension_for_transformers.neural_chat.pipeline.plugins.prompt import generate_qa_prompt, generate_prompt
 from intel_extension_for_transformers.neural_chat.plugins import register_plugin
 
@@ -33,7 +33,7 @@ def __init__(self, persist_dir="./output", process=True, input_path=None,
         self.model = None
         self.tokenizer = None
         self.retrieval_type = retrieval_type
-
+        self.retriever = None
         self.intent_detector = IntentDetector()
         if os.path.exists(input_path):
             self.doc_parser = DocumentIndexing(retrieval_type=self.retrieval_type, document_store=document_store,
@@ -53,7 +53,10 @@ def pre_llm_inference_actions(self, model_name, query):
             prompt = generate_prompt(query)
         else:
             print("Chat with QA agent.")
-            context = self.retriever.get_context(query)
-            prompt = generate_qa_prompt(query, context)
+            if self.retriever:
+                context = self.retriever.get_context(query)
+                prompt = generate_qa_prompt(query, context)
+            else:
+                prompt = generate_prompt(query)
         return prompt
 
diff --git a/intel_extension_for_transformers/neural_chat/pipeline/plugins/security/safety_checker.py b/intel_extension_for_transformers/neural_chat/pipeline/plugins/security/safety_checker.py
index 2f221931aa9..b6cfc2f1231 100644
--- a/intel_extension_for_transformers/neural_chat/pipeline/plugins/security/safety_checker.py
+++ b/intel_extension_for_transformers/neural_chat/pipeline/plugins/security/safety_checker.py
@@ -20,6 +20,7 @@
 import os
 from ....plugins import register_plugin
 
+doc_path = "/intel-extension-for-transformers/intel_extension_for_transformers/neural_chat/pipeline/plugins/security/"
 def convert_fullwidth_to_halfwidth(query):
     """Converting Full-width Characters to Half-width Characters."""
     content = ""
@@ -37,9 +38,19 @@ class SafetyChecker:
     def __init__(self, dict_path=None, matchType=2):
         if dict_path == None or (not os.path.exists(dict_path)):
             dict_path = os.path.dirname(os.path.abspath(__file__))
-        f = open(os.path.join(dict_path, "stopword.txt"), encoding="utf8")
+        if os.path.exists(os.path.join(dict_path, "stopword.txt")):
+            f = open(os.path.join(dict_path, "stopword.txt"), encoding="utf8")
+        elif os.path.exists(os.path.join(doc_path, "stopword.txt")):
+            f = open(os.path.join(doc_path, "stopword.txt"), encoding="utf8")
+        else:
+            print("Can't find stopword.txt")
         self.Stopwords = [i.split('\n')[0] for i in f.readlines()]
-        f1 = open(os.path.join(dict_path, "dict.txt"), encoding="utf8")
+        if os.path.exists(os.path.join(dict_path, "dict.txt")):
+            f1 = open(os.path.join(dict_path, "dict.txt"), encoding="utf8")
+        elif os.path.exists(os.path.join(doc_path, "dict.txt")):
+            f1 = open(os.path.join(doc_path, "dict.txt"), encoding="utf8")
+        else:
+            print("Can't find dict.txt")
         lst = f1.readlines()
         self.sensitiveWordSet = [i.split("\n")[0].split("\t") for i in lst]
         self.sensitiveWordMap = self._initSensitiveWordMap()
diff --git a/intel_extension_for_transformers/neural_chat/requirements.txt b/intel_extension_for_transformers/neural_chat/requirements.txt
index 5d655e30959..5f615d3a59d 100644
--- a/intel_extension_for_transformers/neural_chat/requirements.txt
+++ b/intel_extension_for_transformers/neural_chat/requirements.txt
@@ -1,4 +1,4 @@
-transformers>=4.31.0
+transformers>=4.32.0
 peft
 fschat
 torch
@@ -19,7 +19,6 @@ scikit-learn
 farm-haystack
 numpy==1.23.5
 librosa
-docx
 beautifulsoup4
 InstructorEmbedding
 chromadb
@@ -30,6 +29,7 @@ yacs
 uvicorn
 optimum
 optimum[habana]
-#py-cpuinfo
-#oneccl_bind_pt==2.0.100 -f https://developer.intel.com/ipex-whl-stable-xpu
-#deepspeed
+sentence_transformers
+unstructured
+markdown
+rouge_score
diff --git a/intel_extension_for_transformers/neural_chat/tests/api/test_chatbot_build_api.py b/intel_extension_for_transformers/neural_chat/tests/api/test_chatbot_build_api.py
index 2fcd7b5cfa1..838b7bdcea7 100644
--- a/intel_extension_for_transformers/neural_chat/tests/api/test_chatbot_build_api.py
+++ b/intel_extension_for_transformers/neural_chat/tests/api/test_chatbot_build_api.py
@@ -21,6 +21,7 @@
 from intel_extension_for_transformers.neural_chat import PipelineConfig, GenerationConfig
 from intel_extension_for_transformers.neural_chat import plugins
 
+# All UT cases use 'facebook/opt-125m' to reduce test time.
 class TestChatbotBuilder(unittest.TestCase):
     def setUp(self):
         return super().setUp()
@@ -29,14 +30,17 @@ def tearDown(self) -> None:
         return super().tearDown()
 
     def test_build_chatbot_with_default_config(self):
-        chatbot = build_chatbot()
+        config = PipelineConfig(model_name_or_path="facebook/opt-125m")
+        chatbot = build_chatbot(config)
         self.assertIsNotNone(chatbot)
         response = chatbot.predict(query="Tell me about Intel Xeon Scalable Processors.")
         print(response)
         self.assertIsNotNone(response)
 
     def test_build_chatbot_with_customized_pipelinecfg(self):
-        config = PipelineConfig(model_name_or_path="mosaicml/mpt-7b-chat")
+        config = PipelineConfig(model_name_or_path="facebook/opt-125m",
+                                tokenizer_name_or_path="EleutherAI/gpt-neox-20b",
+                                device="cpu")
         chatbot = build_chatbot(config)
         self.assertIsNotNone(chatbot)
         response = chatbot.predict(query="Tell me about Intel Xeon Scalable Processors.")
@@ -44,7 +48,8 @@ def test_build_chatbot_with_customized_pipelinecfg(self):
         self.assertIsNotNone(response)
 
     def test_build_chatbot_with_customized_generationcfg(self):
-        chatbot = build_chatbot()
+        config = PipelineConfig(model_name_or_path="facebook/opt-125m")
+        chatbot = build_chatbot(config)
         self.assertIsNotNone(chatbot)
         config = GenerationConfig(max_new_tokens=512, temperature=0.1)
         response = chatbot.predict(query="Tell me about Intel Xeon Scalable Processors.", config=config)
@@ -54,18 +59,22 @@ def test_build_chatbot_with_customized_generationcfg(self):
     def test_build_chatbot_with_audio_plugin(self):
         plugins.tts.enable = True
         plugins.tts.args["output_audio_path"]="./output_audio.wav"
-        pipeline_config = PipelineConfig(plugins=plugins)
+        pipeline_config = PipelineConfig(model_name_or_path="facebook/opt-125m",
+                                         plugins=plugins)
         chatbot = build_chatbot(pipeline_config)
         self.assertIsNotNone(chatbot)
         gen_config = GenerationConfig(max_new_tokens=64)
-        response = chatbot.predict(query="../../assets/audio/pat.wav", config=gen_config)
+        response = chatbot.predict(query= \
+          "/intel-extension-for-transformers/intel_extension_for_transformers/neural_chat/assets/audio/sample.wav", \
+          config=gen_config)
         self.assertIsNotNone(response)
         print("output audio path: ", response)
         self.assertTrue(os.path.exists(plugins.tts.args["output_audio_path"]))
 
     def test_build_chatbot_with_safety_checker_plugin(self):
         plugins.safety_checker.enable = True
-        pipeline_config = PipelineConfig(plugins=plugins)
+        pipeline_config = PipelineConfig(model_name_or_path="facebook/opt-125m",
+                                         plugins=plugins)
         chatbot = build_chatbot(pipeline_config)
         self.assertIsNotNone(chatbot)
         response = chatbot.predict(query="蔡英文是谁？")
@@ -74,11 +83,12 @@ def test_build_chatbot_with_safety_checker_plugin(self):
 
     def test_build_chatbot_with_retrieval_plugin(self):
         plugins.retrieval.enable = True
-        plugins.retrieval.args["input_path"] = "../../assets/docs/"
-        pipeline_config = PipelineConfig(plugins=plugins)
+        plugins.retrieval.args["input_path"] = "../../../../README.md"
+        pipeline_config = PipelineConfig(model_name_or_path="facebook/opt-125m",
+                                         plugins=plugins)
         chatbot = build_chatbot(pipeline_config)
         self.assertIsNotNone(chatbot)
-        response = chatbot.predict(query="What are total cores of Intel® Xeon® Platinum 8480+ Processor?")
+        response = chatbot.predict(query="What is Intel extension for transformers?")
         print("response: ", response)
 
 if __name__ == '__main__':
diff --git a/intel_extension_for_transformers/neural_chat/tests/api/test_inference.py b/intel_extension_for_transformers/neural_chat/tests/api/test_inference.py
index 3aa16d18172..1a3a28bf4ab 100644
--- a/intel_extension_for_transformers/neural_chat/tests/api/test_inference.py
+++ b/intel_extension_for_transformers/neural_chat/tests/api/test_inference.py
@@ -19,8 +19,7 @@
 import unittest
 from intel_extension_for_transformers.neural_chat.chatbot import build_chatbot, finetune_model, optimize_model
 from intel_extension_for_transformers.neural_chat.config import (
-    PipelineConfig, GenerationConfig, FinetuningConfig, OptimizationConfig,
-    ModelArguments, DataArguments, TrainingArguments, FinetuningArguments
+    PipelineConfig, GenerationConfig, AMPConfig,
 )
 from intel_extension_for_transformers.neural_chat import plugins
 
@@ -32,14 +31,17 @@ def tearDown(self) -> None:
         return super().tearDown()
 
     def test_text_chat(self):
-        config = PipelineConfig()
+        config = PipelineConfig(model_name_or_path="facebook/opt-125m")
         chatbot = build_chatbot(config)
         response = chatbot.predict("Tell me about Intel Xeon Scalable Processors.")
         print(response)
         self.assertIsNotNone(response)
 
     def test_retrieval(self):
-        config = PipelineConfig(retrieval_type="sparse", retrieval_document_path="../../assets/docs/")
+        plugins.retrieval.enable = True
+        plugins.retrieval.args["input_path"] = "../../assets/docs/"
+        config = PipelineConfig(model_name_or_path="facebook/opt-125m",
+                                plugins=plugins)
         chatbot = build_chatbot(config)
         response = chatbot.predict("Tell me about Intel Xeon Scalable Processors.")
         print(response)
@@ -48,7 +50,8 @@ def test_retrieval(self):
     def test_voice_chat(self):
         plugins.tts.enable = True
         plugins.tts.args["output_audio_path"] = "./response.wav"
-        pipeline_config = PipelineConfig(plugins=plugins)
+        pipeline_config = PipelineConfig(model_name_or_path="facebook/opt-125m",
+                                         plugins=plugins)
         chatbot = build_chatbot(config=pipeline_config)
         gen_config = GenerationConfig(max_new_tokens=64)
         response = chatbot.predict(query="Nice to meet you!", config=gen_config)
@@ -57,21 +60,9 @@ def test_voice_chat(self):
         print("output audio path: ", response)
         self.assertTrue(os.path.exists("./response.wav"))
 
-    def test_finetuning(self):
-        model_args = ModelArguments(model_name_or_path="meta-llama/Llama-2-7b-chat-hf", use_fast_tokenizer=False)
-        data_args = DataArguments(train_file='./alpaca_data.json', dataset_concatenation=True)
-        training_args = TrainingArguments(gradient_accumulation_steps=1,
-                                          do_train=True, learning_rate=1e-4, num_train_epochs=1,
-                                          logging_steps=100, save_total_limit=2, overwrite_output_dir=True,
-                                          log_level='info', save_strategy='epoch', max_steps=3,
-                                          output_dir='./saved_model', no_cuda=True)
-        finetune_args = FinetuningArguments(peft='lora')
-        config = FinetuningConfig(model_args, data_args, training_args, finetune_args)
-        finetune_model(config)
-
     def test_quantization(self):
-        config = OptimizationConfig()
-        optimize_model(config)
+        config = AMPConfig()
+        optimize_model(model="facebook/opt-125m", config=config)
 
 if __name__ == '__main__':
     unittest.main()
diff --git a/intel_extension_for_transformers/neural_chat/tests/audio/test_asr.py b/intel_extension_for_transformers/neural_chat/tests/audio/test_asr.py
index dc0fef0041d..2de1f452f5d 100644
--- a/intel_extension_for_transformers/neural_chat/tests/audio/test_asr.py
+++ b/intel_extension_for_transformers/neural_chat/tests/audio/test_asr.py
@@ -29,14 +29,14 @@ def setUpClass(self):
             self.asr_bf16 = AudioSpeechRecognition("openai/whisper-small", bf16=True)
 
     def test_audio2text(self):
-        audio_path = "../../assets/audio/pat.wav"
+        audio_path = "/intel-extension-for-transformers/intel_extension_for_transformers/neural_chat/assets/audio/welcome.wav"
         text = self.asr.audio2text(audio_path)
         self.assertEqual(text.lower(), "Welcome to Neural Chat".lower())
 
     def test_audio2text_bf16(self):
         if torch.cuda.is_available():
             return
-        audio_path = "../../assets/audio/pat.wav"
+        audio_path = "/intel-extension-for-transformers/intel_extension_for_transformers/neural_chat/assets/audio/welcome.wav"
         text = self.asr_bf16.audio2text(audio_path)
         self.assertEqual(text.lower(), "Welcome to Neural Chat".lower())
 
diff --git a/intel_extension_for_transformers/neural_chat/tests/audio/test_tts.py b/intel_extension_for_transformers/neural_chat/tests/audio/test_tts.py
index a5b5c14e050..ab6761df5b3 100644
--- a/intel_extension_for_transformers/neural_chat/tests/audio/test_tts.py
+++ b/intel_extension_for_transformers/neural_chat/tests/audio/test_tts.py
@@ -54,7 +54,8 @@ def text_generate():
             self.assertTrue(os.path.exists(result_path))
 
     def test_create_speaker_embedding(self):
-        driven_audio_path = "../../assets/audio/pat.wav"
+        driven_audio_path = \
+           "/intel-extension-for-transformers/intel_extension_for_transformers/neural_chat/assets/audio/sample.wav"
         spk_embed = self.tts.create_speaker_embedding(driven_audio_path)
         self.assertEqual(spk_embed.shape[0], 1)
         self.assertEqual(spk_embed.shape[1], 512)
diff --git a/intel_extension_for_transformers/neural_chat/tests/cli/test_cli.py b/intel_extension_for_transformers/neural_chat/tests/cli/test_cli.py
index 173f03a77be..f3bad881386 100644
--- a/intel_extension_for_transformers/neural_chat/tests/cli/test_cli.py
+++ b/intel_extension_for_transformers/neural_chat/tests/cli/test_cli.py
@@ -26,34 +26,39 @@ def test_text_chat(self):
         logger.info(f'Testing CLI request === Text Chat ===')
         command = 'neuralchat textchat \
                     --query "Tell me about Intel." \
-                    --model_name_or_path "./Llama-2-7b-chat-hf"'
+                    --model_name_or_path "facebook/opt-125m"'
+        result = None
         try:
-            result = subprocess.run(command, check=True)
+            result = subprocess.run(command, capture_output = True, check=True,
+                                    universal_newlines=True, shell=True) # nosec
         except subprocess.CalledProcessError as e:
             print("Error while executing command:", e)
-        print(result.stdout)
-        self.assertEqual(result.stdout, 0, msg="Textchat command line test failed.")
+        self.assertIn("model loaded", result.stdout)
 
     def test_help(self):
         logger.info(f'Testing CLI request === Help ===')
         command = 'neuralchat help'
+        result = None
         try:
-            result = subprocess.run(command, check=True)
+            result = subprocess.run(command, capture_output = True, check=True,
+                                    universal_newlines=True, shell=True) # nosec
         except subprocess.CalledProcessError as e:
             print("Error while executing command:", e)
-        self.assertEqual(result.stdout, 0, msg="Textchat command line test failed.")
+        self.assertIn("Show help for neuralchat commands.", result.stdout)
 
     def test_voice_chat(self):
         logger.info(f'Testing CLI request === Voice Chat ===')
         command = 'neuralchat voicechat \
                     --query "Tell me about Intel Xeon Scalable Processors." \
                     --audio_output_path "./response.wav" \
-                    --model_name_or_path "./Llama-2-7b-chat-hf"'
+                    --model_name_or_path "facebook/opt-125m"'
+        result = None
         try:
-            result = subprocess.run(command, check=True)
+            result = subprocess.run(command, capture_output = True, check=True,
+                                    universal_newlines=True, shell=True) # nosec
         except subprocess.CalledProcessError as e:
             print("Error while executing command:", e)
-        self.assertEqual(result.stdout, 0, msg="Textchat command line test failed.")
+        self.assertIn("model loaded", result.stdout)
 
 
 if __name__ == "__main__":
diff --git a/intel_extension_for_transformers/neural_chat/tests/optimization/test_optimization.py b/intel_extension_for_transformers/neural_chat/tests/optimization/test_optimization.py
index 9503c3ef1c6..262f3436511 100644
--- a/intel_extension_for_transformers/neural_chat/tests/optimization/test_optimization.py
+++ b/intel_extension_for_transformers/neural_chat/tests/optimization/test_optimization.py
@@ -31,7 +31,7 @@ def tearDown(self) -> None:
         return super().tearDown()
 
     def test_build_chatbot_with_AMP(self):
-        config = PipelineConfig()
+        config = PipelineConfig(model_name_or_path="facebook/opt-125m")
         chatbot = build_chatbot(config)
         self.assertIsNotNone(chatbot)
         response = chatbot.predict(query="Tell me about Intel Xeon Scalable Processors.")
@@ -39,7 +39,7 @@ def test_build_chatbot_with_AMP(self):
         self.assertIsNotNone(response)
 
     def test_build_chatbot_with_weight_only_quant(self):
-        config = PipelineConfig(
+        config = PipelineConfig(model_name_or_path="facebook/opt-125m",
             optimization_config=WeightOnlyQuantizationConfig()
         )
         chatbot = build_chatbot(config)
@@ -51,6 +51,7 @@ def test_build_chatbot_with_weight_only_quant(self):
     def test_build_chatbot_with_bitsandbytes_quant(self):
         if is_bitsandbytes_available() and torch.cuda.is_available():
             config = PipelineConfig(
+                model_name_or_path="facebook/opt-125m",
                 device='cuda',
                 optimization_config=BitsAndBytesConfig(
                         load_in_4bit=True,
diff --git a/intel_extension_for_transformers/neural_chat/tests/restful/config.py b/intel_extension_for_transformers/neural_chat/tests/restful/config.py
deleted file mode 100644
index 964f95c2aac..00000000000
--- a/intel_extension_for_transformers/neural_chat/tests/restful/config.py
+++ /dev/null
@@ -1,41 +0,0 @@
-# !/usr/bin/env python
-# -*- coding: utf-8 -*-
-#
-# Copyright (c) 2023 Intel Corporation
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#   http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-
-import os
-
-# Get the host and port from the environment variables
-host = os.environ.get('MY_HOST')
-port = os.environ.get('MY_PORT')
-
-# Check if the environment variables are set and not empty
-if host and port:
-    # Combine the host and port to form the full URL
-    HOST = f"http://{host}:{port}"
-    API_COMPLETION = '/v1/completions'
-    API_CHAT_COMPLETION = '/v1/chat/completions'
-    API_AUDIO = '/v1/voicechat/completions'
-    API_FINETUNE = '/v1/finetune'
-    API_TEXT2IMAGE = '/v1/text2image'
-
-    print("HOST URL:", HOST)
-    print("Completions Endpoint:", API_COMPLETION)
-    print("Chat completions Endpoint:", API_CHAT_COMPLETION)
-    print("Voice cbat Endpoint:", API_AUDIO)
-    print("Finerune Endpoint:", API_FINETUNE)
-    print("Text to image Endpoint:", API_TEXT2IMAGE)
-else:
-    raise("Please set the environment variables MY_HOST and MY_PORT.")
\ No newline at end of file
diff --git a/intel_extension_for_transformers/neural_chat/tests/restful/stress_test/README.md b/intel_extension_for_transformers/neural_chat/tests/restful/stress_test/README.md
deleted file mode 100644
index 68a784033cd..00000000000
--- a/intel_extension_for_transformers/neural_chat/tests/restful/stress_test/README.md
+++ /dev/null
@@ -1,34 +0,0 @@
-Stress Test for Restful API of Neural Chat
-====================
-
-The [Locust](https://github.com/locustio/locust) framework is used for stress test.
-
-## Preperations
-
-### Modify Configurations
-
-Before running stress test, you could modify `stress_test/locust.conf`:
-- locustfile: The stress test script to run.
-- headless: Do not use web UI, show test result in terminal directly.
-- host: The host IP and port of the backend server.
-- users: Number of users you want to mock.
-- spawn-rate: Rate to spawn users.
-- run-time: Stop after the specified amount of time, e.g. (300s,
-                        20m, 3h, 1h30m, etc.).
-
-For more configuration settings, refer to the [official documentations](https://docs.locust.io/en/stable/configuration.html) of locust.
-
-### install locust package
-
-```bash
-pip install locust
-```
-
-
-
-## Run Stress Test
-```
-cd stress_test
-locust
-```
-![stress test result](https://i.imgur.com/iCWkUQ6.jpeg)
\ No newline at end of file
diff --git a/intel_extension_for_transformers/neural_chat/tests/restful/stress_test/locust.conf b/intel_extension_for_transformers/neural_chat/tests/restful/stress_test/locust.conf
deleted file mode 100644
index 0d05f42605b..00000000000
--- a/intel_extension_for_transformers/neural_chat/tests/restful/stress_test/locust.conf
+++ /dev/null
@@ -1,23 +0,0 @@
-# !/usr/bin/env python
-# -*- coding: utf-8 -*-
-#
-# Copyright (c) 2023 Intel Corporation
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#   http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-
-locustfile = locust.py
-headless = true
-host = http://ip_address:port # change the ip address and port number according to your server
-users = 100
-spawn-rate = 10
-run-time = 1m # set upper-bound for time of tress test
\ No newline at end of file
diff --git a/intel_extension_for_transformers/neural_chat/tests/restful/stress_test/locust.py b/intel_extension_for_transformers/neural_chat/tests/restful/stress_test/locust.py
deleted file mode 100644
index 58d5e8de444..00000000000
--- a/intel_extension_for_transformers/neural_chat/tests/restful/stress_test/locust.py
+++ /dev/null
@@ -1,71 +0,0 @@
-# !/usr/bin/env python
-# -*- coding: utf-8 -*-
-#
-# Copyright (c) 2023 Intel Corporation
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#   http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-
-from locust import HttpUser, task, between
-from intel_extension_for_transformers.neural_chat.tests.restful.config import API_COMPLETION, API_CHAT_COMPLETION, API_ASR, API_TTS, API_FINETUNE, API_TEXT2IMAGE
-import time
-from intel_extension_for_transformers.neural_chat.server.restful.openai_protocol import CompletionRequest, ChatCompletionRequest
-from datasets import Dataset, Audio
-
-
-# locust will create a FeedbackUser instance for each user
-class FeedbackUser(HttpUser):
-    # each simulated user will wait 1~2 seconds for the next operation
-    wait_time = between(0.5, 2)
-
-    @task
-    def test_completions(self):
-        time.sleep(0.01)
-        request = CompletionRequest(
-            model="mpt-7b-chat",
-            prompt="This is a test."
-        )
-        self.client.post(API_COMPLETION, data=request)
-
-    @task
-    def test_chat_completions(self):
-        time.sleep(0.01)
-        request = ChatCompletionRequest(
-            model="mpt-7b-chat",
-            messages=[
-                {"role": "system","content": "You are a helpful assistant."},
-                {"role": "user","content": "Hello!"}
-            ]
-        )
-        self.client.post(API_CHAT_COMPLETION, data=request)
-
-    @task
-    def test_asr(self):
-        audio_path = "../../../assets/audio/pat.wav"
-        audio_dataset = Dataset.from_dict({"audio": [audio_path]}).cast_column("audio", Audio(sampling_rate=16000))
-        waveform = audio_dataset[0]["audio"]['array']
-        self.client.post(API_ASR, data=waveform)
-
-    @task
-    def test_tts(self):
-        text = "Welcome to Neural Chat"
-        self.client.post(API_TTS, data=text)
-
-    @task
-    def test_text2image(self):
-        text = "A running horse."
-        self.client.post(API_TEXT2IMAGE, data=text)
-
-    @task
-    def test_finetune(self):
-        self.client.post(API_FINETUNE)
-        time.sleep(2)
\ No newline at end of file
diff --git a/intel_extension_for_transformers/neural_chat/tests/restful/unit_test/test_finetune_api.py b/intel_extension_for_transformers/neural_chat/tests/restful/unit_test/test_finetune_api.py
deleted file mode 100644
index a75fd147c68..00000000000
--- a/intel_extension_for_transformers/neural_chat/tests/restful/unit_test/test_finetune_api.py
+++ /dev/null
@@ -1,66 +0,0 @@
-# !/usr/bin/env python
-# -*- coding: utf-8 -*-
-#
-# Copyright (c) 2023 Intel Corporation
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#   http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-
-import requests
-import unittest
-import shutil
-import os
-import json
-from intel_extension_for_transformers.neural_chat.tests.restful.config import HOST, API_FINETUNE
-from intel_extension_for_transformers.neural_chat.cli.log import logger
-
-
-json_data = \
-"""
-[
-    {"instruction": "Generate a slogan for a software company", "input": "", "output": "The Future of Software is Here"},
-    {"instruction": "Provide the word that comes immediately after the.", "input": "He threw the ball over the fence.", "output": "fence."}
-]
-"""
-test_data_file = '/test.json'
-
-
-class UnitTest(unittest.TestCase):
-
-    @classmethod
-    def setUpClass(self):
-        with open(test_data_file, mode='w') as f:
-            f.write(json_data)
-
-    @classmethod
-    def tearDownClass(self):
-        shutil.rmtree('./tmp', ignore_errors=True)
-        os.remove(test_data_file)
-
-    def __init__(self, *args):
-        super(UnitTest, self).__init__(*args)
-        self.host = HOST
-
-    def test_finetune(self):
-        logger.info(f'Testing POST request: {self.host+API_FINETUNE}')
-        request = {
-            "model_name_or_path": "facebook/opt-125m",
-            "train_file": "/test.json"
-        }
-        response = requests.post(self.host+API_FINETUNE, json.dumps(request))
-        logger.info('Response status code: {}'.format(response.status_code))
-        logger.info('Response text: {}'.format(response.text))
-        self.assertEqual(response.status_code, 200, msg="Abnormal response status code.")
-
-
-if __name__ == "__main__":
-    unittest.main()
\ No newline at end of file
diff --git a/intel_extension_for_transformers/neural_chat/tests/restful/unit_test/test_textchat_api.py b/intel_extension_for_transformers/neural_chat/tests/restful/unit_test/test_textchat_api.py
deleted file mode 100644
index 27633ad86ff..00000000000
--- a/intel_extension_for_transformers/neural_chat/tests/restful/unit_test/test_textchat_api.py
+++ /dev/null
@@ -1,55 +0,0 @@
-# !/usr/bin/env python
-# -*- coding: utf-8 -*-
-#
-# Copyright (c) 2023 Intel Corporation
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#   http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-
-import json
-import requests
-import unittest
-from intel_extension_for_transformers.neural_chat.tests.restful.config import HOST, API_COMPLETION, API_CHAT_COMPLETION
-from intel_extension_for_transformers.neural_chat.cli.log import logger
-
-
-class UnitTest(unittest.TestCase):
-
-    def __init__(self, *args):
-        super(UnitTest, self).__init__(*args)
-        self.host = HOST
-
-    def test_completions(self):
-        logger.info(f'Testing POST request: {self.host+API_COMPLETION}')
-        request = {
-            "prompt": "Tell me about Intel Xeon Scalable Processors."
-        }
-        response = requests.post(self.host+API_COMPLETION, json.dumps(request))
-        response_dict = response.json()
-        logger.info('Response status code: {}'.format(response.status_code))
-        logger.info('Response text: {}'.format(response_dict['response']))
-        self.assertEqual(response.status_code, 200, msg="Abnormal response status code.")
-
-    def test_chat_completions(self):
-        logger.info(f'Testing POST request: {self.host+API_CHAT_COMPLETION}')
-        request = {
-            "prompt": "Tell me about Intel Xeon Scalable Processors."
-        }
-        response = requests.post(self.host+API_CHAT_COMPLETION, json.dumps(request))
-        response_dict = response.json()
-        logger.info('Response status code: {}'.format(response.status_code))
-        logger.info('Response text: {}'.format(response_dict['response']))
-        self.assertEqual(response.status_code, 200, msg="Abnormal response status code.")
-
-
-if __name__ == "__main__":
-    unittest.main()
\ No newline at end of file
diff --git a/intel_extension_for_transformers/neural_chat/tests/restful/unit_test/test_voicechat_api.py b/intel_extension_for_transformers/neural_chat/tests/restful/unit_test/test_voicechat_api.py
deleted file mode 100644
index b2b1a8e9677..00000000000
--- a/intel_extension_for_transformers/neural_chat/tests/restful/unit_test/test_voicechat_api.py
+++ /dev/null
@@ -1,65 +0,0 @@
-# !/usr/bin/env python
-# -*- coding: utf-8 -*-
-#
-# Copyright (c) 2023 Intel Corporation
-#
-# Licensed under the Apache License, Version 2.0 (the "License");
-# you may not use this file except in compliance with the License.
-# You may obtain a copy of the License at
-#
-#   http://www.apache.org/licenses/LICENSE-2.0
-#
-# Unless required by applicable law or agreed to in writing, software
-# distributed under the License is distributed on an "AS IS" BASIS,
-# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-# See the License for the specific language governing permissions and
-# limitations under the License.
-
-import requests
-import unittest
-from datasets import Dataset, Audio
-from intel_extension_for_transformers.neural_chat.tests.restful.config import HOST, API_AUDIO
-from intel_extension_for_transformers.neural_chat.cli.log import logger
-
-
-class UnitTest(unittest.TestCase):
-
-    def __init__(self, *args):
-        super(UnitTest, self).__init__(*args)
-        self.host = HOST
-
-    def test_voicechat_text_out(self):
-        logger.info(f'Testing POST request: {self.host+API_AUDIO} with text output.')
-        audio_path = "../../../assets/audio/pat.wav"
-
-        with open(audio_path, "rb") as wav_file:
-            files = {
-                "file": ("audio.wav", wav_file, "audio/wav"),
-                "voice": (None, "pat"),
-                "audio_output_path": (None, " ")
-            }
-            response = requests.post(self.host+API_AUDIO, files=files)
-
-            logger.info('Response status code: {}'.format(response.status_code))
-            logger.info('Response text: {}'.format(response.text))
-            self.assertEqual(response.status_code, 200, msg="Abnormal response status code.")
-
-    def test_voicechat_audio_out(self):
-        logger.info(f'Testing POST request: {self.host+API_AUDIO} with audio output.')
-        audio_path = "../../../assets/audio/pat.wav"
-
-        with open(audio_path, "rb") as wav_file:
-            files = {
-                "file": ("audio.wav", wav_file, "audio/wav"),
-                "voice": (None, "pat"),
-                "audio_output_path": (None, "./response.wav")
-            }
-            response = requests.post(self.host+API_AUDIO, files=files)
-
-            logger.info('Response status code: {}'.format(response.status_code))
-            logger.info('Response text: {}'.format(response.text))
-            self.assertEqual(response.status_code, 200, msg="Abnormal response status code.")
-            
-
-if __name__ == "__main__":
-    unittest.main()