diff --git a/.github/workflows/release_code.yml b/.github/workflows/release_code.yml
index b340cc4..a11312c 100644
--- a/.github/workflows/release_code.yml
+++ b/.github/workflows/release_code.yml
@@ -32,20 +32,8 @@ on:
       - 'CMakeLists.txt'
       - 'csrc/**'
   push:
-    branches:
-      - 'main'
-      - '*-dev'
     tags:
       - 'v*'
-    paths:
-      - '.github/workflows/release_code.yml'
-      - 'vllm_ascend/**'
-      - 'setup.py'
-      - 'pyproject.toml'
-      - 'requirements.txt'
-      - 'cmake/**'
-      - 'CMakeLists.txt'
-      - 'csrc/**'
 
 jobs:
   build:
diff --git a/.github/workflows/release_whl.yml b/.github/workflows/release_whl.yml
index 8ed29ef..2e6a44b 100644
--- a/.github/workflows/release_whl.yml
+++ b/.github/workflows/release_whl.yml
@@ -36,21 +36,8 @@ on:
       - 'CMakeLists.txt'
       - 'csrc/**'
   push:
-    branches:
-      - 'main'
-      - '*-dev'
     tags:
       - 'v*'
-    paths:
-      - '.github/workflows/release_whl.yml'
-      - '.github/Dockerfile.buildwheel'
-      - 'vllm_ascend/**'
-      - 'setup.py'
-      - 'pyproject.toml'
-      - 'requirements.txt'
-      - 'cmake/**'
-      - 'CMakeLists.txt'
-      - 'csrc/**'
 
 jobs:
   build:
diff --git a/.github/workflows/vllm_ascend_test.yaml b/.github/workflows/vllm_ascend_test.yaml
index ee6bb5f..ec23fce 100644
--- a/.github/workflows/vllm_ascend_test.yaml
+++ b/.github/workflows/vllm_ascend_test.yaml
@@ -19,38 +19,12 @@ name: 'test'
 
 on:
   push:
-    # Enable merged test per commit
     branches:
       - 'main'
-    paths:
-      - '*.txt'
-      - '**/*.py'
-      - '.github/workflows/vllm_ascend_test.yaml'
-      - '!docs/**'
-      - 'pytest.ini'
-      - '!benchmarks/**'
-      - 'tools/mypy.sh'
-      - 'mypy.ini'
-      - '.github/workflows/*.ya?ml'
-      - '.github/workflows/actionlint.*'
-      - '.github/workflows/matchers/actionlint.json'
   pull_request:
     branches:
       - 'main'
       - '*-dev'
-    paths:
-      - '*.txt'
-      - '**/*.py'
-      - '.github/workflows/vllm_ascend_test.yaml'
-      - '!docs/**'
-      - '!examples/**'
-      - 'pytest.ini'
-      - '!benchmarks/**'
-      - 'tools/mypy.sh'
-      - 'mypy.ini'
-      - '.github/workflows/*.ya?ml'
-      - '.github/workflows/actionlint.*'
-      - '.github/workflows/matchers/actionlint.json'
 
 # Bash shells do not use ~/.profile or ~/.bashrc so these shells need to be explicitly
 # declared as "shell: bash -el {0}" on steps that need to be properly activated.
@@ -71,11 +45,35 @@ jobs:
     if: ${{ github.event_name == 'pull_request' }}
     uses: ./.github/workflows/pre-commit.yml
 
+  changes:
+    # Only trigger changes on pull request
+    if: ${{ github.event_name == 'pull_request' }}
+    runs-on: ubuntu-latest
+    permissions:
+      pull-requests: read
+    outputs:
+      e2e_tracker: ${{ steps.filter.outputs.e2e_tracker }}
+      ut_tracker: ${{ steps.filter.outputs.ut_tracker }}
+    steps:
+    - uses: dorny/paths-filter@v3
+      id: filter
+      with:
+        filters: |
+          e2e_tracker:
+            - 'vllm_ascend/**'
+            - 'csrc/**'
+            - 'cmake/**'
+            - 'tests/e2e/**'
+            - 'tests/conftest.py'
+            - 'tests/model_utils.py'
+            - 'tests/utils.py'
+          ut_tracker:
+            - 'tests/ut/**'
   ut:
-    needs: [lint]
+    needs: [lint, changes]
     name: unit test
-    # only trigger e2e test on [pull request after lint passed] and [merged commit]
-    if: ${{ needs.lint.result == 'success' || github.event_name == 'push' }}
+    # only trigger unit test after lint passed and the change is e2e and ut related. Or the PR is merged.
+    if: ${{ github.event_name == 'push' || (needs.lint.result == 'success' && (needs.changes.outputs.e2e_tracker == 'true' || needs.changes.outputs.ut_tracker == 'true')) }}
     runs-on: ubuntu-latest
     container:
       image: quay.io/ascend/cann:8.1.rc1-910b-ubuntu22.04-py3.10
@@ -134,9 +132,9 @@ jobs:
           verbose: true
 
   e2e:
-    needs: [lint]
-    # only trigger e2e test on pull request after lint passed
-    if: ${{ needs.lint.result == 'success' && github.event_name == 'pull_request' }}
+    needs: [lint, changes]
+    # only trigger e2e test after lint passed and the change is e2e related.
+    if: ${{ needs.lint.result == 'success' && needs.changes.outputs.e2e_tracker == 'true' }}
     strategy:
       max-parallel: 2
       matrix:
diff --git a/pytest.ini b/pytest.ini
deleted file mode 100644
index e59921e..0000000
--- a/pytest.ini
+++ /dev/null
@@ -1,68 +0,0 @@
-[pytest]
-minversion = 6.0
-norecursedirs = 
-    vllm-empty/tests/prefix_caching
-    vllm-empty/tests/weight_loading
-    vllm-empty/tests/samplers
-    vllm-empty/tests/kernels
-    vllm-empty/tests/quantization
-    vllm-empty/tests/tool_use
-    vllm-empty/tests/runai_model_streamer_test
-    vllm-empty/tests/kv_transfer
-    vllm-empty/tests/plugins
-    vllm-empty/tests/plugins_tests
-    vllm-empty/tests/prompt_adapter
-    vllm-empty/tests/compile
-    vllm-empty/tests/lora
-    vllm-empty/tests/models
-    vllm-empty/tests/mistral_tool_use
-    vllm-empty/tests/standalone_tests
-    vllm-empty/tests/async_engine
-    vllm-empty/tests/mq_llm_engine
-    vllm-empty/tests/tokenization
-    vllm-empty/tests/core
-    vllm-empty/tests/tracing
-    vllm-empty/tests/engine
-    vllm-empty/tests/tensorizer_loader
-    vllm-empty/tests/entrypoints
-    vllm-empty/tests/model_executor
-    vllm-empty/tests/encoder_decoder
-    vllm-empty/tests/v1
-    vllm-empty/tests/spec_decode
-    vllm-empty/tests/multi_step
-    vllm-empty/tests/vllm_test_utils
-    vllm-empty/tests/tpu
-    vllm-empty/tests/distributed
-    vllm-empty/tests/basic_correctness
-    vllm-empty/tests/worker
-    vllm-empty/tests/metrics
-    vllm-empty/tests/neuron
-  ; fastsafetensors not support npu now
-    vllm-empty/tests/fastsafetensors_loader
-
-addopts = --ignore=vllm-empty/tests/test_utils.py
-          --ignore=vllm-empty/tests/test_config.py
-          --ignore=vllm-empty/tests/test_scalartype.py
-          --ignore=vllm-empty/tests/test_version.py
-          --ignore=vllm-empty/tests/test_embedded_commit.py
-          --ignore=vllm-empty/tests/test_inputs.py
-          --ignore=vllm-empty/tests/test_sharded_state_loader.py
-          --ignore=vllm-empty/tests/test_logger.py
-          --ignore=vllm-empty/tests/test_logits_processor.py
-          --ignore=vllm-empty/tests/test_regression.py
-          --ignore=vllm-empty/tests/models/decoder_only/language/test_aqlm.py
-          --ignore=vllm-empty/tests/models/decoder_only/language/test_gptq_marlin.py
-          --ignore=vllm-empty/tests/models/decoder_only/language/test_gptq_marlin_24.py
-          --ignore=vllm-empty/tests/multimodal/test_processing.py
-          --ignore=vllm-empty/tests/multimodal/test_processor_kwargs.py
-          --ignore=vllm-empty/tests/multimodal/test_utils.py
-        ;   Both V1 and V0 engine will be run in detokenizer/test_stop_reason
-        ;   VLLM_USE_V1=1 is not supported with device type=npu.
-          --ignore=vllm-empty/tests/detokenizer/test_stop_reason.py
-        ;   oom on llama-2-7b-hf
-          --ignore=vllm-empty/tests/detokenizer/test_stop_strings.py
-        ; no need to run on vllm-ascend
-          --ignore=vllm-empty/tests/test_vllm_port.py
-
-testpaths =
-    vllm-empty/tests