Fix readthedocs (#2134 )

* [Fix]: limit urllib3 for readthedocs (#2070) * fix readthedocs for zh_cn * fix
add deform conv v3 plugin (#1872 )
2023-05-31 15:18:47 +08:00 · 2023-05-23 10:22:47 +08:00 · 2023-05-04 12:51:05 +08:00 · 2023-04-05 14:28:00 +08:00 · 2023-04-04 11:00:53 +08:00 · 2023-03-31 14:08:07 +08:00
927 changed files with 36907 additions and 12346 deletions
--- a/.circleci/scripts/linux/build.sh
+++ b/.circleci/scripts/linux/build.sh
@ -6,11 +6,7 @@ cd mmdeploy
 MMDEPLOY_DIR=$(pwd)
 mkdir -p build && cd build
 cmake .. -DMMDEPLOY_BUILD_SDK=ON -DMMDEPLOY_BUILD_TEST=ON -DMMDEPLOY_BUILD_SDK_PYTHON_API=ON \
-         -DMMDEPLOY_BUILD_SDK_CXX_API=ON -DMMDEPLOY_BUILD_SDK_CSHARP_API=ON \
+         -DMMDEPLOY_BUILD_EXAMPLES=ON -DMMDEPLOY_BUILD_SDK_CSHARP_API=ON \
         -DMMDEPLOY_TARGET_DEVICES="$1" -DMMDEPLOY_TARGET_BACKENDS="$2" "${ARGS[@]:2}"

 make -j$(nproc) && make install
-cd install/example
-mkdir -p build
-cd build
-cmake ../cpp -DMMDeploy_DIR="$MMDEPLOY_DIR"/build/install/lib/cmake/MMDeploy "${ARGS[@]:2}" && make -j$(nproc)
--- a/.circleci/scripts/windows/install_opencv.ps1
+++ b/.circleci/scripts/windows/install_opencv.ps1
@ -1,3 +1,3 @@
-Invoke-WebRequest -Uri https://download.openmmlab.com/mmdeploy/library/opencv-4.5.5.zip -OutFile opencv.zip
+Invoke-WebRequest -Uri https://github.com/irexyc/mmdeploy-ci-resource/releases/download/opencv/opencv-win-amd64-4.5.5-vc16.zip -OutFile opencv.zip
 Expand-Archive opencv.zip .
 Move-Item opencv-4.5.5 opencv
--- a/.circleci/test.yml
+++ b/.circleci/test.yml
@ -74,7 +74,7 @@ commands:
      - run:
          name: Install mmcv-full
          command: |
-            python -m pip install opencv-python==4.5.4.60
+            python -m pip install opencv-python==4.5.4.60 opencv-contrib-python==4.5.4.60 opencv-python-headless==4.5.4.60
            python -m pip install mmcv-full==<< parameters.version >> -f https://download.openmmlab.com/mmcv/dist/cpu/torch<< parameters.torch >>/index.html
  install_mmcv_cuda:
    parameters:
@ -91,7 +91,7 @@ commands:
      - run:
          name: Install mmcv-full
          command: |
-            python -m pip install opencv-python==4.5.4.60
+            python -m pip install opencv-python==4.5.4.60 opencv-contrib-python==4.5.4.60 opencv-python-headless==4.5.4.60
            python -m pip install mmcv-full==<< parameters.version >> -f https://download.openmmlab.com/mmcv/dist/<< parameters.cuda >>/torch<< parameters.torch >>/index.html
  install_mmdeploy:
    description: "Install MMDeploy"
@ -216,19 +216,12 @@ jobs:
             -DMMDEPLOY_BUILD_SDK=ON `
             -DMMDEPLOY_BUILD_TEST=ON `
             -DMMDEPLOY_BUILD_SDK_PYTHON_API=ON `
-             -DMMDEPLOY_BUILD_SDK_CXX_API=ON `
+             -DMMDEPLOY_BUILD_EXAMPLES=ON `
             -DMMDEPLOY_BUILD_SDK_CSHARP_API=ON `
             -DMMDEPLOY_TARGET_BACKENDS="ort" `
             -DOpenCV_DIR="$env:OPENCV_PACKAGE_DIR"
            cmake --build . --config Release -- /m
            cmake --install . --config Release
-            cd install/example
-            mkdir build -ErrorAction SilentlyContinue
-            cd build
-            cmake ../cpp -G "Visual Studio 16 2019" -A x64 -T v142 `
-             -DMMDeploy_DIR="$env:MMDEPLOY_DIR/build/install/lib/cmake/MMDeploy" `
-             -DOpenCV_DIR="$env:OPENCV_PACKAGE_DIR"
-            cmake --build . --config Release -- /m
      - install_mmdeploy
      - install_model_converter_req
      - perform_model_converter_ut
@ -280,7 +273,7 @@ jobs:
      - run:
          name: Inference model by SDK
          command: |
-            mmdeploy/build/install/example/build/image_classification cpu mmdeploy-models/mmcls/onnxruntime mmclassification/demo/demo.JPEG
+            ./mmdeploy/build/bin/image_classification cpu mmdeploy-models/mmcls/onnxruntime mmclassification/demo/demo.JPEG


 # See: https://circleci.com/docs/2.0/configuration-reference/#workflows
--- a/.codespell_ignore.txt
+++ b/.codespell_ignore.txt
@ -1,2 +1,3 @@
 cann
 CANN
+nd
--- a/.github/ISSUE_TEMPLATE/1-bug-report.yml
+++ b/.github/ISSUE_TEMPLATE/1-bug-report.yml
@ -1,6 +1,7 @@
-name: Bug report
-description: Create a report to help us improve
-
+name: 🐞 Bug report
+description: Create a report to help us reproduce and fix the bug
+title: "[Bug] "
+labels: ['Bug']

 body:
 - type: checkboxes
@ -52,5 +53,3 @@ body:
     If you have already identified the reason, you can provide the information here. If you are willing to create a PR to fix it, please also leave a comment here and that would be much appreciated!

     Thanks for your bug report. We appreciate it a lot.
-
-labels: ['Bug']
--- a/.github/ISSUE_TEMPLATE/2-feature_request.yaml
+++ b/.github/ISSUE_TEMPLATE/2-feature_request.yaml
@ -1,11 +1,15 @@
-name: Feature request
+name: 🚀 Feature request
 description: Suggest an idea for this project
+title: "[Feature] "

 body:
 - type: markdown
  attributes:
-    value: >
-      ## Describe the feature
+    value: |
+      We strongly appreciate you creating a PR to implement this feature [here](https://github.com/open-mmlab/mmdeploy/pulls)!
+      If you need our help, please fill in as much of the following form as you're able to.
+
+      **The less clear the description, the longer it will take to solve it.**
 - type: textarea
  attributes:
    label: Motivation
--- a/.github/ISSUE_TEMPLATE/3-documentation.yml
+++ b/.github/ISSUE_TEMPLATE/3-documentation.yml
@ -0,0 +1,23 @@
+name: 📚 Documentation
+description: Report an issue related to the documentation.
+labels: "kind/doc,status/unconfirmed"
+title: "[Docs] "
+
+body:
+- type: textarea
+  attributes:
+    label: 📚 The doc issue
+    description: >
+      A clear and concise description the issue.
+  validations:
+    required: true
+
+- type: textarea
+  attributes:
+    label: Suggest a potential alternative/fix
+    description: >
+      Tell us how we could improve the documentation in this regard.
+- type: markdown
+  attributes:
+    value: >
+      Thanks for contributing 🎉!
--- a/.github/ISSUE_TEMPLATE/config.yml
+++ b/.github/ISSUE_TEMPLATE/config.yml
@ -1,6 +1,12 @@
 blank_issues_enabled: false

 contact_links:
-  - name: Common Issues
+  - name: 💥 FAQ
    url: https://github.com/open-mmlab/mmdeploy/blob/master/docs/en/faq.md
    about: Check if your issue already has solutions
+  - name: 💬 Forum
+    url: https://github.com/open-mmlab/mmdeploy/discussions
+    about: Ask general usage questions and discuss with other MMDeploy community members
+  - name: 🌐 Explore OpenMMLab
+    url: https://openmmlab.com/
+    about: Get know more about OpenMMLab
--- a/.github/ISSUE_TEMPLATE/general_questions.md
+++ b/.github/ISSUE_TEMPLATE/general_questions.md
@ -1,7 +0,0 @@
---
-name: General questions
-about: Ask general questions to get help
-title: ''
-labels: ''
-assignees: ''
---
--- a/.github/release.yml
+++ b/.github/release.yml
@ -0,0 +1,32 @@
+changelog:
+  categories:
+    - title: 🚀 Features
+      labels:
+        - feature
+        - enhancement
+    - title: 💥 Improvements
+      labels:
+        - improvement
+    - title: 🐞 Bug fixes
+      labels:
+        - bug
+        - Bug:P0
+        - Bug:P1
+        - Bug:P2
+        - Bug:P3
+    - title: 📚 Documentations
+      labels:
+        - documentation
+    - title: 🌐 Other
+      labels:
+        - '*'
+      exclude:
+        labels:
+          - feature
+          - enhancement
+          - bug
+          - documentation
+          - Bug:P0
+          - Bug:P1
+          - Bug:P2
+          - Bug:P3
--- a/.github/scripts/test_java_demo.py
+++ b/.github/scripts/test_java_demo.py
@ -47,6 +47,13 @@ PARAMS = [
        'configs': [
            'https://media.githubusercontent.com/media/hanrui1sensetime/mmdeploy-javaapi-testdata/master/litehrnet.tar'  # noqa: E501
        ]
+    },
+    {
+        'task':
+        'RotatedDetection',
+        'configs': [
+            'https://media.githubusercontent.com/media/hanrui1sensetime/mmdeploy-javaapi-testdata/master/gliding-vertex.tar'  # noqa: E501
+        ]
    }
 ]

--- a/.github/workflows/backend-ascend.yml
+++ b/.github/workflows/backend-ascend.yml
@ -16,15 +16,18 @@ concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: true

+permissions:
+  contents: read
+
 jobs:
  build_sdk_demo:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
    strategy:
      matrix:
        python-version: [3.7]
    steps:
      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v2
+        uses: actions/setup-python@v4
        with:
          python-version: ${{ matrix.python-version }}
      - name: Checkout repository
@ -37,9 +40,7 @@ jobs:
        run: |
          sudo apt update
          sudo apt install -y ffmpeg libsm6 libxext6 git ninja-build libglib2.0-0 libxrender-dev libc++1-9 libc++abi1-9
-          sudo add-apt-repository ppa:ignaciovizzo/opencv3-nonfree
          sudo apt install libopencv-dev
-          pkg-config --libs opencv
      - name: Install Ascend Toolkit
        run: |
          mkdir -p $GITHUB_WORKSPACE/Ascend
@ -50,5 +51,5 @@ jobs:
          mkdir -p build && pushd build
          source $GITHUB_WORKSPACE/Ascend/ascend-toolkit/set_env.sh
          export LD_LIBRARY_PATH=$GITHUB_WORKSPACE/Ascend/ascend-toolkit/latest/runtime/lib64/stub:$LD_LIBRARY_PATH
-          cmake .. -DCMAKE_CXX_COMPILER=g++-7    -DMMDEPLOY_SHARED_LIBS=ON   -DMMDEPLOY_BUILD_SDK=ON       -DMMDEPLOY_BUILD_SDK_PYTHON_API=OFF  -DMMDEPLOY_TARGET_DEVICES=cpu   -DMMDEPLOY_BUILD_EXAMPLES=ON    -DMMDEPLOY_TARGET_BACKENDS=acl  -DMMDEPLOY_CODEBASES=all
+          cmake .. -DMMDEPLOY_BUILD_SDK=ON -DMMDEPLOY_BUILD_EXAMPLES=ON -DMMDEPLOY_TARGET_BACKENDS=acl
          make install -j4
--- a/.github/workflows/backend-coreml.yml
+++ b/.github/workflows/backend-coreml.yml
@ -39,22 +39,21 @@ jobs:
          wget https://github.com/irexyc/mmdeploy-ci-resource/releases/download/libtorch/libtorch-osx-arm64-1.8.0.tar.gz
          mkdir $GITHUB_WORKSPACE/libtorch-install
          tar xf libtorch-osx-arm64-1.8.0.tar.gz -C $GITHUB_WORKSPACE/libtorch-install
-      - name: build
+      - name: build-static-lib
        run: |
          mkdir build && cd build
          cmake .. -DCMAKE_OSX_ARCHITECTURES="arm64" \
            -DCMAKE_SYSTEM_PROCESSOR="arm64" \
            -DMMDEPLOY_BUILD_SDK=ON \
-            -DMMDEPLOY_TARGET_DEVICES="cpu" \
-            -DMMDEPLOY_CODEBASES=all \
            -DOpenCV_DIR=$GITHUB_WORKSPACE/opencv-install/lib/cmake/opencv4 \
            -DTorch_DIR=$GITHUB_WORKSPACE/libtorch-install/share/cmake/Torch \
            -DMMDEPLOY_TARGET_BACKENDS="coreml" \
            -DMMDEPLOY_BUILD_EXAMPLES=ON \
+            -DMMDEPLOY_BUILD_SDK_MONOLITHIC=OFF \
            -DMMDEPLOY_SHARED_LIBS=OFF
          cmake --build . -j 3
          cmake --build . --target install
-      - name: build-shared
+      - name: build-monolithic-lib
        run: |
          mkdir build-shared && cd build-shared
          cmake .. -DCMAKE_OSX_ARCHITECTURES="arm64" \
@ -65,7 +64,8 @@ jobs:
            -DOpenCV_DIR=$GITHUB_WORKSPACE/opencv-install/lib/cmake/opencv4 \
            -DTorch_DIR=$GITHUB_WORKSPACE/libtorch-install/share/cmake/Torch \
            -DMMDEPLOY_TARGET_BACKENDS="coreml" \
-            -DMMDEPLOY_BUILD_EXAMPLES=ON \
-            -DMMDEPLOY_SHARED_LIBS=ON
+            -DMMDEPLOY_BUILD_SDK_MONOLITHIC=ON \
+            -DMMDEPLOY_SHARED_LIBS=OFF \
+            -DMMDEPLOY_BUILD_EXAMPLES=ON
          cmake --build . -j 3
          cmake --build . --target install
--- a/.github/workflows/backend-ncnn.yml
+++ b/.github/workflows/backend-ncnn.yml
@ -2,11 +2,19 @@ name: backend-ncnn

 on:
  push:
+    branches:
+      - main
+      - master
+      - dev-1.x
    paths-ignore:
      - "demo/**"
      - "tools/**"

  pull_request:
+    branches:
+      - main
+      - master
+      - dev-1.x
    paths-ignore:
      - "demo/**"
      - "tools/**"
@ -16,9 +24,12 @@ concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: true

+permissions:
+  contents: read
+
 jobs:
  test_onnx2ncnn:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
    strategy:
      matrix:
        python-version: [3.7]
@ -28,7 +39,7 @@ jobs:
        with:
          submodules: 'recursive'
      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v2
+        uses: actions/setup-python@v4
        with:
          python-version: ${{ matrix.python-version }}
      - name: Install unittest dependencies
@ -71,12 +82,12 @@ jobs:
        with:
          submodules: 'recursive'
      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v2
+        uses: actions/setup-python@v4
        with:
          python-version: ${{ matrix.python-version }}
      - name: Install mmdeploy
        run: |
-          python3 tools/scripts/build_ubuntu_x64_ncnn.py
          python3 -m pip install torch==1.8.2 torchvision==0.9.2 --extra-index-url https://download.pytorch.org/whl/lts/1.8/cpu
          python3 -m pip install mmcv-full==1.5.1  -f https://download.openmmlab.com/mmcv/dist/cpu/torch1.8.0/index.html
-          python3 -c 'import mmdeploy.apis.ncnn as ncnn_api; assert ncnn_api.is_available() and ncnn_api.is_custom_ops_available()'
+          python3 tools/scripts/build_ubuntu_x64_ncnn.py 8
+          python3 -c 'import mmdeploy.apis.ncnn as ncnn_api; assert ncnn_api.is_available(with_custom_ops=True)'
--- a/.github/workflows/backend-ort.yml
+++ b/.github/workflows/backend-ort.yml
@ -16,6 +16,9 @@ concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: true

+permissions:
+  contents: read
+
 jobs:
  script_install:
    runs-on: ubuntu-20.04
@ -28,15 +31,15 @@ jobs:
        with:
          submodules: 'recursive'
      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v2
+        uses: actions/setup-python@v4
        with:
          python-version: ${{ matrix.python-version }}
      - name: Install mmdeploy
        run: |
-          python3 tools/scripts/build_ubuntu_x64_ort.py
          python3 -m pip install torch==1.8.2 torchvision==0.9.2 --extra-index-url https://download.pytorch.org/whl/lts/1.8/cpu
          python3 -m pip install mmcv-full==1.5.1  -f https://download.openmmlab.com/mmcv/dist/cpu/torch1.8.0/index.html
-          python3 -c 'import mmdeploy.apis.onnxruntime as ort_api; assert ort_api.is_available() and ort_api.is_custom_ops_available()'
+          python3 tools/scripts/build_ubuntu_x64_ort.py 8
+          python3 -c 'import mmdeploy.apis.onnxruntime as ort_api; assert ort_api.is_available(with_custom_ops=True)'
      - name: test mmcls full pipeline
        run: |
          pip install openmim
--- a/.github/workflows/backend-pplnn.yml
+++ b/.github/workflows/backend-pplnn.yml
@ -16,24 +16,20 @@ concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: true

+permissions:
+  contents: read
+
 jobs:
  script_install:
-    runs-on: ubuntu-18.04
-    strategy:
-      matrix:
-        python-version: [3.7]
+    runs-on: ubuntu-20.04
    steps:
      - name: Checkout repository
        uses: actions/checkout@v3
        with:
          submodules: 'recursive'
-      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v2
-        with:
-          python-version: ${{ matrix.python-version }}
      - name: Install mmdeploy
        run: |
-          python3 tools/scripts/build_ubuntu_x64_pplnn.py
          python3 -m pip install torch==1.8.2 torchvision==0.9.2 --extra-index-url https://download.pytorch.org/whl/lts/1.8/cpu
          python3 -m pip install mmcv-full==1.5.1  -f https://download.openmmlab.com/mmcv/dist/cpu/torch1.8.0/index.html
+          python3 tools/scripts/build_ubuntu_x64_pplnn.py 8
          python3 -c 'import mmdeploy.apis.pplnn as pplnn_api; assert pplnn_api.is_available()'
--- a/.github/workflows/backend-rknn.yml
+++ b/.github/workflows/backend-rknn.yml
@ -0,0 +1,47 @@
+name: backend-rknn
+
+on:
+  push:
+    paths:
+      - "csrc/**"
+      - "demo/csrc/**"
+      - "CMakeLists.txt"
+
+  pull_request:
+    paths-ignore:
+      - "csrc/**"
+      - "demo/csrc/**"
+      - "CMakeLists.txt"
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
+permissions:
+  contents: read
+
+jobs:
+  build_rknpu2:
+    runs-on: ubuntu-20.04
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v3
+        with:
+          submodules: 'recursive'
+      - name: update
+        run: sudo apt update
+      - name: cross compile
+        run: |
+          sh -x tools/scripts/ubuntu_cross_build_rknn.sh rk3588
+  build_rknpu:
+    runs-on: ubuntu-20.04
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v3
+        with:
+          submodules: 'recursive'
+      - name: update
+        run: sudo apt update
+      - name: cross compile
+        run: |
+          sh -x tools/scripts/ubuntu_cross_build_rknn.sh rv1126
--- a/.github/workflows/backend-snpe.yml
+++ b/.github/workflows/backend-snpe.yml
@ -16,9 +16,12 @@ concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: true

+permissions:
+  contents: read
+
 jobs:
  build_sdk_demo:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
    steps:
      - name: Checkout repository
        uses: actions/checkout@v3
@ -31,9 +34,7 @@ jobs:
          sudo apt install wget libprotobuf-dev protobuf-compiler
          sudo apt update
          sudo apt install -y ffmpeg libsm6 libxext6 git ninja-build libglib2.0-0 libxrender-dev libc++1-9 libc++abi1-9
-          sudo add-apt-repository ppa:ignaciovizzo/opencv3-nonfree
          sudo apt install libopencv-dev
-          pkg-config --libs opencv
      - name: Install snpe
        run: |
          wget https://media.githubusercontent.com/media/tpoisonooo/mmdeploy_snpe_testdata/main/snpe-1.59.tar.gz
@ -47,7 +48,7 @@ jobs:
          export SNPE_ROOT=/home/runner/work/mmdeploy/mmdeploy/snpe-1.59.0.3230
          export LD_LIBRARY_PATH=${SNPE_ROOT}/lib/x86_64-linux-clang:${LD_LIBRARY_PATH}
          export MMDEPLOY_SNPE_X86_CI=1
-          cmake .. -DCMAKE_CXX_COMPILER=g++-7    -DMMDEPLOY_SHARED_LIBS=ON   -DMMDEPLOY_BUILD_SDK=ON       -DMMDEPLOY_BUILD_SDK_PYTHON_API=OFF  -DMMDEPLOY_TARGET_DEVICES=cpu       -DMMDEPLOY_TARGET_BACKENDS=snpe  -DMMDEPLOY_CODEBASES=all
+          cmake .. -DMMDEPLOY_BUILD_SDK=ON -DMMDEPLOY_TARGET_BACKENDS=snpe
          make -j2
          make install
          pushd install/example
--- a/.github/workflows/backend-torchscript.yml
+++ b/.github/workflows/backend-torchscript.yml
@ -16,9 +16,12 @@ concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: true

+permissions:
+  contents: read
+
 jobs:
  script_install:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
    strategy:
      matrix:
        python-version: [3.7]
@ -28,9 +31,9 @@ jobs:
        with:
          submodules: 'recursive'
      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v2
+        uses: actions/setup-python@v4
        with:
          python-version: ${{ matrix.python-version }}
      - name: Install mmdeploy
        run: |
-          python3 tools/scripts/build_ubuntu_x64_torchscript.py
+          python3 tools/scripts/build_ubuntu_x64_torchscript.py 8
--- a/.github/workflows/backend-tvm.yml
+++ b/.github/workflows/backend-tvm.yml
@ -0,0 +1,44 @@
+name: backend-tvm
+
+on:
+  push:
+    paths-ignore:
+      - "demo/**"
+      - "tools/**"
+
+  pull_request:
+    paths-ignore:
+      - "demo/**"
+      - "tools/**"
+      - "docs/**"
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
+permissions:
+  contents: read
+
+jobs:
+  script_install:
+    runs-on: ubuntu-20.04
+    strategy:
+      matrix:
+        python-version: [3.7]
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v3
+        with:
+          submodules: 'recursive'
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v4
+        with:
+          python-version: ${{ matrix.python-version }}
+      - name: Install mmdeploy
+        run: |
+          python3 -m pip install torch==1.8.2 torchvision==0.9.2 --extra-index-url https://download.pytorch.org/whl/lts/1.8/cpu
+          python3 -m pip install mmcv-full==1.5.1  -f https://download.openmmlab.com/mmcv/dist/cpu/torch1.8.0/index.html
+          python3 -m pip install decorator psutil scipy attrs tornado pytest
+          python3 tools/scripts/build_ubuntu_x64_tvm.py 8
+          source ~/mmdeploy.env
+          python3 -c 'import mmdeploy.apis.tvm as tvm_api; assert tvm_api.is_available()'
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@ -20,12 +20,14 @@ concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: true

+permissions:
+  contents: read
+
 jobs:
  build_cpu_model_convert:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
    strategy:
      matrix:
-        python-version: [3.7]
        torch: [1.8.0, 1.9.0]
        mmcv: [1.4.2]
        include:
@ -36,23 +38,23 @@ jobs:
            torch_version: torch1.9
            torchvision: 0.10.0
    steps:
-      - uses: actions/checkout@v2
-      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v2
-        with:
-          python-version: ${{ matrix.python-version }}
+      - uses: actions/checkout@v3
      - name: Install PyTorch
-        run: pip install torch==${{matrix.torch}}+cpu torchvision==${{matrix.torchvision}}+cpu -f https://download.pytorch.org/whl/torch_stable.html
+        run: |
+          python -m pip install --upgrade pip
+          python -V
+          python -m pip install torch==${{matrix.torch}}+cpu torchvision==${{matrix.torchvision}}+cpu -f https://download.pytorch.org/whl/torch_stable.html
      - name: Install MMCV
        run: |
-          pip install mmcv-full==${{matrix.mmcv}} -f https://download.openmmlab.com/mmcv/dist/cpu/${{matrix.torch_version}}/index.html
+          python -m pip install mmcv-full==${{matrix.mmcv}} -f https://download.openmmlab.com/mmcv/dist/cpu/${{matrix.torch_version}}/index.html
          python -c 'import mmcv; print(mmcv.__version__)'
      - name: Install unittest dependencies
        run: |
-          pip install -r requirements.txt
-          pip install -U numpy
+          python -m pip install -U numpy
+          python -m pip install rapidfuzz==2.15.1
+          python -m pip install -r requirements.txt
      - name: Build and install
-        run: rm -rf .eggs && pip install -e .
+        run: rm -rf .eggs && python -m pip install -e .
      - name: Run python unittests and generate coverage report
        run: |
          coverage run --branch --source mmdeploy -m pytest -rsE tests
@ -60,7 +62,7 @@ jobs:
          coverage report -m

  build_cpu_sdk:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
    steps:
      - name: Checkout repository
        uses: actions/checkout@v3
@ -70,16 +72,18 @@ jobs:
        run: sudo apt update
      - name: gcc-multilib
        run: |
-          sudo apt install gcc-multilib g++-multilib wget libprotobuf-dev protobuf-compiler
          sudo apt update
-          sudo apt install -y ffmpeg libsm6 libxext6 git ninja-build libglib2.0-0 libxrender-dev libc++1-9 libc++abi1-9
-          sudo add-apt-repository ppa:ignaciovizzo/opencv3-nonfree
-          sudo apt install libopencv-dev lcov wget
-          pkg-config --libs opencv
+          sudo apt install libopencv-dev lcov wget -y
      - name: Build and run SDK unit test without backend
        run: |
          mkdir -p build && pushd build
-          cmake .. -DCMAKE_CXX_COMPILER=g++-7  -DMMDEPLOY_CODEBASES=all    -DMMDEPLOY_BUILD_SDK=ON     -DMMDEPLOY_BUILD_SDK_PYTHON_API=OFF     -DMMDEPLOY_TARGET_DEVICES=cpu   -DMMDEPLOY_COVERAGE=ON  -DMMDEPLOY_BUILD_TEST=ON
+          cmake .. \
+            -DMMDEPLOY_CODEBASES=all \
+            -DMMDEPLOY_BUILD_SDK=ON \
+            -DMMDEPLOY_BUILD_SDK_PYTHON_API=OFF \
+            -DMMDEPLOY_TARGET_DEVICES=cpu \
+            -DMMDEPLOY_COVERAGE=ON \
+            -DMMDEPLOY_BUILD_TEST=ON
          make -j2
          mkdir -p mmdeploy_test_resources/transform
          cp ../tests/data/tiger.jpeg mmdeploy_test_resources/transform/
@ -88,15 +92,32 @@ jobs:
          ls -lah coverage.info
          cp coverage.info ../

+  cross_build_aarch64:
+    runs-on: ubuntu-20.04
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v3
+        with:
+          submodules: 'recursive'
+      - name: update
+        run: sudo apt update
+      - name: Set up Python
+        uses: actions/setup-python@v4
+        with:
+          python-version: 3.8
+      - name: gcc-multilib
+        run: |
+          sh -x tools/scripts/ubuntu_cross_build_aarch64.sh
+
  build_cuda102:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
    container:
      image: pytorch/pytorch:1.9.0-cuda10.2-cudnn7-devel
    env:
      FORCE_CUDA: 1
    strategy:
      matrix:
-        python-version: [3.7]
+        python-version: [3.8]
        torch: [1.9.0+cu102]
        mmcv: [1.4.2]
        include:
@ -104,26 +125,20 @@ jobs:
            torch_version: torch1.9
            torchvision: 0.10.0+cu102
    steps:
-      - uses: actions/checkout@v2
-      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v2
-        with:
-          python-version: ${{ matrix.python-version }}
+      - uses: actions/checkout@v3
      - name: Install system dependencies
        run: |
          apt-key adv --keyserver keyserver.ubuntu.com --recv-keys A4B469963BF863CC
-          apt-get update && apt-get install -y ffmpeg libsm6 libxext6 git ninja-build libglib2.0-0 libxrender-dev python${{matrix.python-version}}-dev
-          apt-get clean
-          rm -rf /var/lib/apt/lists/*
+          apt-get update && apt-get install -y git
      - name: Install PyTorch
        run: python -m pip install torch==${{matrix.torch}} torchvision==${{matrix.torchvision}} -f https://download.pytorch.org/whl/torch_stable.html
      - name: Install dependencies
        run: |
          python -V
+          python -m pip install -U pip
          python -m pip install mmcv-full==${{matrix.mmcv}} -f https://download.openmmlab.com/mmcv/dist/cu102/${{matrix.torch_version}}/index.html
-          CFLAGS=`python -c 'import sysconfig;print("-I"+sysconfig.get_paths()["include"])'` python -m pip install -r requirements.txt
-          pip install -U pycuda
-          python -m pip install -U numpy
+          python -m pip install -r requirements.txt
+          python -m pip install rapidfuzz==2.15.1
      - name: Build and install
        run: |
          rm -rf .eggs && python -m pip install -e .
@ -135,41 +150,31 @@ jobs:
          coverage report -m

  build_cuda111:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
    container:
      image: pytorch/pytorch:1.8.0-cuda11.1-cudnn8-devel
-
    strategy:
      matrix:
-        python-version: [3.7]
+        python-version: [3.8]
        torch: [1.8.0+cu111]
        mmcv: [1.4.2]
        include:
          - torch: 1.8.0+cu111
            torch_version: torch1.8
            torchvision: 0.9.0+cu111
-
    steps:
-      - uses: actions/checkout@v2
-      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v2
-        with:
-          python-version: ${{ matrix.python-version }}
+      - uses: actions/checkout@v3
      - name: Install system dependencies
        run: |
          apt-key adv --keyserver keyserver.ubuntu.com --recv-keys A4B469963BF863CC
-          apt-get update && apt-get install -y ffmpeg libsm6 libxext6 git ninja-build libglib2.0-0 libxrender-dev python${{matrix.python-version}}-dev
-          apt-get clean
-          rm -rf /var/lib/apt/lists/*
-      - name: Install PyTorch
-        run: python -m pip install torch==${{matrix.torch}} torchvision==${{matrix.torchvision}} -f https://download.pytorch.org/whl/torch_stable.html
+          apt-get update && apt-get install -y git
      - name: Install dependencies
        run: |
          python -V
+          python -m pip install -U pip
          python -m pip install mmcv-full==${{matrix.mmcv}} -f https://download.openmmlab.com/mmcv/dist/cu111/${{matrix.torch_version}}/index.html
-          CFLAGS=`python -c 'import sysconfig;print("-I"+sysconfig.get_paths()["include"])'` python -m pip install -r requirements.txt
-          pip install -U pycuda
-          python -m pip install -U numpy
+          python -m pip install -r requirements.txt
+          python -m pip install rapidfuzz==2.15.1
      - name: Build and install
        run: |
          rm -rf .eggs && python -m pip install -e .
--- a/.github/workflows/java_api.yml
+++ b/.github/workflows/java_api.yml
@ -14,16 +14,19 @@ concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: true

+permissions:
+  contents: read
+
 jobs:
  test_java_api:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
    steps:
      - name: Checkout repository
        uses: actions/checkout@v3
        with:
          submodules: 'recursive'
      - name: Set up Python 3.7
-        uses: actions/setup-python@v2
+        uses: actions/setup-python@v4
        with:
          python-version: 3.7
      - name: Install unittest dependencies
--- a/.github/workflows/lint.yml
+++ b/.github/workflows/lint.yml
@ -2,13 +2,16 @@ name: lint

 on: [push, pull_request]

+permissions:
+  contents: read
+
 jobs:
  lint:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
    steps:
-      - uses: actions/checkout@v2
+      - uses: actions/checkout@v3
      - name: Set up Python 3.7
-        uses: actions/setup-python@v2
+        uses: actions/setup-python@v4
        with:
          python-version: 3.7
      - name: Install pre-commit hook
--- a/.github/workflows/linux-riscv64-gcc.yml
+++ b/.github/workflows/linux-riscv64-gcc.yml
@ -17,6 +17,9 @@ concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: true

+permissions:
+  contents: read
+
 jobs:
  build_riscv64_gcc:
    runs-on: ubuntu-20.04
@ -45,12 +48,9 @@ jobs:
          cmake .. \
            -DCMAKE_TOOLCHAIN_FILE=../cmake/toolchains/riscv64-linux-gnu.cmake \
            -DMMDEPLOY_BUILD_SDK=ON \
-            -DMMDEPLOY_SHARED_LIBS=ON \
            -DMMDEPLOY_BUILD_EXAMPLES=ON \
-            -DMMDEPLOY_TARGET_DEVICES="cpu" \
            -DMMDEPLOY_TARGET_BACKENDS="ncnn" \
            -Dncnn_DIR=$GITHUB_WORKSPACE/ncnn-install/lib/cmake/ncnn/ \
-            -DMMDEPLOY_CODEBASES=all \
            -DOpenCV_DIR=$GITHUB_WORKSPACE/opencv-install/lib/cmake/opencv4
          make -j$(nproc)
          make install
--- a/.github/workflows/prebuild.yml
+++ b/.github/workflows/prebuild.yml
@ -0,0 +1,277 @@
+name: prebuild
+
+on:
+  push:
+    branches:
+      - main
+      - dev-1.x
+      - master
+    paths:
+      - "mmdeploy/version.py"
+
+permissions: write-all
+
+jobs:
+  linux_build:
+    runs-on: [self-hosted, linux-3090]
+    container:
+      image: openmmlab/mmdeploy:manylinux2014_x86_64-cuda11.3
+      options: "--gpus=all --ipc=host"
+      volumes:
+        - /data2/actions-runner/prebuild:/__w/mmdeploy/prebuild
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v3
+        with:
+          submodules: recursive
+      - name: Get mmdeploy version
+        run: |
+          export MMDEPLOY_VERSION=$(python3 -c "import sys; sys.path.append('mmdeploy');from version import __version__;print(__version__)")
+          echo $MMDEPLOY_VERSION
+          echo "MMDEPLOY_VERSION=$MMDEPLOY_VERSION"  >> $GITHUB_ENV
+          echo "OUTPUT_DIR=$MMDEPLOY_VERSION-$GITHUB_RUN_ID" >> $GITHUB_ENV
+      - name: Build MMDeploy
+        run: |
+          source activate mmdeploy-3.6
+          pip install pyyaml packaging setuptools wheel
+          mkdir pack; cd pack
+          python ../tools/package_tools/generate_build_config.py --backend 'trt;ort' \
+            --system linux --output config.yml --build-mmdeploy
+          python ../tools/package_tools/mmdeploy_builder.py --config config.yml
+      - name: Build sdk cpu backend
+        run: |
+          source activate mmdeploy-3.6
+          cd pack
+          python ../tools/package_tools/generate_build_config.py --backend 'ort' \
+            --system linux --output config.yml --device cpu --build-sdk --build-sdk-monolithic \
+            --build-sdk-python --sdk-dynamic-net
+          python ../tools/package_tools/mmdeploy_builder.py --config config.yml
+      - name: Build sdk cuda backend
+        run: |
+          source activate mmdeploy-3.6
+          cd pack
+          python ../tools/package_tools/generate_build_config.py --backend 'ort;trt' \
+            --system linux --output config.yml --device cuda --build-sdk --build-sdk-monolithic \
+            --build-sdk-python --sdk-dynamic-net --onnxruntime-dir=$ONNXRUNTIME_GPU_DIR
+          python ../tools/package_tools/mmdeploy_builder.py --config config.yml
+      - name: Move artifact
+        run: |
+          mkdir -p /__w/mmdeploy/prebuild/$OUTPUT_DIR
+          cp -r pack/* /__w/mmdeploy/prebuild/$OUTPUT_DIR
+
+  linux_build_cxx11abi:
+    runs-on: [self-hosted, linux-3090]
+    container:
+      image: openmmlab/mmdeploy:build-ubuntu16.04-cuda11.3
+      options: "--gpus=all --ipc=host"
+      volumes:
+        - /data2/actions-runner/prebuild:/__w/mmdeploy/prebuild
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v3
+        with:
+          submodules: recursive
+      - name: Get mmdeploy version
+        run: |
+          export MMDEPLOY_VERSION=$(python3 -c "import sys; sys.path.append('mmdeploy');from version import __version__;print(__version__)")
+          echo $MMDEPLOY_VERSION
+          echo "MMDEPLOY_VERSION=$MMDEPLOY_VERSION"  >> $GITHUB_ENV
+          echo "OUTPUT_DIR=$MMDEPLOY_VERSION-$GITHUB_RUN_ID" >> $GITHUB_ENV
+      - name: Build sdk cpu backend
+        run: |
+          mkdir pack; cd pack
+          python ../tools/package_tools/generate_build_config.py --backend 'ort' \
+            --system linux --output config.yml --device cpu --build-sdk --build-sdk-monolithic \
+            --sdk-dynamic-net --cxx11abi
+          python ../tools/package_tools/mmdeploy_builder.py --config config.yml
+      - name: Build sdk cuda backend
+        run: |
+          cd pack
+          python ../tools/package_tools/generate_build_config.py --backend 'ort;trt' \
+            --system linux --output config.yml --device cuda --build-sdk --build-sdk-monolithic \
+            --sdk-dynamic-net --cxx11abi --onnxruntime-dir=$ONNXRUNTIME_GPU_DIR --cudnn-dir /usr
+          python ../tools/package_tools/mmdeploy_builder.py --config config.yml
+      - name: Move artifact
+        run: |
+          mkdir -p /__w/mmdeploy/prebuild/$OUTPUT_DIR
+          cp -r pack/* /__w/mmdeploy/prebuild/$OUTPUT_DIR
+
+  linux_test:
+    runs-on: [self-hosted, linux-3090]
+    needs:
+      - linux_build
+      - linux_build_cxx11abi
+    container:
+      image: openmmlab/mmdeploy:ubuntu20.04-cuda11.3
+      options: "--gpus=all --ipc=host"
+      volumes:
+        - /data2/actions-runner/prebuild:/__w/mmdeploy/prebuild
+        - /data2/actions-runner/testmodel:/__w/mmdeploy/testmodel
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v3
+      - name: Get mmdeploy version
+        run: |
+          export MMDEPLOY_VERSION=$(python3 -c "import sys; sys.path.append('mmdeploy');from version import __version__;print(__version__)")
+          echo $MMDEPLOY_VERSION
+          echo "MMDEPLOY_VERSION=$MMDEPLOY_VERSION"  >> $GITHUB_ENV
+          echo "OUTPUT_DIR=$MMDEPLOY_VERSION-$GITHUB_RUN_ID" >> $GITHUB_ENV
+      - name: Test python
+        run: |
+          cd /__w/mmdeploy/prebuild/$OUTPUT_DIR
+          bash $GITHUB_WORKSPACE/tools/package_tools/test/test_sdk_python.sh
+      - name: Test c/cpp
+        run: |
+          cd /__w/mmdeploy/prebuild/$OUTPUT_DIR
+          bash $GITHUB_WORKSPACE/tools/package_tools/test/test_sdk.sh
+
+  linux_upload:
+    runs-on: [self-hosted, linux-3090]
+    if: startsWith(github.ref, 'refs/tags/')
+    environment: 'prod'
+    needs: linux_test
+    env:
+      PREBUILD_DIR: /data2/actions-runner/prebuild
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v3
+      - name: Get mmdeploy version
+        run: |
+          export MMDEPLOY_VERSION=$(python3 -c "import sys; sys.path.append('mmdeploy');from version import __version__;print(__version__)")
+          echo $MMDEPLOY_VERSION
+          echo "MMDEPLOY_VERSION=$MMDEPLOY_VERSION"  >> $GITHUB_ENV
+          echo "OUTPUT_DIR=$MMDEPLOY_VERSION-$GITHUB_RUN_ID" >> $GITHUB_ENV
+      - name: Upload mmdeploy
+        run: |
+          cd $PREBUILD_DIR/$OUTPUT_DIR/mmdeploy
+          pip install twine
+          # twine upload * --repository testpypi -u __token__ -p ${{ secrets.test_pypi_password }}
+          twine upload * -u __token__ -p ${{ secrets.pypi_password }}
+      - name: Upload mmdeploy_runtime
+        run: |
+          cd $PREBUILD_DIR/$OUTPUT_DIR/mmdeploy_runtime
+          # twine upload * --repository testpypi -u __token__ -p ${{ secrets.test_pypi_password }}
+          twine upload * -u __token__ -p ${{ secrets.pypi_password }}
+      - name: Zip mmdeploy sdk
+        run: |
+          cd $PREBUILD_DIR/$OUTPUT_DIR/sdk
+          for folder in *
+          do
+            tar czf $folder.tar.gz $folder
+          done
+      - name: Upload mmdeploy sdk
+        uses: softprops/action-gh-release@v1
+        with:
+          files: |
+              $PREBUILD_DIR/$OUTPUT_DIR/sdk/*.tar.gz
+
+
+  windows_build:
+    runs-on: [self-hosted, win10-3080]
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v3
+        with:
+          submodules: recursive
+      - name: Get mmdeploy version
+        run: |
+          conda activate mmdeploy-3.8
+          $env:MMDEPLOY_VERSION=(python -c "import sys; sys.path.append('mmdeploy');from version import __version__;print(__version__)")
+          echo $env:MMDEPLOY_VERSION
+          echo "MMDEPLOY_VERSION=$env:MMDEPLOY_VERSION"  >> $env:GITHUB_ENV
+          echo "OUTPUT_DIR=$env:MMDEPLOY_VERSION-$env:GITHUB_RUN_ID" >> $env:GITHUB_ENV
+      - name: Build MMDeploy
+        run: |
+          . D:\DEPS\cienv\prebuild_gpu_env.ps1
+          conda activate mmdeploy-3.6
+          mkdir pack; cd pack
+          python ../tools/package_tools/generate_build_config.py --backend 'trt;ort' `
+            --system windows --output config.yml --build-mmdeploy
+          python ../tools/package_tools/mmdeploy_builder.py --config config.yml
+      - name: Build sdk cpu backend
+        run: |
+          . D:\DEPS\cienv\prebuild_cpu_env.ps1
+          conda activate mmdeploy-3.6
+          cd pack
+          python ../tools/package_tools/generate_build_config.py --backend 'ort' `
+            --system windows --output config.yml --device cpu --build-sdk --build-sdk-monolithic `
+            --build-sdk-python --sdk-dynamic-net
+          python ../tools/package_tools/mmdeploy_builder.py --config config.yml
+      - name: Build sdk cuda backend
+        run: |
+          . D:\DEPS\cienv\prebuild_gpu_env.ps1
+          conda activate mmdeploy-3.6
+          cd pack
+          python ../tools/package_tools/generate_build_config.py --backend 'ort;trt' `
+            --system windows --output config.yml --device cuda --build-sdk --build-sdk-monolithic `
+            --build-sdk-python --sdk-dynamic-net
+          python ../tools/package_tools/mmdeploy_builder.py --config config.yml
+      - name: Move artifact
+        run: |
+          New-Item "D:/DEPS/ciartifact/$env:OUTPUT_DIR" -ItemType Directory -Force
+          Move-Item pack/* "D:/DEPS/ciartifact/$env:OUTPUT_DIR"
+
+  windows_test:
+    runs-on: [self-hosted, win10-3080]
+    needs: windows_build
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v3
+      - name: Get mmdeploy version
+        run: |
+          conda activate mmdeploy-3.8
+          $env:MMDEPLOY_VERSION=(python -c "import sys; sys.path.append('mmdeploy');from version import __version__;print(__version__)")
+          echo $env:MMDEPLOY_VERSION
+          echo "MMDEPLOY_VERSION=$env:MMDEPLOY_VERSION"  >> $env:GITHUB_ENV
+          echo "OUTPUT_DIR=$env:MMDEPLOY_VERSION-$env:GITHUB_RUN_ID" >> $env:GITHUB_ENV
+      - name: Test python
+        run: |
+          cd "D:/DEPS/ciartifact/$env:OUTPUT_DIR"
+          . D:\DEPS\cienv\prebuild_cpu_env.ps1
+          conda activate ci-test
+          & "$env:GITHUB_WORKSPACE/tools/package_tools/test/test_sdk_python.ps1"
+      - name: Test c/cpp
+        run: |
+          cd "D:/DEPS/ciartifact/$env:OUTPUT_DIR"
+          . D:\DEPS\cienv\prebuild_cpu_env.ps1
+          & "$env:GITHUB_WORKSPACE/tools/package_tools/test/test_sdk.ps1"
+
+  windows_upload:
+    runs-on: [self-hosted, win10-3080]
+    if: startsWith(github.ref, 'refs/tags/')
+    environment: 'prod'
+    needs: windows_test
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v3
+      - name: Get mmdeploy version
+        run: |
+          conda activate mmdeploy-3.8
+          $env:MMDEPLOY_VERSION=(python -c "import sys; sys.path.append('mmdeploy');from version import __version__;print(__version__)")
+          echo $env:MMDEPLOY_VERSION
+          echo "MMDEPLOY_VERSION=$env:MMDEPLOY_VERSION"  >> $env:GITHUB_ENV
+          echo "OUTPUT_DIR=$env:MMDEPLOY_VERSION-$env:GITHUB_RUN_ID" >> $env:GITHUB_ENV
+      - name: Upload mmdeploy
+        run: |
+          cd "D:/DEPS/ciartifact/$env:OUTPUT_DIR/mmdeploy"
+          conda activate mmdeploy-3.8
+          # twine upload * --repository testpypi -u __token__ -p ${{ secrets.test_pypi_password }}
+          twine upload * -u __token__ -p ${{ secrets.pypi_password }}
+      - name: Upload mmdeploy_runtime
+        run: |
+          cd "D:/DEPS/ciartifact/$env:OUTPUT_DIR/mmdeploy_runtime"
+          conda activate mmdeploy-3.8
+          # twine upload * --repository testpypi -u __token__ -p ${{ secrets.test_pypi_password }}
+          twine upload * -u __token__ -p ${{ secrets.pypi_password }}
+      - name: Zip mmdeploy sdk
+        run: |
+          cd "D:/DEPS/ciartifact/$env:OUTPUT_DIR/sdk"
+          $folders = $(ls).Name
+          foreach ($folder in $folders) {
+            Compress-Archive -Path $folder -DestinationPath "$folder.zip"
+          }
+      - name: Upload mmdeploy sdk
+        uses: softprops/action-gh-release@v1
+        with:
+          files: |
+              D:/DEPS/ciartifact/$env:OUTPUT_DIR/sdk/*.zip
--- a/.github/workflows/quantize.yml
+++ b/.github/workflows/quantize.yml
@ -16,9 +16,12 @@ concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: true

+permissions:
+  contents: read
+
 jobs:
  test_ncnn_PTQ:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
    container:
      image: pytorch/pytorch:1.8.0-cuda11.1-cudnn8-devel

@ -33,45 +36,31 @@ jobs:
            torchvision: 0.9.0+cu111

    steps:
-      - uses: actions/checkout@v2
-      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v2
-        with:
-          python-version: ${{ matrix.python-version }}
+      - uses: actions/checkout@v3
      - name: Install system dependencies
        run: |
          apt-key adv --keyserver keyserver.ubuntu.com --recv-keys A4B469963BF863CC
-          apt-get update && apt-get install -y wget ffmpeg libsm6 libxext6 git ninja-build libglib2.0-0 libxrender-dev python${{matrix.python-version}}-dev
-          apt-get clean
-          rm -rf /var/lib/apt/lists/*
-      - name: Install PyTorch
-        run: python -m pip install torch==${{matrix.torch}} torchvision==${{matrix.torchvision}} -f https://download.pytorch.org/whl/torch_stable.html
+          apt-get update && apt-get install -y git wget
      - name: Install dependencies
        run: |
          python -V
+          python -m pip install -U pip
          python -m pip install mmcv-full==${{matrix.mmcv}} -f https://download.openmmlab.com/mmcv/dist/cu111/${{matrix.torch_version}}/index.html
-          CFLAGS=`python -c 'import sysconfig;print("-I"+sysconfig.get_paths()["include"])'` python -m pip install -r requirements.txt
-          python -m pip install -U numpy
-
+          python -m pip install -r requirements.txt
+          python -m pip install rapidfuzz==2.15.1
      - name: Install mmcls
        run: |
          cd ~
-          git clone https://github.com/open-mmlab/mmclassification.git
+          git clone -b v0.23.0 --depth 1 https://github.com/open-mmlab/mmclassification.git
          cd mmclassification
-          git checkout v0.23.0
          python3 -m pip install -e .
-          cd -
      - name: Install ppq
        run: |
-          cd ~
-          python -m pip install protobuf==3.20.0
-          git clone https://github.com/openppl-public/ppq
+          git clone -b v0.6.6 --depth 1 https://github.com/openppl-public/ppq
          cd ppq
-          git checkout edbecf44c7b203515640e4f4119c000a1b66b33a
          python3 -m pip install -r requirements.txt
          python3 setup.py install
-          cd -
      - name: Run tests
        run: |
-          echo $(pwd)
+          export PYTHONPATH=${PWD}/ppq:${PYTHONPATH}
          python3 .github/scripts/quantize_to_ncnn.py
--- a/.github/workflows/rust_api.yml
+++ b/.github/workflows/rust_api.yml
@ -14,9 +14,12 @@ concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: true

+permissions:
+  contents: read
+
 jobs:
  test_rust_api:
-    runs-on: ubuntu-18.04
+    runs-on: ubuntu-20.04
    steps:
      - name: Checkout repository
        uses: actions/checkout@v3
@ -24,7 +27,7 @@ jobs:
          submodules: 'recursive'

      - name: Set up Python 3.7
-        uses: actions/setup-python@v2
+        uses: actions/setup-python@v4
        with:
          python-version: 3.7
      - name: Install latest nightly Rust
--- a/.github/workflows/stale.yml
+++ b/.github/workflows/stale.yml
@ -0,0 +1,32 @@
+name: 'Close stale issues and PRs'
+
+on:
+  schedule:
+    # check issue and pull request once at 01:30 a.m. every day
+    - cron: '30 1 * * *'
+
+permissions:
+  contents: read
+
+jobs:
+  stale:
+    permissions:
+      issues: write
+      pull-requests: write
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/stale@v7
+        with:
+          stale-issue-message: 'This issue is marked as stale because it has been marked as invalid or awaiting response for 7 days without any further response. It will be closed in 5 days if the stale label is not removed or if there is no further response.'
+          stale-pr-message: 'This PR is marked as stale because there has been no activity in the past 45 days. It will be closed in 10 days if the stale label is not removed or if there is no further updates.'
+          close-issue-message: 'This issue is closed because it has been stale for 5 days. Please open a new issue if you have similar issues or you have any new updates now.'
+          close-pr-message: 'This PR is closed because it has been stale for 10 days. Please reopen this PR if you have any updates and want to keep contributing the code.'
+          # only issues/PRS with following labels are checked
+          any-of-labels: 'invalid, awaiting response, duplicate'
+          days-before-issue-stale: 7
+          days-before-pr-stale: 45
+          days-before-issue-close: 5
+          days-before-pr-close: 10
+          # automatically remove the stale label when the issues or the pull reqquests are updated or commented
+          remove-stale-when-updated: true
+          operations-per-run: 50
--- a/.gitignore
+++ b/.gitignore
@ -164,3 +164,6 @@ service/snpe/grpc_cpp_plugin
 csrc/mmdeploy/preprocess/elena/json
 csrc/mmdeploy/preprocess/elena/cpu_kernel/*
 csrc/mmdeploy/preprocess/elena/cuda_kernel/*
+
+# c#
+demo/csharp/*/Properties
--- a/.gitmodules
+++ b/.gitmodules
@ -1,9 +1,9 @@
 [submodule "third_party/cub"]
-	path = third_party/cub
-	url = https://github.com/NVIDIA/cub.git
+path = third_party/cub
+url = https://github.com/NVIDIA/cub.git
 [submodule "third_party/pybind11"]
-	path = third_party/pybind11
-	url = https://github.com/pybind/pybind11.git
+path = third_party/pybind11
+url = https://github.com/pybind/pybind11.git
 [submodule "third_party/spdlog"]
-	path = third_party/spdlog
-	url = https://github.com/gabime/spdlog.git
+path = third_party/spdlog
+url = https://github.com/gabime/spdlog.git
--- a/.pre-commit-config.yaml
+++ b/.pre-commit-config.yaml
@ -3,9 +3,11 @@ repos:
    rev: 4.0.1
    hooks:
      - id: flake8
-        args: ["--exclude=*/client/inference_pb2.py,*/client/inference_pb2_grpc.py"]
+        args: ["--exclude=*/client/inference_pb2.py, \
+                */client/inference_pb2_grpc.py, \
+                tools/package_tools/packaging/setup.py"]
  - repo: https://github.com/PyCQA/isort
-    rev: 5.10.1
+    rev: 5.11.5
    hooks:
      - id: isort
  - repo: https://github.com/pre-commit/mirrors-yapf
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@ -5,7 +5,7 @@ endif ()
 message(STATUS "CMAKE_INSTALL_PREFIX: ${CMAKE_INSTALL_PREFIX}")

 cmake_minimum_required(VERSION 3.14)
-project(MMDeploy VERSION 0.9.0)
+project(MMDeploy VERSION 0.14.0)

 set(CMAKE_CXX_STANDARD 17)

@ -22,12 +22,12 @@ endif ()
 set(CMAKE_RUNTIME_OUTPUT_DIRECTORY ${CMAKE_BINARY_DIR}/bin)

 # options
-option(MMDEPLOY_SHARED_LIBS "build shared libs" ON)
+option(MMDEPLOY_SHARED_LIBS "build shared libs" OFF)
 option(MMDEPLOY_BUILD_SDK "build MMDeploy SDK" OFF)
-option(MMDEPLOY_BUILD_SDK_MONOLITHIC "build single lib for SDK API" OFF)
+option(MMDEPLOY_DYNAMIC_BACKEND "dynamic load backend" OFF)
+option(MMDEPLOY_BUILD_SDK_MONOLITHIC "build single lib for SDK API" ON)
 option(MMDEPLOY_BUILD_TEST "build unittests" OFF)
 option(MMDEPLOY_BUILD_SDK_PYTHON_API "build SDK Python API" OFF)
-option(MMDEPLOY_BUILD_SDK_CXX_API "build SDK C++ API" OFF)
 option(MMDEPLOY_BUILD_SDK_CSHARP_API "build SDK C# API support" OFF)
 option(MMDEPLOY_BUILD_SDK_JAVA_API "build SDK JAVA API" OFF)
 option(MMDEPLOY_BUILD_EXAMPLES "build examples" OFF)
@ -40,6 +40,10 @@ set(MMDEPLOY_TARGET_DEVICES "cpu" CACHE STRING "target devices to support")
 set(MMDEPLOY_TARGET_BACKENDS "" CACHE STRING "target inference engines to support")
 set(MMDEPLOY_CODEBASES "all" CACHE STRING "select OpenMMLab codebases")

+if ((NOT MMDEPLOY_BUILD_SDK_MONOLITHIC) AND MMDEPLOY_DYNAMIC_BACKEND)
+    set(MMDEPLOY_DYNAMIC_BACKEND OFF)
+endif ()
+
 if (NOT CMAKE_BUILD_TYPE)
    set(CMAKE_BUILD_TYPE Release CACHE STRING "choose 'Release' as default build type" FORCE)
 endif ()
@ -73,8 +77,6 @@ endif ()

 if (MSVC)
    add_compile_options($<$<COMPILE_LANGUAGE:CXX>:/diagnostics:classic>)
-    add_compile_options($<$<COMPILE_LANGUAGE:CXX>:/Zc:preprocessor>) # /experimental:preprocessor on VS2017
-    add_compile_options($<$<COMPILE_LANGUAGE:CXX>:/Zc:__cplusplus>)
    add_compile_options($<$<COMPILE_LANGUAGE:CXX>:/wd4251>)
 endif ()

@ -97,10 +99,12 @@ include(cmake/MMDeploy.cmake)
 add_subdirectory(csrc/mmdeploy)

 if (MMDEPLOY_BUILD_SDK)
-    install(TARGETS MMDeployStaticModules
-            MMDeployDynamicModules
-            MMDeployLibs
-            EXPORT MMDeployTargets)
+    if (NOT MMDEPLOY_BUILD_SDK_MONOLITHIC)
+        install(TARGETS MMDeployStaticModules
+                MMDeployDynamicModules
+                MMDeployLibs
+                EXPORT MMDeployTargets)
+    endif ()

    if (MMDEPLOY_BUILD_TEST)
        add_subdirectory(tests/test_csrc)
@ -128,6 +132,7 @@ if (MMDEPLOY_BUILD_SDK)
        mmdeploy_add_deps(pplnn BACKENDS ${MMDEPLOY_TARGET_BACKENDS} DEPS pplnn)
    endif ()
    mmdeploy_add_deps(snpe BACKENDS ${MMDEPLOY_TARGET_BACKENDS} DEPS snpe)
+    mmdeploy_add_deps(rknn BACKENDS ${MMDEPLOY_TARGET_BACKENDS} DEPS rknn)

    include(CMakePackageConfigHelpers)
    # generate the config file that is includes the exports
--- a/MANIFEST.in
+++ b/MANIFEST.in
@ -2,7 +2,9 @@ include requirements/*.txt
 include mmdeploy/backend/ncnn/*.so
 include mmdeploy/backend/ncnn/*.dll
 include mmdeploy/backend/ncnn/*.pyd
+include mmdeploy/backend/ncnn/mmdeploy_onnx2ncnn*
 include mmdeploy/lib/*.so
+include mmdeploy/lib/*.so*
 include mmdeploy/lib/*.dll
 include mmdeploy/lib/*.pyd
 include mmdeploy/backend/torchscript/*.so
--- a/README.md
+++ b/README.md
@ -28,6 +28,16 @@

 English | [简体中文](README_zh-CN.md)

+## Highlights
+
+The MMDeploy 1.x has been released, which is adapted to upstream codebases from OpenMMLab 2.0. Please **align the version** when using it.
+The default branch has been switched to `main` from `master`. MMDeploy 0.x (`master`) will be deprecated and new features will only be added to MMDeploy 1.x (`main`) in future.
+
+| mmdeploy | mmengine |   mmcv   |  mmdet   | others |
+| :------: | :------: | :------: | :------: | :----: |
+|  0.x.y   |    -     | \<=1.x.y | \<=2.x.y | 0.x.y  |
+|  1.x.y   |  0.x.y   |  2.x.y   |  3.x.y   | 1.x.y  |
+
 ## Introduction

 MMDeploy is an open-source deep learning model deployment toolset. It is a part of the [OpenMMLab](https://openmmlab.com/) project.
@ -50,6 +60,7 @@ The currently supported codebases and models are as follows, and more will be in
 - [mmpose](docs/en/04-supported-codebases/mmpose.md)
 - [mmdet3d](docs/en/04-supported-codebases/mmdet3d.md)
 - [mmrotate](docs/en/04-supported-codebases/mmrotate.md)
+- [mmaction2](docs/en/04-supported-codebases/mmaction2.md)

 ### Multiple inference backends are available

@ -57,18 +68,22 @@ The supported Device-Platform-InferenceBackend matrix is presented as following,

 The benchmark can be found from [here](docs/en/03-benchmark/benchmark.md)

-| Device / Platform | Linux                                                           | Windows                                 | macOS    | Android          |
-| ----------------- | --------------------------------------------------------------- | --------------------------------------- | -------- | ---------------- |
-| x86_64 CPU        | ✔️ONNX Runtime<br>✔️pplnn<br>✔️ncnn<br>✔️OpenVINO<br>✔️LibTorch | ✔️ONNX Runtime<br>✔️OpenVINO            | -        | -                |
-| ARM CPU           | ✔️ncnn                                                          | -                                       | -        | ✔️ncnn           |
-| RISC-V            | ✔️ncnn                                                          | -                                       | -        | -                |
-| NVIDIA GPU        | ✔️ONNX Runtime<br>✔️TensorRT<br>✔️pplnn<br>✔️LibTorch           | ✔️ONNX Runtime<br>✔️TensorRT<br>✔️pplnn | -        | -                |
-| NVIDIA Jetson     | ✔️TensorRT                                                      | ✔️TensorRT                              | -        | -                |
-| Huawei ascend310  | ✔️CANN                                                          | -                                       | -        | -                |
-| Rockchip          | ✔️RKNN                                                          | -                                       | -        | -                |
-| Apple M1          | -                                                               | -                                       | ✔️CoreML | -                |
-| Adreno GPU        | -                                                               | -                                       | -        | ✔️ncnn<br>✔️SNPE |
-| Hexagon DSP       | -                                                               | -                                       | -        | ✔️SNPE           |
+| Device / Platform | Linux                                                                                                                                                                                                                                                                                                                                                                                           | Windows                                                                                                                       | macOS                                                           | Android                                                                                                              |
+| ----------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------- |
+| x86_64 CPU        | [![Build Status][pass-backend-ort]][ci-backend-ort]ONNXRuntime<br>[![Build Status][pass-backend-pplnn]][ci-backend-pplnn]pplnn<br>[![Build Status][pass-backend-ncnn]][ci-backend-ncnn]ncnn<br>[![Build Status][pass-backend-torchscript]][ci-backend-torchscript]LibTorch<br>[![Build Status][pass-build-rknpu]][ci-build-rknpu]OpenVINO<br>[![Build Status][pass-build-tvm]][ci-build-tvm]TVM | ![][pass-no-status]ONNXRuntime<br>![][pass-no-status]OpenVINO                                                                 | -                                                               | -                                                                                                                    |
+| ARM CPU           | [![Build Status][pass-build-rknpu]][ci-build-rknpu]ncnn                                                                                                                                                                                                                                                                                                                                         | -                                                                                                                             | -                                                               | [![Build Status][pass-build-rknpu]][ci-build-rknpu]ncnn                                                              |
+| RISC-V            | [![Build Status][pass-build-riscv64-gcc]][ci-build-riscv64-gcc]ncnn                                                                                                                                                                                                                                                                                                                             | -                                                                                                                             | -                                                               | -                                                                                                                    |
+| NVIDIA GPU        | ![Build Status][pass-no-status]ONNXRuntime<br>![Build Status][pass-no-status]TensorRT<br>![Build Status][pass-no-status]pplnn<br>![Build Status][pass-no-status]LibTorch<br>![Build Status][pass-no-status]TVM                                                                                                                                                                                  | ![Build Status][pass-no-status]ONNXRuntime<br>![Build Status][pass-no-status]TensorRT<br>![Build Status][pass-no-status]pplnn | -                                                               | -                                                                                                                    |
+| NVIDIA Jetson     | [![Build Status][pass-build-rknpu]][ci-build-rknpu]TensorRT                                                                                                                                                                                                                                                                                                                                     | -                                                                                                                             | -                                                               | -                                                                                                                    |
+| Huawei ascend310  | [![Build Status][pass-backend-ascend]][ci-backend-ascend]CANN                                                                                                                                                                                                                                                                                                                                   | -                                                                                                                             | -                                                               | -                                                                                                                    |
+| Rockchip          | [![Build Status][pass-backend-rknn]][ci-backend-rknn]RKNN                                                                                                                                                                                                                                                                                                                                       | -                                                                                                                             | -                                                               | -                                                                                                                    |
+| Apple M1          | -                                                                                                                                                                                                                                                                                                                                                                                               | -                                                                                                                             | [![Build Status][pass-backend-coreml]][ci-backend-coreml]CoreML | -                                                                                                                    |
+| Adreno GPU        | -                                                                                                                                                                                                                                                                                                                                                                                               | -                                                                                                                             | -                                                               | [![Build Status][pass-backend-snpe]][ci-backend-snpe]SNPE<br>[![Build Status][pass-build-rknpu]][ci-build-rknpu]ncnn |
+| Hexagon DSP       | -                                                                                                                                                                                                                                                                                                                                                                                               | -                                                                                                                             | -                                                               | [![Build Status][pass-backend-snpe]][ci-backend-snpe]SNPE                                                            |
+
+```
+                                                     |
+```

 ### Efficient and scalable C/C++ SDK Framework

@ -82,11 +97,14 @@ Please read [getting_started](docs/en/get_started.md) for the basic usage of MMD
  - [Build from Docker](docs/en/01-how-to-build/build_from_docker.md)
  - [Build from Script](docs/en/01-how-to-build/build_from_script.md)
  - [Build for Linux](docs/en/01-how-to-build/linux-x86_64.md)
+  - [Build for Windows](docs/en/01-how-to-build/windows.md)
  - [Build for macOS](docs/en/01-how-to-build/macos-arm64.md)
  - [Build for Win10](docs/en/01-how-to-build/windows.md)
  - [Build for Android](docs/en/01-how-to-build/android.md)
  - [Build for Jetson](docs/en/01-how-to-build/jetsons.md)
  - [Build for SNPE](docs/en/01-how-to-build/snpe.md)
+  - [Build for Rockchip](docs/en/01-how-to-build/rockchip.md)
+  - [Cross Build for aarch64](docs/en/01-how-to-build/cross_build_ncnn_aarch64.md)
 - User Guide
  - [How to convert model](docs/en/02-how-to-run/convert_model.md)
  - [How to write config](docs/en/02-how-to-run/write_config.md)
@ -148,6 +166,7 @@ This project is released under the [Apache 2.0 license](LICENSE).
 - [MMClassification](https://github.com/open-mmlab/mmclassification): OpenMMLab image classification toolbox and benchmark.
 - [MMDetection](https://github.com/open-mmlab/mmdetection): OpenMMLab detection toolbox and benchmark.
 - [MMDetection3D](https://github.com/open-mmlab/mmdetection3d): OpenMMLab's next-generation platform for general 3D object detection.
+- [MMYOLO](https://github.com/open-mmlab/mmyolo): OpenMMLab YOLO series toolbox and benchmark
 - [MMRotate](https://github.com/open-mmlab/mmrotate): OpenMMLab rotated object detection toolbox and benchmark.
 - [MMSegmentation](https://github.com/open-mmlab/mmsegmentation): OpenMMLab semantic segmentation toolbox and benchmark.
 - [MMOCR](https://github.com/open-mmlab/mmocr): OpenMMLab text detection, recognition, and understanding toolbox.
@ -162,3 +181,27 @@ This project is released under the [Apache 2.0 license](LICENSE).
 - [MMEditing](https://github.com/open-mmlab/mmediting): OpenMMLab image and video editing toolbox.
 - [MMGeneration](https://github.com/open-mmlab/mmgeneration): OpenMMLab image and video generative models toolbox.
 - [MMDeploy](https://github.com/open-mmlab/mmdeploy): OpenMMLab model deployment framework.
+
+[ci-backend-ascend]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-ascend.yml
+[ci-backend-coreml]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-coreml.yml
+[ci-backend-ncnn]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-ncnn.yml
+[ci-backend-ort]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-ort.yml
+[ci-backend-pplnn]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-pplnn.yml
+[ci-backend-rknn]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-rknn.yml
+[ci-backend-snpe]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-snpe.yml
+[ci-backend-torchscript]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-torchscript.yml
+[ci-build-riscv64-gcc]: https://github.com/open-mmlab/mmdeploy/actions/workflows/linux-riscv64-gcc.yml
+[ci-build-rknpu]: https://github.com/open-mmlab/mmdeploy/actions/workflows/linux-rknpu.yml
+[ci-build-tvm]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-tvm.yml
+[pass-backend-ascend]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-ascend.yml?branch=master
+[pass-backend-coreml]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-coreml.yml?branch=master
+[pass-backend-ncnn]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-ncnn.yml?branch=master
+[pass-backend-ort]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-ort.yml?branch=master
+[pass-backend-pplnn]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-pplnn.yml?branch=master
+[pass-backend-rknn]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-rknn.yml?branch=master
+[pass-backend-snpe]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-snpe.yml?branch=master
+[pass-backend-torchscript]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-ort.yml?branch=master
+[pass-build-riscv64-gcc]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/linux-riscv64-gcc.yml?branch=master
+[pass-build-rknpu]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-rknn.yml?branch=master
+[pass-build-tvm]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-tvm.yml?branch=master
+[pass-no-status]: https://img.shields.io/badge/build-no%20status-lightgrey
--- a/README_zh-CN.md
+++ b/README_zh-CN.md
@ -28,6 +28,16 @@

 [English](README.md) | 简体中文

+## MMDeploy 1.x 版本
+
+全新的 MMDeploy 1.x 已发布，该版本适配OpenMMLab 2.0生态体系，使用时务必**对齐版本**。
+MMDeploy 代码库默认分支从`master`切换至`main`。 MMDeploy 0.x (`master`)将逐步废弃，新特性将只添加到 MMDeploy 1.x (`main`)。
+
+| mmdeploy | mmengine |   mmcv   |  mmdet   | mmcls and others |
+| :------: | :------: | :------: | :------: | :--------------: |
+|  0.x.y   |    -     | \<=1.x.y | \<=2.x.y |      0.x.y       |
+|  1.x.y   |  0.x.y   |  2.x.y   |  3.x.y   |      1.x.y       |
+
 ## 介绍

 MMDeploy 是 [OpenMMLab](https://openmmlab.com/) 模型部署工具箱，**为各算法库提供统一的部署体验**。基于 MMDeploy，开发者可以轻松从训练 repo 生成指定硬件所需 SDK，省去大量适配时间。
@ -50,23 +60,28 @@ MMDeploy 是 [OpenMMLab](https://openmmlab.com/) 模型部署工具箱，**为
 - [mmpose](docs/zh_cn/04-supported-codebases/mmpose.md)
 - [mmdet3d](docs/zh_cn/04-supported-codebases/mmdet3d.md)
 - [mmrotate](docs/zh_cn/04-supported-codebases/mmrotate.md)
+- [mmaction2](docs/zh_cn/04-supported-codebases/mmaction2.md)

 ### 支持多种推理后端

 支持的设备平台和推理引擎如下表所示。benchmark请参考[这里](docs/zh_cn/03-benchmark/benchmark.md)

-| Device / Platform | Linux                                                           | Windows                                 | macOS    | Android          |
-| ----------------- | --------------------------------------------------------------- | --------------------------------------- | -------- | ---------------- |
-| x86_64 CPU        | ✔️ONNX Runtime<br>✔️pplnn<br>✔️ncnn<br>✔️OpenVINO<br>✔️LibTorch | ✔️ONNX Runtime<br>✔️OpenVINO            | -        | -                |
-| ARM CPU           | ✔️ncnn                                                          | -                                       | -        | ✔️ncnn           |
-| RISC-V            | ✔️ncnn                                                          | -                                       | -        | -                |
-| NVIDIA GPU        | ✔️ONNX Runtime<br>✔️TensorRT<br>✔️pplnn<br>✔️LibTorch           | ✔️ONNX Runtime<br>✔️TensorRT<br>✔️pplnn | -        | -                |
-| NVIDIA Jetson     | ✔️TensorRT                                                      | ✔️TensorRT                              | -        | -                |
-| Huawei ascend310  | ✔️CANN                                                          | -                                       | -        | -                |
-| Rockchip          | ✔️RKNN                                                          | -                                       | -        | -                |
-| Apple M1          | -                                                               | -                                       | ✔️CoreML | -                |
-| Adreno GPU        | -                                                               | -                                       | -        | ✔️ncnn<br>✔️SNPE |
-| Hexagon DSP       | -                                                               | -                                       | -        | ✔️SNPE           |
+| Device / Platform | Linux                                                                                                                                                                                                                                                                                                                                                                                           | Windows                                                                                                                       | macOS                                                           | Android                                                                                                              |
+| ----------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------- | --------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------- |
+| x86_64 CPU        | [![Build Status][pass-backend-ort]][ci-backend-ort]ONNXRuntime<br>[![Build Status][pass-backend-pplnn]][ci-backend-pplnn]pplnn<br>[![Build Status][pass-backend-ncnn]][ci-backend-ncnn]ncnn<br>[![Build Status][pass-backend-torchscript]][ci-backend-torchscript]LibTorch<br>[![Build Status][pass-build-rknpu]][ci-build-rknpu]OpenVINO<br>[![Build Status][pass-build-tvm]][ci-build-tvm]TVM | ![][pass-no-status]ONNXRuntime<br>![][pass-no-status]OpenVINO                                                                 | -                                                               | -                                                                                                                    |
+| ARM CPU           | [![Build Status][pass-build-rknpu]][ci-build-rknpu]ncnn                                                                                                                                                                                                                                                                                                                                         | -                                                                                                                             | -                                                               | [![Build Status][pass-build-rknpu]][ci-build-rknpu]ncnn                                                              |
+| RISC-V            | [![Build Status][pass-build-riscv64-gcc]][ci-build-riscv64-gcc]ncnn                                                                                                                                                                                                                                                                                                                             | -                                                                                                                             | -                                                               | -                                                                                                                    |
+| NVIDIA GPU        | ![Build Status][pass-no-status]ONNXRuntime<br>![Build Status][pass-no-status]TensorRT<br>![Build Status][pass-no-status]pplnn<br>![Build Status][pass-no-status]LibTorch<br>![Build Status][pass-no-status]TVM                                                                                                                                                                                  | ![Build Status][pass-no-status]ONNXRuntime<br>![Build Status][pass-no-status]TensorRT<br>![Build Status][pass-no-status]pplnn | -                                                               | -                                                                                                                    |
+| NVIDIA Jetson     | [![Build Status][pass-build-rknpu]][ci-build-rknpu]TensorRT                                                                                                                                                                                                                                                                                                                                     | -                                                                                                                             | -                                                               | -                                                                                                                    |
+| Huawei ascend310  | [![Build Status][pass-backend-ascend]][ci-backend-ascend]CANN                                                                                                                                                                                                                                                                                                                                   | -                                                                                                                             | -                                                               | -                                                                                                                    |
+| Rockchip          | [![Build Status][pass-backend-rknn]][ci-backend-rknn]RKNN                                                                                                                                                                                                                                                                                                                                       | -                                                                                                                             | -                                                               | -                                                                                                                    |
+| Apple M1          | -                                                                                                                                                                                                                                                                                                                                                                                               | -                                                                                                                             | [![Build Status][pass-backend-coreml]][ci-backend-coreml]CoreML | -                                                                                                                    |
+| Adreno GPU        | -                                                                                                                                                                                                                                                                                                                                                                                               | -                                                                                                                             | -                                                               | [![Build Status][pass-backend-snpe]][ci-backend-snpe]SNPE<br>[![Build Status][pass-build-rknpu]][ci-build-rknpu]ncnn |
+| Hexagon DSP       | -                                                                                                                                                                                                                                                                                                                                                                                               | -                                                                                                                             | -                                                               | [![Build Status][pass-backend-snpe]][ci-backend-snpe]SNPE                                                            |
+
+```
+                                                     |
+```

 ### SDK 可高度定制化

@ -81,11 +96,14 @@ MMDeploy 是 [OpenMMLab](https://openmmlab.com/) 模型部署工具箱，**为
  - [一键式脚本安装](docs/zh_cn/01-how-to-build/build_from_script.md)
  - [Build from Docker](docs/zh_cn/01-how-to-build/build_from_docker.md)
  - [Build for Linux](docs/zh_cn/01-how-to-build/linux-x86_64.md)
+  - [Build for Windows](docs/zh_cn/01-how-to-build/windows.md)
  - [Build for macOS](docs/zh_cn/01-how-to-build/macos-arm64.md)
  - [Build for Win10](docs/zh_cn/01-how-to-build/windows.md)
  - [Build for Android](docs/zh_cn/01-how-to-build/android.md)
  - [Build for Jetson](docs/zh_cn/01-how-to-build/jetsons.md)
  - [Build for SNPE](docs/zh_cn/01-how-to-build/snpe.md)
+  - [Build for Rockchip](docs/zh_cn/01-how-to-build/rockchip.md)
+  - [Cross Build for aarch64](docs/zh_cn/01-how-to-build/cross_build_ncnn_aarch64.md)
 - 使用
  - [把模型转换到推理 Backend](docs/zh_cn/02-how-to-run/convert_model.md)
  - [配置转换参数](docs/zh_cn/02-how-to-run/write_config.md)
@ -153,6 +171,7 @@ MMDeploy 是 [OpenMMLab](https://openmmlab.com/) 模型部署工具箱，**为
 - [MMClassification](https://github.com/open-mmlab/mmclassification): OpenMMLab 图像分类工具箱
 - [MMDetection](https://github.com/open-mmlab/mmdetection): OpenMMLab 目标检测工具箱
 - [MMDetection3D](https://github.com/open-mmlab/mmdetection3d): OpenMMLab 新一代通用 3D 目标检测平台
+- [MMYOLO](https://github.com/open-mmlab/mmyolo): OpenMMLab YOLO 系列工具箱和基准测试
 - [MMRotate](https://github.com/open-mmlab/mmrotate): OpenMMLab 旋转框检测工具箱与测试基准
 - [MMSegmentation](https://github.com/open-mmlab/mmsegmentation): OpenMMLab 语义分割工具箱
 - [MMOCR](https://github.com/open-mmlab/mmocr): OpenMMLab 全流程文字检测识别理解工具包
@ -188,3 +207,27 @@ MMDeploy 是 [OpenMMLab](https://openmmlab.com/) 模型部署工具箱，**为
 - 🔥 提供与各行各业开发者充分交流的平台

 干货满满 📘，等您来撩 💗，OpenMMLab 社区期待您的加入 👬
+
+[ci-backend-ascend]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-ascend.yml
+[ci-backend-coreml]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-coreml.yml
+[ci-backend-ncnn]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-ncnn.yml
+[ci-backend-ort]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-ort.yml
+[ci-backend-pplnn]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-pplnn.yml
+[ci-backend-rknn]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-rknn.yml
+[ci-backend-snpe]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-snpe.yml
+[ci-backend-torchscript]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-torchscript.yml
+[ci-build-riscv64-gcc]: https://github.com/open-mmlab/mmdeploy/actions/workflows/linux-riscv64-gcc.yml
+[ci-build-rknpu]: https://github.com/open-mmlab/mmdeploy/actions/workflows/linux-rknpu.yml
+[ci-build-tvm]: https://github.com/open-mmlab/mmdeploy/actions/workflows/backend-tvm.yml
+[pass-backend-ascend]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-ascend.yml?branch=master
+[pass-backend-coreml]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-coreml.yml?branch=master
+[pass-backend-ncnn]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-ncnn.yml?branch=master
+[pass-backend-ort]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-ort.yml?branch=master
+[pass-backend-pplnn]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-pplnn.yml?branch=master
+[pass-backend-rknn]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-rknn.yml?branch=master
+[pass-backend-snpe]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-snpe.yml?branch=master
+[pass-backend-torchscript]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-ort.yml?branch=master
+[pass-build-riscv64-gcc]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/linux-riscv64-gcc.yml?branch=master
+[pass-build-rknpu]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-rknn.yml?branch=master
+[pass-build-tvm]: https://img.shields.io/github/actions/workflow/status/open-mmlab/mmdeploy/backend-tvm.yml?branch=master
+[pass-no-status]: https://img.shields.io/badge/build-no%20status-lightgrey
--- a/cmake/MMDeploy.cmake
+++ b/cmake/MMDeploy.cmake
@ -1,6 +1,6 @@
 # Copyright (c) OpenMMLab. All rights reserved.

-function (mmdeploy_export NAME)
+function (mmdeploy_export_impl NAME)
    set(_LIB_DIR lib)
    if (MSVC)
        set(_LIB_DIR bin)
@ -12,6 +12,24 @@ function (mmdeploy_export NAME)
            RUNTIME DESTINATION bin)
 endfunction ()

+macro(mmdeploy_add_net NAME)
+    if (MMDEPLOY_DYNAMIC_BACKEND)
+        mmdeploy_add_library(${NAME} SHARED ${ARGN})
+        # DYNAMIC_BACKEND implies BUILD_SDK_MONOLITHIC
+        mmdeploy_export_impl(${NAME})
+        target_link_libraries(${PROJECT_NAME} PRIVATE mmdeploy)
+        set(BACKEND_LIB_NAMES ${BACKEND_LIB_NAMES} ${PROJECT_NAME} PARENT_SCOPE)
+    else ()
+        mmdeploy_add_module(${NAME} ${ARGN})
+    endif ()
+endmacro()
+
+function (mmdeploy_export NAME)
+    if (NOT MMDEPLOY_BUILD_SDK_MONOLITHIC)
+        mmdeploy_export_impl(${NAME})
+    endif ()
+endfunction ()
+

 function (mmdeploy_add_library NAME)
    # EXCLUDE: exclude from registering & exporting
--- a/cmake/MMDeployConfig.cmake.in
+++ b/cmake/MMDeployConfig.cmake.in
@ -10,13 +10,12 @@ set(MMDEPLOY_TARGET_DEVICES @MMDEPLOY_TARGET_DEVICES@)
 set(MMDEPLOY_TARGET_BACKENDS @MMDEPLOY_TARGET_BACKENDS@)
 set(MMDEPLOY_BUILD_TYPE @CMAKE_BUILD_TYPE@)
 set(MMDEPLOY_BUILD_SHARED @MMDEPLOY_SHARED_LIBS@)
-set(MMDEPLOY_BUILD_SDK_CXX_API @MMDEPLOY_BUILD_SDK_CXX_API@)
 set(MMDEPLOY_BUILD_SDK_MONOLITHIC @MMDEPLOY_BUILD_SDK_MONOLITHIC@)
 set(MMDEPLOY_VERSION_MAJOR @MMDEPLOY_VERSION_MAJOR@)
 set(MMDEPLOY_VERSION_MINOR @MMDEPLOY_VERSION_MINOR@)
 set(MMDEPLOY_VERSION_PATCH @MMDEPLOY_VERSION_PATCH@)

-if (NOT MMDEPLOY_BUILD_SHARED)
+if (NOT MMDEPLOY_BUILD_SHARED AND NOT MMDEPLOY_BUILD_SDK_MONOLITHIC)
    if ("cuda" IN_LIST MMDEPLOY_TARGET_DEVICES)
        find_package(CUDA REQUIRED)
        if(MSVC)
--- a/cmake/cuda.cmake
+++ b/cmake/cuda.cmake
@ -11,6 +11,37 @@ if (MSVC OR (NOT DEFINED CMAKE_CUDA_RUNTIME_LIBRARY))
    set(CUDA_USE_STATIC_CUDA_RUNTIME OFF)
 endif ()

+if (MSVC)
+    # no plugin in BuildCustomizations and no specify cuda toolset
+    if (NOT CMAKE_VS_PLATFORM_TOOLSET_CUDA)
+        message(FATAL_ERROR "Please install CUDA MSBuildExtensions")
+    endif ()
+
+    if (CMAKE_VS_PLATFORM_TOOLSET_CUDA_CUSTOM_DIR)
+        # find_package(CUDA) required ENV{CUDA_PATH}
+        set(ENV{CUDA_PATH} ${CMAKE_VS_PLATFORM_TOOLSET_CUDA_CUSTOM_DIR})
+    else ()
+        # we use CUDA_PATH and ignore nvcc.exe
+        # cmake will import highest cuda props version, which may not equal to CUDA_PATH
+        if (NOT (DEFINED ENV{CUDA_PATH}))
+            message(FATAL_ERROR "Please set CUDA_PATH environment variable")
+        endif ()
+
+        string(REGEX REPLACE ".*v([0-9]+)\\..*" "\\1" _MAJOR $ENV{CUDA_PATH})
+        string(REGEX REPLACE ".*v[0-9]+\\.([0-9]+).*" "\\1" _MINOR $ENV{CUDA_PATH})
+        if (NOT (${CMAKE_VS_PLATFORM_TOOLSET_CUDA} STREQUAL "${_MAJOR}.${_MINOR}"))
+            message(FATAL_ERROR "Auto detected cuda version ${CMAKE_VS_PLATFORM_TOOLSET_CUDA}"
+                " is mismatch with ENV{CUDA_PATH} $ENV{CUDA_PATH}. Please modify CUDA_PATH"
+                " to match ${CMAKE_VS_PLATFORM_TOOLSET_CUDA} or specify cuda toolset by"
+                " cmake -T cuda=/path/to/cuda ..")
+        endif ()
+
+        if (NOT (DEFINED ENV{CUDA_PATH_V${_MAJOR}_${_MINOR}}))
+            message(FATAL_ERROR "Please set CUDA_PATH_V${_MAJOR}_${_MINOR} environment variable")
+        endif ()
+    endif ()
+endif ()
+
 # nvcc compiler settings
 find_package(CUDA REQUIRED)

@ -42,6 +73,7 @@ if (NOT CMAKE_CUDA_ARCHITECTURES)
    if (CUDA_VERSION_MAJOR VERSION_GREATER_EQUAL "8")
        set(_NVCC_FLAGS "${_NVCC_FLAGS} -gencode arch=compute_60,code=sm_60")
        set(_NVCC_FLAGS "${_NVCC_FLAGS} -gencode arch=compute_61,code=sm_61")
+        set(_NVCC_FLAGS "${_NVCC_FLAGS} -gencode arch=compute_62,code=sm_62")
    endif ()
    if (CUDA_VERSION_MAJOR VERSION_GREATER_EQUAL "9")
        set(_NVCC_FLAGS "${_NVCC_FLAGS} -gencode arch=compute_70,code=sm_70")
--- a/cmake/loader.cpp.in
+++ b/cmake/loader.cpp.in
@ -1,15 +1,53 @@
 // Copyright (c) OpenMMLab. All rights reserved.

-#include <Windows.h>
-
+#include <string>
 #include <cstdio>

+#ifdef _WIN32
+#include <Windows.h>
+#else
+#include <dlfcn.h>
+#endif
+
+#ifdef _WIN32
+#define LIBPREFIX ""
+#define LIBSUFFIX ".dll"
+#elif defined(__APPLE__)
+#define LIBPREFIX "lib"
+#define LIBSUFFIX ".dylib"
+#else
+#define LIBPREFIX "lib"
+#define LIBSUFFIX ".so"
+#endif
+
 namespace mmdeploy {
 namespace {

+#ifdef _WIN32
+inline static const std::wstring GetDllPath() {
+    HMODULE hm = NULL;
+    GetModuleHandleExW(GET_MODULE_HANDLE_EX_FLAG_FROM_ADDRESS | GET_MODULE_HANDLE_EX_FLAG_UNCHANGED_REFCOUNT,
+      (LPWSTR)&GetDllPath, &hm);
+    std::wstring ret;
+    ret.resize(MAX_PATH);
+    GetModuleFileNameW(hm, &ret[0], ret.size());
+    ret = ret.substr(0, ret.find_last_of(L"/\\"));
+    return ret;
+}
+#endif
+
 void* mmdeploy_load_library(const char* name) {
  fprintf(stderr, "loading %s ...\n", name);
-  auto handle = LoadLibraryA(name);
+
+#ifdef _WIN32
+  auto handle = LoadLibraryExA(name, NULL, LOAD_LIBRARY_SEARCH_USER_DIRS);
+  if (handle == NULL) {
+    handle = LoadLibraryExA(name, NULL, LOAD_WITH_ALTERED_SEARCH_PATH);
+  }
+#else
+  auto handle = dlopen(name, RTLD_NOW | RTLD_GLOBAL);
+#endif
+
  if (!handle) {
    fprintf(stderr, "failed to load library %s\n", name);
    return nullptr;
@ -22,11 +60,15 @@ void* mmdeploy_load_library(const char* name) {
 class Loader {
 public:
  Loader() {
+#ifdef _WIN32
+    AddDllDirectory(GetDllPath().c_str());
+#endif
    const char* modules[] = {
        @_MMDEPLOY_DYNAMIC_MODULES@
    };
    for (const auto name : modules) {
-      mmdeploy_load_library(name);
+      std::string libname = std::string{} + LIBPREFIX + name + LIBSUFFIX;
+      mmdeploy_load_library(libname.c_str());
    }
  }
 };
--- a/cmake/modules/FindTVM.cmake
+++ b/cmake/modules/FindTVM.cmake
@ -0,0 +1,47 @@
+# Copyright (c) OpenMMLab. All rights reserved.
+
+if (NOT DEFINED TVM_DIR)
+    set(TVM_DIR $ENV{TVM_DIR})
+endif ()
+if (NOT TVM_DIR)
+    message(FATAL_ERROR "Please set TVM_DIR with cmake -D option.")
+endif()
+
+find_path(
+    TVM_INCLUDE_DIR tvm/runtime/c_runtime_api.h
+    HINTS ${TVM_DIR}
+    PATH_SUFFIXES include)
+
+find_path(
+    DMLC_CORE_INCLUDE_DIR  dmlc/io.h
+    HINTS ${TVM_DIR}/3rdparty/dmlc-core
+    PATH_SUFFIXES include)
+
+find_path(
+    DLPACK_INCLUDE_DIR dlpack/dlpack.h
+    HINTS ${TVM_DIR}/3rdparty/dlpack
+    PATH_SUFFIXES include)
+
+find_library(
+    TVM_LIBRARY_PATH tvm_runtime
+    HINTS ${TVM_DIR}
+    PATH_SUFFIXES build lib build/${CMAKE_BUILD_TYPE})
+if (NOT (TVM_INCLUDE_DIR AND DMLC_CORE_INCLUDE_DIR AND DLPACK_INCLUDE_DIR AND TVM_LIBRARY_PATH))
+    message(FATAL_ERROR "Couldn't find tvm in TVM_DIR: "
+        "${TVM_DIR}, please check if the path is correct.")
+endif()
+
+add_library(tvm_runtime SHARED IMPORTED)
+set_property(TARGET tvm_runtime APPEND PROPERTY IMPORTED_CONFIGURATIONS RELEASE)
+if (MSVC)
+    set_target_properties(tvm_runtime PROPERTIES
+        IMPORTED_IMPLIB_RELEASE ${TVM_LIBRARY_PATH}
+        INTERFACE_INCLUDE_DIRECTORIES ${TVM_INCLUDE_DIR} ${DMLC_CORE_INCLUDE_DIR} ${DLPACK_INCLUDE_DIR}
+    )
+
+else()
+    set_target_properties(tvm_runtime PROPERTIES
+        IMPORTED_LOCATION_RELEASE ${TVM_LIBRARY_PATH}
+        INTERFACE_INCLUDE_DIRECTORIES ${TVM_INCLUDE_DIR} ${DMLC_CORE_INCLUDE_DIR} ${DLPACK_INCLUDE_DIR}
+    )
+endif()
--- a/cmake/tensorrt.cmake
+++ b/cmake/tensorrt.cmake
@ -29,6 +29,7 @@ else ()
    message(FATAL_ERROR "Cannot find TensorRT libs")
 endif ()

+include(FindPackageHandleStandardArgs)
 find_package_handle_standard_args(TENSORRT DEFAULT_MSG TENSORRT_INCLUDE_DIR
        TENSORRT_LIBRARY)
 if (NOT TENSORRT_FOUND)
--- a/cmake/toolchains/aarch64-linux-gnu.cmake
+++ b/cmake/toolchains/aarch64-linux-gnu.cmake
@ -0,0 +1,17 @@
+set(CMAKE_SYSTEM_NAME Linux)
+set(CMAKE_SYSTEM_PROCESSOR aarch64)
+
+set(CMAKE_C_COMPILER "aarch64-linux-gnu-gcc")
+set(CMAKE_CXX_COMPILER "aarch64-linux-gnu-g++")
+set(CMAKE_LINKER "aarch64-linux-gnu-ld")
+
+set(CMAKE_FIND_ROOT_PATH_MODE_PROGRAM NEVER)
+set(CMAKE_FIND_ROOT_PATH_MODE_LIBRARY ONLY)
+set(CMAKE_FIND_ROOT_PATH_MODE_INCLUDE ONLY)
+
+set(CMAKE_C_FLAGS "-march=armv8-a")
+set(CMAKE_CXX_FLAGS "-march=armv8-a")
+
+# cache flags
+set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS}" CACHE STRING "c flags")
+set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS}" CACHE STRING "c++ flags")
--- a/cmake/toolchains/arm-linux-gnueabihf.cmake
+++ b/cmake/toolchains/arm-linux-gnueabihf.cmake
@ -0,0 +1,16 @@
+set(CMAKE_SYSTEM_NAME Linux)
+set(CMAKE_SYSTEM_PROCESSOR arm)
+
+set(CMAKE_C_COMPILER "arm-linux-gnueabihf-gcc")
+set(CMAKE_CXX_COMPILER "arm-linux-gnueabihf-g++")
+
+set(CMAKE_FIND_ROOT_PATH_MODE_PROGRAM NEVER)
+set(CMAKE_FIND_ROOT_PATH_MODE_LIBRARY ONLY)
+set(CMAKE_FIND_ROOT_PATH_MODE_INCLUDE ONLY)
+
+set(CMAKE_C_FLAGS "-march=armv7-a -mfloat-abi=hard -mfpu=neon")
+set(CMAKE_CXX_FLAGS "-march=armv7-a -mfloat-abi=hard -mfpu=neon")
+
+# cache flags
+set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS}" CACHE STRING "c flags")
+set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS}" CACHE STRING "c++ flags")
--- a/cmake/toolchains/rknpu2-linux-gnu.cmake
+++ b/cmake/toolchains/rknpu2-linux-gnu.cmake
@ -0,0 +1,23 @@
+set(CMAKE_SYSTEM_NAME Linux)
+set(CMAKE_SYSTEM_PROCESSOR rockchip)
+
+if(DEFINED ENV{RKNN_TOOL_CHAIN})
+    file(TO_CMAKE_PATH $ENV{RKNN_TOOL_CHAIN} RKNN_TOOL_CHAIN)
+else()
+    message(FATAL_ERROR "RKNN_TOOL_CHAIN env must be defined")
+endif()
+
+set(CMAKE_C_COMPILER ${RKNN_TOOL_CHAIN}/bin/aarch64-rockchip-linux-gnu-gcc)
+set(CMAKE_CXX_COMPILER ${RKNN_TOOL_CHAIN}/bin/aarch64-rockchip-linux-gnu-g++)
+
+set(CMAKE_FIND_ROOT_PATH_MODE_PROGRAM NEVER)
+set(CMAKE_FIND_ROOT_PATH_MODE_LIBRARY ONLY)
+set(CMAKE_FIND_ROOT_PATH_MODE_INCLUDE ONLY)
+set(CMAKE_FIND_ROOT_PATH_MODE_PACKAGE ONLY)
+
+set(CMAKE_C_FLAGS "-Wl,--allow-shlib-undefined")
+set(CMAKE_CXX_FLAGS "-Wl,--allow-shlib-undefined")
+
+# cache flags
+set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS}" CACHE STRING "c flags")
+set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS}" CACHE STRING "c++ flags")
--- a/configs/_base_/backends/coreml.py
+++ b/configs/_base_/backends/coreml.py
@ -1 +1,11 @@
-backend_config = dict(type='coreml', convert_to='mlprogram')
+backend_config = dict(
+    type='coreml',
+    # mlprogram or neuralnetwork
+    convert_to='mlprogram',
+    common_config=dict(
+        # FLOAT16 or FLOAT32, see coremltools.precision
+        compute_precision='FLOAT32',
+        # iOS15, iOS16, etc., see coremltools.target
+        minimum_deployment_target='iOS16',
+        skip_model_load=False),
+)
--- a/configs/_base_/backends/rknn.py
+++ b/configs/_base_/backends/rknn.py
@ -1,8 +1,6 @@
 backend_config = dict(
    type='rknn',
    common_config=dict(
-        mean_values=None,
-        std_values=None,
-        target_platform='rk3588',
-        optimization_level=3),
-    quantization_config=dict(do_quantization=False, dataset=None))
+        target_platform='rv1126',  # 'rk3588'
+        optimization_level=1),
+    quantization_config=dict(do_quantization=True, dataset=None))
--- a/configs/_base_/backends/tvm.py
+++ b/configs/_base_/backends/tvm.py
@ -0,0 +1 @@
+backend_config = dict(type='tvm')
--- a/configs/mmaction/video-recognition/video-recognition_2d_dynamic.py
+++ b/configs/mmaction/video-recognition/video-recognition_2d_dynamic.py
@ -0,0 +1,15 @@
+_base_ = ['./video-recognition_static.py']
+
+onnx_config = dict(
+    dynamic_axes={
+        'input': {
+            0: 'batch',
+            1: 'num_crops * num_segs',
+            3: 'height',
+            4: 'width'
+        },
+        'output': {
+            0: 'batch',
+        }
+    },
+    input_shape=None)
--- a/configs/mmaction/video-recognition/video-recognition_2d_tensorrt_static-224x224.py
+++ b/configs/mmaction/video-recognition/video-recognition_2d_tensorrt_static-224x224.py
@ -0,0 +1,14 @@
+_base_ = ['./video-recognition_static.py', '../../_base_/backends/tensorrt.py']
+
+onnx_config = dict(input_shape=[224, 224])
+
+backend_config = dict(
+    common_config=dict(max_workspace_size=1 << 30),
+    model_inputs=[
+        dict(
+            input_shapes=dict(
+                input=dict(
+                    min_shape=[1, 250, 3, 224, 224],
+                    opt_shape=[1, 250, 3, 224, 224],
+                    max_shape=[1, 250, 3, 224, 224])))
+    ])
--- a/configs/mmaction/video-recognition/video-recognition_3d_dynamic.py
+++ b/configs/mmaction/video-recognition/video-recognition_3d_dynamic.py
@ -0,0 +1,16 @@
+_base_ = ['./video-recognition_static.py']
+
+onnx_config = dict(
+    dynamic_axes={
+        'input': {
+            0: 'batch',
+            1: 'num_crops * num_segs',
+            3: 'time',
+            4: 'height',
+            5: 'width'
+        },
+        'output': {
+            0: 'batch',
+        }
+    },
+    input_shape=None)
--- a/configs/mmaction/video-recognition/video-recognition_3d_tensorrt_static-256x256.py
+++ b/configs/mmaction/video-recognition/video-recognition_3d_tensorrt_static-256x256.py
@ -0,0 +1,14 @@
+_base_ = ['./video-recognition_static.py', '../../_base_/backends/tensorrt.py']
+
+onnx_config = dict(input_shape=[256, 256])
+
+backend_config = dict(
+    common_config=dict(max_workspace_size=1 << 30),
+    model_inputs=[
+        dict(
+            input_shapes=dict(
+                input=dict(
+                    min_shape=[1, 30, 3, 32, 256, 256],
+                    opt_shape=[1, 30, 3, 32, 256, 256],
+                    max_shape=[1, 30, 3, 32, 256, 256])))
+    ])
--- a/configs/mmaction/video-recognition/video-recognition_onnxruntime_static.py
+++ b/configs/mmaction/video-recognition/video-recognition_onnxruntime_static.py
@ -0,0 +1,5 @@
+_base_ = [
+    './video-recognition_static.py', '../../_base_/backends/onnxruntime.py'
+]
+
+onnx_config = dict(input_shape=None)
--- a/configs/mmaction/video-recognition/video-recognition_sdk.py
+++ b/configs/mmaction/video-recognition/video-recognition_sdk.py
@ -0,0 +1,19 @@
+_base_ = ['./video-recognition_static.py', '../../_base_/backends/sdk.py']
+
+codebase_config = dict(model_type='sdk')
+
+# It will read SampleFrames from model_cfg to info
+# in pipeline of backend_config
+
+backend_config = dict(pipeline=[
+    dict(type='OpenCVInit', num_threads=1),
+    dict(
+        type='SampleFrames',
+        clip_len=1,
+        frame_interval=1,
+        num_clips=25,
+        test_mode=True),
+    dict(type='OpenCVDecode'),
+    dict(type='Collect', keys=['imgs'], meta_keys=[]),
+    dict(type='ListToNumpy', keys=['imgs'])
+])
--- a/configs/mmaction/video-recognition/video-recognition_static.py
+++ b/configs/mmaction/video-recognition/video-recognition_static.py
@ -0,0 +1,3 @@
+_base_ = ['../../_base_/onnx_config.py']
+
+codebase_config = dict(type='mmaction', task='VideoRecognition')
--- a/configs/mmcls/classification_coreml_static-384x384.py
+++ b/configs/mmcls/classification_coreml_static-384x384.py
@ -0,0 +1,11 @@
+_base_ = ['./classification_coreml_dynamic-224x224-224x224.py']
+
+ir_config = dict(input_shape=(384, 384))
+backend_config = dict(model_inputs=[
+    dict(
+        input_shapes=dict(
+            input=dict(
+                min_shape=[1, 3, 384, 384],
+                max_shape=[1, 3, 384, 384],
+                default_shape=[1, 3, 384, 384])))
+])
--- a/configs/mmcls/classification_rknn-fp16_static-224x224.py
+++ b/configs/mmcls/classification_rknn-fp16_static-224x224.py
@ -0,0 +1,7 @@
+_base_ = ['./classification_static.py', '../_base_/backends/rknn.py']
+
+onnx_config = dict(input_shape=[224, 224])
+codebase_config = dict(model_type='end2end')
+backend_config = dict(
+    input_size_list=[[3, 224, 224]],
+    quantization_config=dict(do_quantization=False))
--- a/configs/mmcls/classification_rknn-int8_static-224x224.py
+++ b/configs/mmcls/classification_rknn-int8_static-224x224.py
@ -1,5 +1,5 @@
 _base_ = ['./classification_static.py', '../_base_/backends/rknn.py']

 onnx_config = dict(input_shape=[224, 224])
-codebase_config = dict(model_type='rknn')
+codebase_config = dict(model_type='end2end')
 backend_config = dict(input_size_list=[[3, 224, 224]])
--- a/configs/mmcls/classification_tvm-ansor_static-224x224.py
+++ b/configs/mmcls/classification_tvm-ansor_static-224x224.py
@ -0,0 +1,12 @@
+_base_ = ['./classification_static.py', '../_base_/backends/tvm.py']
+
+onnx_config = dict(input_shape=[224, 224])
+backend_config = dict(model_inputs=[
+    dict(
+        shape=dict(input=[1, 3, 224, 224]),
+        dtype=dict(input='float32'),
+        tuner=dict(
+            type='AutoScheduleTuner',
+            log_file='tvm_tune_log.log',
+            num_measure_trials=2000))
+])
--- a/configs/mmcls/classification_tvm-autotvm-int8_static-224x224.py
+++ b/configs/mmcls/classification_tvm-autotvm-int8_static-224x224.py
@ -0,0 +1,16 @@
+_base_ = ['./classification_tvm-autotvm_static-224x224.py']
+
+calib_config = dict(create_calib=True, calib_file='calib_data.h5')
+backend_config = dict(model_inputs=[
+    dict(
+        shape=dict(input=[1, 3, 224, 224]),
+        dtype=dict(input='float32'),
+        tuner=dict(
+            type='AutoTVMTuner',
+            log_file='tvm_tune_log.log',
+            n_trial=1000,
+            tuner=dict(type='XGBTuner'),
+        ),
+        qconfig=dict(calibrate_mode='kl_divergence', weight_scale='max'),
+    )
+])
--- a/configs/mmcls/classification_tvm-autotvm_static-224x224.py
+++ b/configs/mmcls/classification_tvm-autotvm_static-224x224.py
@ -0,0 +1,13 @@
+_base_ = ['./classification_static.py', '../_base_/backends/tvm.py']
+
+onnx_config = dict(input_shape=[224, 224])
+backend_config = dict(model_inputs=[
+    dict(
+        shape=dict(input=[1, 3, 224, 224]),
+        dtype=dict(input='float32'),
+        tuner=dict(
+            type='AutoTVMTuner',
+            log_file='tvm_tune_log.log',
+            n_trial=1000,
+            tuner=dict(type='XGBTuner')))
+])
--- a/configs/mmdet/_base_/base_instance-seg_coreml_static-800x1344.py
+++ b/configs/mmdet/_base_/base_instance-seg_coreml_static-800x1344.py
@ -0,0 +1,16 @@
+_base_ = ['./base_torchscript.py', '../../_base_/backends/coreml.py']
+
+ir_config = dict(
+    input_shape=(1344, 800), output_names=['dets', 'labels', 'masks'])
+
+backend_config = dict(model_inputs=[
+    dict(
+        input_shapes=dict(
+            input=dict(
+                min_shape=[1, 3, 800, 1344],
+                max_shape=[1, 3, 800, 1344],
+                default_shape=[1, 3, 800, 1344])))
+])
+
+# Don't know if this is necessary
+codebase_config = dict(post_processing=dict(export_postprocess_mask=False))
--- a/configs/mmdet/detection/detection_coreml_static-608x608.py
+++ b/configs/mmdet/detection/detection_coreml_static-608x608.py
@ -0,0 +1,11 @@
+_base_ = ['../_base_/base_torchscript.py', '../../_base_/backends/coreml.py']
+
+ir_config = dict(input_shape=(608, 608))
+backend_config = dict(model_inputs=[
+    dict(
+        input_shapes=dict(
+            input=dict(
+                min_shape=[1, 3, 608, 608],
+                max_shape=[1, 3, 608, 608],
+                default_shape=[1, 3, 608, 608])))
+])
--- a/configs/mmdet/detection/detection_rknn-fp16_static-320x320.py
+++ b/configs/mmdet/detection/detection_rknn-fp16_static-320x320.py
@ -0,0 +1,34 @@
+_base_ = ['../_base_/base_static.py', '../../_base_/backends/rknn.py']
+
+onnx_config = dict(input_shape=[320, 320])
+
+codebase_config = dict(model_type='rknn')
+
+backend_config = dict(
+    input_size_list=[[3, 320, 320]],
+    quantization_config=dict(do_quantization=False))
+
+# # yolov3, yolox for rknn-toolkit and rknn-toolkit2
+# partition_config = dict(
+#     type='rknn',  # the partition policy name
+#     apply_marks=True,  # should always be set to True
+#     partition_cfg=[
+#         dict(
+#             save_file='model.onnx',  # name to save the partitioned onnx
+#             start=['detector_forward:input'],  # [mark_name:input, ...]
+#             end=['yolo_head:input'],  # [mark_name:output, ...]
+#             output_names=[f'pred_maps.{i}' for i in range(3)])  # out names
+#     ])
+
+# # retinanet, ssd, fsaf for rknn-toolkit2
+# partition_config = dict(
+#     type='rknn',  # the partition policy name
+#     apply_marks=True,
+#     partition_cfg=[
+#         dict(
+#             save_file='model.onnx',
+#             start='detector_forward:input',
+#             end=['BaseDenseHead:output'],
+#             output_names=[f'BaseDenseHead.cls.{i}' for i in range(5)] +
+#             [f'BaseDenseHead.loc.{i}' for i in range(5)])
+#     ])
--- a/configs/mmdet/detection/detection_rknn-int8_static-320x320.py
+++ b/configs/mmdet/detection/detection_rknn-int8_static-320x320.py
@ -0,0 +1,32 @@
+_base_ = ['../_base_/base_static.py', '../../_base_/backends/rknn.py']
+
+onnx_config = dict(input_shape=[320, 320])
+
+codebase_config = dict(model_type='rknn')
+
+backend_config = dict(input_size_list=[[3, 320, 320]])
+
+# # yolov3, yolox for rknn-toolkit and rknn-toolkit2
+# partition_config = dict(
+#     type='rknn',  # the partition policy name
+#     apply_marks=True,  # should always be set to True
+#     partition_cfg=[
+#         dict(
+#             save_file='model.onnx',  # name to save the partitioned onnx
+#             start=['detector_forward:input'],  # [mark_name:input, ...]
+#             end=['yolo_head:input'],  # [mark_name:output, ...]
+#             output_names=[f'pred_maps.{i}' for i in range(3)])  # out names
+#     ])
+
+# # retinanet, ssd, fsaf for rknn-toolkit2
+# partition_config = dict(
+#     type='rknn',  # the partition policy name
+#     apply_marks=True,
+#     partition_cfg=[
+#         dict(
+#             save_file='model.onnx',
+#             start='detector_forward:input',
+#             end=['BaseDenseHead:output'],
+#             output_names=[f'BaseDenseHead.cls.{i}' for i in range(5)] +
+#             [f'BaseDenseHead.loc.{i}' for i in range(5)])
+#     ])
--- a/configs/mmdet/detection/detection_rknn_static.py
+++ b/configs/mmdet/detection/detection_rknn_static.py
@ -1,17 +0,0 @@
-_base_ = ['../_base_/base_static.py', '../../_base_/backends/rknn.py']
-
-onnx_config = dict(input_shape=[640, 640])
-
-codebase_config = dict(model_type='rknn')
-
-backend_config = dict(input_size_list=[[3, 640, 640]])
-
-partition_config = dict(
-    type='rknn',  # the partition policy name
-    apply_marks=True,  # should always be set to True
-    partition_cfg=[
-        dict(
-            save_file='model.onnx',  # name to save the partitioned onnx model
-            start=['detector_forward:input'],  # [mark_name:input/output, ...]
-            end=['yolo_head:input'])  # [mark_name:input/output, ...]
-    ])
--- a/configs/mmdet/detection/detection_tvm-ansor_static-800x1344.py
+++ b/configs/mmdet/detection/detection_tvm-ansor_static-800x1344.py
@ -0,0 +1,13 @@
+_base_ = ['../_base_/base_static.py', '../../_base_/backends/tvm.py']
+
+onnx_config = dict(input_shape=[1344, 800])
+backend_config = dict(model_inputs=[
+    dict(
+        use_vm=True,
+        shape=dict(input=[1, 3, 800, 1344]),
+        dtype=dict(input='float32'),
+        tuner=dict(
+            type='AutoScheduleTuner',
+            log_file='tvm_tune_log.log',
+            num_measure_trials=2000))
+])
--- a/configs/mmdet/detection/detection_tvm-autotvm_static-300x300.py
+++ b/configs/mmdet/detection/detection_tvm-autotvm_static-300x300.py
@ -0,0 +1,15 @@
+_base_ = ['../_base_/base_static.py', '../../_base_/backends/tvm.py']
+
+onnx_config = dict(input_shape=[300, 300])
+backend_config = dict(model_inputs=[
+    dict(
+        use_vm=True,
+        shape=dict(input=[1, 3, 300, 300]),
+        dtype=dict(input='float32'),
+        tuner=dict(
+            type='AutoTVMTuner',
+            log_file='tvm_tune_log.log',
+            n_trial=1000,
+            tuner=dict(type='XGBTuner'),
+        ))
+])
--- a/configs/mmdet/detection/detection_tvm-autotvm_static-800x1344.py
+++ b/configs/mmdet/detection/detection_tvm-autotvm_static-800x1344.py
@ -0,0 +1,15 @@
+_base_ = ['../_base_/base_static.py', '../../_base_/backends/tvm.py']
+
+onnx_config = dict(input_shape=[1344, 800])
+backend_config = dict(model_inputs=[
+    dict(
+        use_vm=True,
+        shape=dict(input=[1, 3, 800, 1344]),
+        dtype=dict(input='float32'),
+        tuner=dict(
+            type='AutoTVMTuner',
+            log_file='tvm_tune_log.log',
+            n_trial=1000,
+            tuner=dict(type='XGBTuner'),
+        ))
+])
--- a/configs/mmdet/detection/yolov3_partition_onnxruntime_static.py
+++ b/configs/mmdet/detection/yolov3_partition_onnxruntime_static.py
@ -8,5 +8,6 @@ partition_config = dict(
        dict(
            save_file='yolov3.onnx',
            start=['detector_forward:input'],
-            end=['yolo_head:input'])
+            end=['yolo_head:input'],
+            output_names=[f'pred_maps.{i}' for i in range(3)])
    ])
--- a/configs/mmdet/instance-seg/instance-seg_coreml_static-800x1344.py
+++ b/configs/mmdet/instance-seg/instance-seg_coreml_static-800x1344.py
@ -0,0 +1 @@
+_base_ = ['../_base_/base_instance-seg_coreml_static-800x1344.py']
--- a/configs/mmdet/instance-seg/instance-seg_tvm-ansor_static-800x1344.py
+++ b/configs/mmdet/instance-seg/instance-seg_tvm-ansor_static-800x1344.py
@ -0,0 +1,15 @@
+_base_ = [
+    '../_base_/base_instance-seg_static.py', '../../_base_/backends/tvm.py'
+]
+
+onnx_config = dict(input_shape=[1344, 800])
+backend_config = dict(model_inputs=[
+    dict(
+        use_vm=True,
+        shape=dict(input=[1, 3, 800, 1344]),
+        dtype=dict(input='float32'),
+        tuner=dict(
+            type='AutoScheduleTuner',
+            log_file='tvm_tune_log.log',
+            num_measure_trials=20000))
+])
--- a/configs/mmdet/instance-seg/instance-seg_tvm-autotvm_static-800x1344.py
+++ b/configs/mmdet/instance-seg/instance-seg_tvm-autotvm_static-800x1344.py
@ -0,0 +1,17 @@
+_base_ = [
+    '../_base_/base_instance-seg_static.py', '../../_base_/backends/tvm.py'
+]
+
+onnx_config = dict(input_shape=[1344, 800])
+backend_config = dict(model_inputs=[
+    dict(
+        use_vm=True,
+        shape=dict(input=[1, 3, 800, 1344]),
+        dtype=dict(input='float32'),
+        tuner=dict(
+            type='AutoTVMTuner',
+            log_file='tvm_tune_log.log',
+            n_trial=10000,
+            tuner=dict(type='XGBTuner'),
+        ))
+])
--- a/configs/mmdet3d/monocular-detection/monocular-detection_dynamic.py
+++ b/configs/mmdet3d/monocular-detection/monocular-detection_dynamic.py
@ -0,0 +1,24 @@
+_base_ = ['./monocular-detection_static.py']
+
+onnx_config = dict(
+    dynamic_axes={
+        'img': {
+            2: 'height',
+            3: 'width',
+        },
+        'bboxes': {
+            1: 'num_dets',
+        },
+        'scores': {
+            1: 'num_dets'
+        },
+        'labels': {
+            1: 'num_dets'
+        },
+        'dir_scores': {
+            1: 'num_dets'
+        },
+        'attrs': {
+            1: 'num_dets'
+        }
+    }, )
--- a/configs/mmdet3d/monocular-detection/monocular-detection_onnxruntime_dynamic.py
+++ b/configs/mmdet3d/monocular-detection/monocular-detection_onnxruntime_dynamic.py
@ -0,0 +1,3 @@
+_base_ = [
+    './monocular-detection_dynamic.py', '../../_base_/backends/onnxruntime.py'
+]
--- a/configs/mmdet3d/monocular-detection/monocular-detection_onnxruntime_static.py
+++ b/configs/mmdet3d/monocular-detection/monocular-detection_onnxruntime_static.py
@ -0,0 +1,3 @@
+_base_ = [
+    './monocular-detection_static.py', '../../_base_/backends/onnxruntime.py'
+]
--- a/configs/mmdet3d/monocular-detection/monocular-detection_static.py
+++ b/configs/mmdet3d/monocular-detection/monocular-detection_static.py
@ -0,0 +1,14 @@
+_base_ = ['../../_base_/onnx_config.py']
+
+onnx_config = dict(
+    input_names=['img', 'cam2img', 'cam2img_inverse'],
+    output_names=['bboxes', 'scores', 'labels', 'dir_scores', 'attrs'],
+    input_shape=None,
+)
+codebase_config = dict(
+    type='mmdet3d',
+    task='MonocularDetection',
+    model_type='end2end',
+    ann_file='tests/test_codebase/test_mmdet3d/data/nuscenes/n015-2018-07-24'
+    '-11-22-45+0800__CAM_BACK__1532402927637525_mono3d.coco.json',
+)
--- a/configs/mmdet3d/monocular-detection/monocular-detection_tensorrt_dynamic-320x320-1600x1600.py
+++ b/configs/mmdet3d/monocular-detection/monocular-detection_tensorrt_dynamic-320x320-1600x1600.py
@ -0,0 +1,16 @@
+_base_ = [
+    './monocular-detection_static.py', '../../_base_/backends/tensorrt.py'
+]
+
+onnx_config = dict(input_shape=(1600, 928))
+
+backend_config = dict(
+    common_config=dict(max_workspace_size=1 << 30),
+    model_inputs=[
+        dict(
+            input_shapes=dict(
+                input=dict(
+                    min_shape=[1, 3, 320, 320],
+                    opt_shape=[1, 3, 928, 1600],
+                    max_shape=[1, 3, 1600, 1600])))
+    ])
--- a/configs/mmdet3d/monocular-detection/monocular-detection_tensorrt_static-928x1600.py
+++ b/configs/mmdet3d/monocular-detection/monocular-detection_tensorrt_static-928x1600.py
@ -0,0 +1,16 @@
+_base_ = [
+    './monocular-detection_static.py', '../../_base_/backends/tensorrt.py'
+]
+
+onnx_config = dict(input_shape=(1600, 928))
+
+backend_config = dict(
+    common_config=dict(max_workspace_size=1 << 30),
+    model_inputs=[
+        dict(
+            input_shapes=dict(
+                input=dict(
+                    min_shape=[1, 3, 928, 1600],
+                    opt_shape=[1, 3, 928, 1600],
+                    max_shape=[1, 3, 928, 1600])))
+    ])
--- a/configs/mmdet3d/voxel-detection/voxel-detection_static.py
+++ b/configs/mmdet3d/voxel-detection/voxel-detection_static.py
@ -3,4 +3,4 @@ codebase_config = dict(
    type='mmdet3d', task='VoxelDetection', model_type='end2end')
 onnx_config = dict(
    input_names=['voxels', 'num_points', 'coors'],
-    output_names=['scores', 'bbox_preds', 'dir_scores'])
+    output_names=['bboxes', 'scores', 'labels'])
--- a/configs/mmedit/inpainting/inpainting_dynamic.py
+++ b/configs/mmedit/inpainting/inpainting_dynamic.py
@ -0,0 +1,20 @@
+_base_ = ['./inpainting_static.py']
+
+onnx_config = dict(
+    dynamic_axes=dict(
+        masked_img={
+            0: 'batch',
+            2: 'height',
+            3: 'width'
+        },
+        mask={
+            0: 'batch',
+            2: 'height',
+            3: 'width'
+        },
+        output={
+            0: 'batch',
+            2: 'height',
+            3: 'width'
+        }),
+    input_shape=None)
--- a/configs/mmedit/inpainting/inpainting_onnxruntime_dynamic.py
+++ b/configs/mmedit/inpainting/inpainting_onnxruntime_dynamic.py
@ -0,0 +1 @@
+_base_ = ['./inpainting_dynamic.py', '../../_base_/backends/onnxruntime.py']
--- a/configs/mmedit/inpainting/inpainting_onnxruntime_static.py
+++ b/configs/mmedit/inpainting/inpainting_onnxruntime_static.py
@ -0,0 +1,3 @@
+_base_ = ['./inpainting_static.py', '../../_base_/backends/onnxruntime.py']
+
+onnx_config = dict(input_shape=[256, 256])
--- a/configs/mmedit/inpainting/inpainting_static.py
+++ b/configs/mmedit/inpainting/inpainting_static.py
@ -0,0 +1,5 @@
+_base_ = ['../../_base_/onnx_config.py']
+
+codebase_config = dict(type='mmedit', task='Inpainting')
+onnx_config = dict(
+    input_names=['masked_img', 'mask'], output_names=['fake_img'])
--- a/configs/mmedit/inpainting/inpainting_tensorrt-fp16_static-256x256.py
+++ b/configs/mmedit/inpainting/inpainting_tensorrt-fp16_static-256x256.py
@ -0,0 +1,17 @@
+_base_ = ['./inpainting_static.py', '../../_base_/backends/tensorrt-fp16.py']
+
+onnx_config = dict(input_shape=[256, 256])
+backend_config = dict(
+    common_config=dict(max_workspace_size=1 << 30),
+    model_inputs=[
+        dict(
+            input_shapes=dict(
+                masked_img=dict(
+                    min_shape=[1, 3, 256, 256],
+                    opt_shape=[1, 3, 256, 256],
+                    max_shape=[1, 3, 256, 256]),
+                mask=dict(
+                    min_shape=[1, 1, 256, 256],
+                    opt_shape=[1, 1, 256, 256],
+                    max_shape=[1, 1, 256, 256])))
+    ])
--- a/configs/mmedit/inpainting/inpainting_tensorrt-int8_static-256x256.py
+++ b/configs/mmedit/inpainting/inpainting_tensorrt-int8_static-256x256.py
@ -0,0 +1,17 @@
+_base_ = ['./inpainting_static.py', '../../_base_/backends/tensorrt-int8.py']
+
+onnx_config = dict(input_shape=[256, 256])
+backend_config = dict(
+    common_config=dict(max_workspace_size=1 << 30),
+    model_inputs=[
+        dict(
+            input_shapes=dict(
+                masked_img=dict(
+                    min_shape=[1, 3, 256, 256],
+                    opt_shape=[1, 3, 256, 256],
+                    max_shape=[1, 3, 256, 256]),
+                mask=dict(
+                    min_shape=[1, 1, 256, 256],
+                    opt_shape=[1, 1, 256, 256],
+                    max_shape=[1, 1, 256, 256])))
+    ])
--- a/configs/mmedit/inpainting/inpainting_tensorrt_static-256x256.py
+++ b/configs/mmedit/inpainting/inpainting_tensorrt_static-256x256.py
@ -0,0 +1,17 @@
+_base_ = ['./inpainting_static.py', '../../_base_/backends/tensorrt.py']
+
+onnx_config = dict(input_shape=[256, 256])
+backend_config = dict(
+    common_config=dict(max_workspace_size=1 << 30),
+    model_inputs=[
+        dict(
+            input_shapes=dict(
+                masked_img=dict(
+                    min_shape=[1, 3, 256, 256],
+                    opt_shape=[1, 3, 256, 256],
+                    max_shape=[1, 3, 256, 256]),
+                mask=dict(
+                    min_shape=[1, 1, 256, 256],
+                    opt_shape=[1, 1, 256, 256],
+                    max_shape=[1, 1, 256, 256])))
+    ])
--- a/configs/mmedit/super-resolution/super-resolution_ncnn-int8_dynamic.py
+++ b/configs/mmedit/super-resolution/super-resolution_ncnn-int8_dynamic.py
@ -0,0 +1,3 @@
+_base_ = [
+    './super-resolution_dynamic.py', '../../_base_/backends/ncnn-int8.py'
+]
--- a/configs/mmocr/text-detection/text-detection_ncnn-int8_static.py
+++ b/configs/mmocr/text-detection/text-detection_ncnn-int8_static.py
--- a/configs/mmocr/text-recognition/text-recognition_coreml_dynamic-32x32-32x640.py
+++ b/configs/mmocr/text-recognition/text-recognition_coreml_dynamic-32x32-32x640.py
@ -0,0 +1,13 @@
+_base_ = [
+    '../../_base_/torchscript_config.py', '../../_base_/backends/coreml.py'
+]
+
+codebase_config = dict(type='mmocr', task='TextRecognition')
+backend_config = dict(model_inputs=[
+    dict(
+        input_shapes=dict(
+            input=dict(
+                min_shape=[1, 3, 32, 32],
+                max_shape=[1, 3, 32, 640],
+                default_shape=[1, 3, 32, 64])))
+])
--- a/configs/mmseg/segmentation_ncnn-int8_static-1024x2048.py
+++ b/configs/mmseg/segmentation_ncnn-int8_static-1024x2048.py
@ -0,0 +1,3 @@
+_base_ = ['./segmentation_static.py', '../_base_/backends/ncnn-int8.py']
+
+onnx_config = dict(input_shape=[2048, 1024])
--- a/configs/mmseg/segmentation_rknn-fp16_static-320x320.py
+++ b/configs/mmseg/segmentation_rknn-fp16_static-320x320.py
@ -0,0 +1,9 @@
+_base_ = ['./segmentation_static.py', '../_base_/backends/rknn.py']
+
+onnx_config = dict(input_shape=[320, 320])
+
+codebase_config = dict(model_type='rknn')
+
+backend_config = dict(
+    input_size_list=[[3, 320, 320]],
+    quantization_config=dict(do_quantization=False))
--- a/configs/mmseg/segmentation_rknn-int8_static-320x320.py
+++ b/configs/mmseg/segmentation_rknn-int8_static-320x320.py
@ -0,0 +1,7 @@
+_base_ = ['./segmentation_static.py', '../_base_/backends/rknn.py']
+
+onnx_config = dict(input_shape=[320, 320])
+
+codebase_config = dict(model_type='rknn', with_argmax=False)
+
+backend_config = dict(input_size_list=[[3, 320, 320]])
--- a/configs/mmseg/segmentation_rknn_static.py
+++ b/configs/mmseg/segmentation_rknn_static.py
@ -1,7 +0,0 @@
-_base_ = ['./segmentation_static.py', '../_base_/backends/rknn.py']
-
-onnx_config = dict(input_shape=[512, 512])
-
-codebase_config = dict(model_type='rknn')
-
-backend_config = dict(input_size_list=[[3, 512, 512]])
--- a/configs/mmseg/segmentation_static.py
+++ b/configs/mmseg/segmentation_static.py
@ -1,2 +1,2 @@
 _base_ = ['../_base_/onnx_config.py']
-codebase_config = dict(type='mmseg', task='Segmentation')
+codebase_config = dict(type='mmseg', task='Segmentation', with_argmax=True)
--- a/configs/mmseg/segmentation_tvm-ansor_static-512x1024.py
+++ b/configs/mmseg/segmentation_tvm-ansor_static-512x1024.py
@ -0,0 +1,12 @@
+_base_ = ['./segmentation_static.py', '../_base_/backends/tvm.py']
+
+onnx_config = dict(input_shape=[1024, 512])
+backend_config = dict(model_inputs=[
+    dict(
+        shape=dict(input=[1, 3, 512, 1024]),
+        dtype=dict(input='float32'),
+        tuner=dict(
+            type='AutoScheduleTuner',
+            log_file='tvm_tune_log.log',
+            num_measure_trials=2000))
+])
--- a/configs/mmseg/segmentation_tvm-autotvm_static-512x1024.py
+++ b/configs/mmseg/segmentation_tvm-autotvm_static-512x1024.py
@ -0,0 +1,13 @@
+_base_ = ['./segmentation_static.py', '../_base_/backends/tvm.py']
+
+onnx_config = dict(input_shape=[1024, 512])
+backend_config = dict(model_inputs=[
+    dict(
+        shape=dict(input=[1, 3, 512, 1024]),
+        dtype=dict(input='float32'),
+        tuner=dict(
+            type='AutoTVMTuner',
+            log_file='tvm_tune_log.log',
+            n_trial=1000,
+            tuner=dict(type='XGBTuner')))
+])
--- a/csrc/mmdeploy/CMakeLists.txt
+++ b/csrc/mmdeploy/CMakeLists.txt
@ -13,6 +13,7 @@ if (MMDEPLOY_BUILD_SDK)
    add_subdirectory(device)
    add_subdirectory(graph)
    add_subdirectory(model)
+    add_subdirectory(operation)
    add_subdirectory(preprocess)
    add_subdirectory(net)
    add_subdirectory(codebase)
--- a/csrc/mmdeploy/apis/CMakeLists.txt
+++ b/csrc/mmdeploy/apis/CMakeLists.txt
@ -1,10 +1,5 @@
 # Copyright (c) OpenMMLab. All rights reserved.

-# Python API depends on C++ API
-if (MMDEPLOY_BUILD_SDK_PYTHON_API)
-    set(MMDEPLOY_BUILD_SDK_CXX_API ON)
-endif ()
-
 add_subdirectory(c)
 add_subdirectory(cxx)
 add_subdirectory(java)
--- a/csrc/mmdeploy/apis/c/CMakeLists.txt
+++ b/csrc/mmdeploy/apis/c/CMakeLists.txt
@ -12,6 +12,9 @@ macro(add_object name)
        target_compile_options(${name} PRIVATE $<$<COMPILE_LANGUAGE:CXX>:-fvisibility=hidden>)
    endif ()
    target_link_libraries(${name} PRIVATE mmdeploy::core)
+    target_include_directories(${name} PUBLIC
+            $<BUILD_INTERFACE:${CMAKE_CURRENT_SOURCE_DIR}>
+            $<INSTALL_INTERFACE:include>)
    set(CAPI_OBJS ${CAPI_OBJS} ${name})
    mmdeploy_export(${name})
 endmacro()
@ -31,6 +34,7 @@ foreach (TASK ${COMMON_LIST})
    mmdeploy_add_library(${TARGET_NAME})
    target_link_libraries(${TARGET_NAME} PRIVATE ${OBJECT_NAME})
    target_include_directories(${TARGET_NAME} PUBLIC
+            $<BUILD_INTERFACE:${CMAKE_CURRENT_SOURCE_DIR}>
            $<INSTALL_INTERFACE:include>)
    install(FILES ${CMAKE_CURRENT_SOURCE_DIR}/mmdeploy/${TASK}.h
            DESTINATION include/mmdeploy)
@ -76,5 +80,14 @@ if (MMDEPLOY_BUILD_SDK_CSHARP_API OR MMDEPLOY_BUILD_SDK_MONOLITHIC)
    set_target_properties(mmdeploy PROPERTIES
            VERSION ${MMDEPLOY_VERSION}
            SOVERSION ${MMDEPLOY_VERSION_MAJOR})
-    mmdeploy_export(mmdeploy)
+    if (APPLE)
+            set_target_properties(mmdeploy PROPERTIES
+                    INSTALL_RPATH "@loader_path"
+                    BUILD_RPATH "@loader_path")
+    else ()
+            set_target_properties(mmdeploy PROPERTIES
+                    INSTALL_RPATH "\$ORIGIN"
+                    BUILD_RPATH "\$ORIGIN")
+    endif ()
+    mmdeploy_export_impl(mmdeploy)
 endif ()
--- a/csrc/mmdeploy/apis/c/mmdeploy/classifier.cpp
+++ b/csrc/mmdeploy/apis/c/mmdeploy/classifier.cpp
@ -1,52 +1,21 @@
 // Copyright (c) OpenMMLab. All rights reserved.

-#include "classifier.h"
+#include "mmdeploy/classifier.h"

 #include <numeric>

-#include "common_internal.h"
-#include "handle.h"
 #include "mmdeploy/archive/value_archive.h"
 #include "mmdeploy/codebase/mmcls/mmcls.h"
+#include "mmdeploy/common_internal.h"
 #include "mmdeploy/core/device.h"
 #include "mmdeploy/core/graph.h"
 #include "mmdeploy/core/utils/formatter.h"
-#include "pipeline.h"
+#include "mmdeploy/handle.h"
+#include "mmdeploy/pipeline.h"

 using namespace mmdeploy;
 using namespace std;

-namespace {
-
-Value config_template(const Model& model) {
-  // clang-format off
-  static Value v{
-    {
-      "pipeline", {
-        {"input", {"img"}},
-        {"output", {"cls"}},
-        {
-          "tasks", {
-            {
-              {"name", "classifier"},
-              {"type", "Inference"},
-              {"params", {{"model", "TBD"}}},
-              {"input", {"img"}},
-              {"output", {"cls"}}
-            }
-          }
-        }
-      }
-    }
-  };
-  // clang-format on
-  auto config = v;
-  config["pipeline"]["tasks"][0]["params"]["model"] = model;
-  return config;
-}
-
-}  // namespace
-
 int mmdeploy_classifier_create(mmdeploy_model_t model, const char* device_name, int device_id,
                               mmdeploy_classifier_t* classifier) {
  mmdeploy_context_t context{};
@ -73,8 +42,7 @@ int mmdeploy_classifier_create_by_path(const char* model_path, const char* devic

 int mmdeploy_classifier_create_v2(mmdeploy_model_t model, mmdeploy_context_t context,
                                  mmdeploy_classifier_t* classifier) {
-  auto config = config_template(*Cast(model));
-  return mmdeploy_pipeline_create_v3(Cast(&config), context, (mmdeploy_pipeline_t*)classifier);
+  return mmdeploy_pipeline_create_from_model(model, context, (mmdeploy_pipeline_t*)classifier);
 }

 int mmdeploy_classifier_create_input(const mmdeploy_mat_t* mats, int mat_count,
--- a/Show More
+++ b/Show More
				`@ -0,0 +1 @@`
				`_base_ = ['../_base_/base_instance-seg_coreml_static-800x1344.py']`
				`@ -0,0 +1 @@`
				`_base_ = ['./inpainting_dynamic.py', '../../_base_/backends/onnxruntime.py']`