mmcls -> mmpretrain

.github/scripts/linux/test_full_pipeline.sh

@@ -14,19 +14,19 @@ mkdir -p $work_dir $mmdeploy_dir/data
 
 model_cfg=$work_dir/resnet18_8xb32_in1k.py
 checkpoint=$work_dir/resnet18_8xb32_in1k_20210831-fbbb1da6.pth
-sdk_cfg=configs/mmcls/classification_sdk_dynamic.py
+sdk_cfg=configs/mmpretrain/classification_sdk_dynamic.py
 input_img=tests/data/tiger.jpeg
 
-python3 -m mim download mmcls --config resnet18_8xb32_in1k --dest $work_dir
+python3 -m mim download mmpretrain --config resnet18_8xb32_in1k --dest $work_dir
 
 if [ $backend == "ort" ]; then
-    deploy_cfg=configs/mmcls/classification_onnxruntime_dynamic.py
+    deploy_cfg=configs/mmpretrain/classification_onnxruntime_dynamic.py
     model=$work_dir/end2end.onnx
 elif [ $backend == "trt" ]; then
-    deploy_cfg=configs/mmcls/classification_tensorrt-fp16_dynamic-224x224-224x224.py
+    deploy_cfg=configs/mmpretrain/classification_tensorrt-fp16_dynamic-224x224-224x224.py
     model=$work_dir/end2end.engine
 elif [ $backend == "ncnn" ]; then
-    deploy_cfg=configs/mmcls/classification_ncnn_static.py
+    deploy_cfg=configs/mmpretrain/classification_ncnn_static.py
     model="$work_dir/end2end.param $work_dir/end2end.bin"
 else
   echo "Unsupported Backend=$backend"

.github/scripts/quantize_to_ncnn.py

@@ -9,17 +9,17 @@
     [
         'hrnet.onnx',
         'https://media.githubusercontent.com/media/tpoisonooo/mmdeploy-onnx2ncnn-testdata/main/hrnet.onnx',  # noqa: E501
-        '~/mmclassification/configs/hrnet/hrnet-w18_4xb32_in1k.py',
+        '~/mmpretrain/configs/hrnet/hrnet-w18_4xb32_in1k.py',
     ],
     [
         'resnet18.onnx',
         'https://media.githubusercontent.com/media/tpoisonooo/mmdeploy-onnx2ncnn-testdata/main/resnet18.onnx',  # noqa: E501
-        '~/mmclassification/configs/resnet/resnet18_8xb16_cifar10.py',
+        '~/mmpretrain/configs/resnet/resnet18_8xb16_cifar10.py',
     ],
     [
         'mobilenet-v2.onnx',
         'https://media.githubusercontent.com/media/tpoisonooo/mmdeploy-onnx2ncnn-testdata/main/mobilenet-v2.onnx',  # noqa: E501
-        '~/mmclassification/configs/mobilenet_v2/mobilenet-v2_8xb32_in1k.py',
+        '~/mmpretrain/configs/mobilenet_v2/mobilenet-v2_8xb32_in1k.py',
     ]
 ]
 
@@ -45,7 +45,7 @@ def main():
         model = conf[0]
         os.system('wget {}'.format(conf[1]))
         model_cfg = conf[2]
-        deploy_cfg = 'configs/mmcls/classification_ncnn-int8_static.py'
+        deploy_cfg = 'configs/mmpretrain/classification_ncnn-int8_static.py'
         quant_cmd = [
             'python', 'tools/onnx2ncnn_quant_table.py', '--onnx', model,
             '--deploy-cfg', deploy_cfg, '--model-cfg', model_cfg, '--out-onnx',

.github/scripts/test_onnx2ncnn.py

@@ -7,22 +7,22 @@
 # list of tuple: config, pretrained model, onnx filename
 CONFIGS = [
     (
-        'mmclassification/configs/vision_transformer/vit-base-p32_ft-64xb64_in1k-384.py',  # noqa: E501
+        'mmpretrain/configs/vision_transformer/vit-base-p32_ft-64xb64_in1k-384.py',  # noqa: E501
         'https://download.openmmlab.com/mmclassification/v0/vit/finetune/vit-base-p32_in21k-pre-3rdparty_ft-64xb64_in1k-384_20210928-9cea8599.pth',  # noqa: E501
         'vit.onnx'),
     (
-        'mmclassification/configs/resnet/resnet50_8xb32_in1k.py',
+        'mmpretrain/configs/resnet/resnet50_8xb32_in1k.py',
         'https://download.openmmlab.com/mmclassification/v0/resnet/resnet50_8xb32_in1k_20210831-ea4938fc.pth',  # noqa: E501
         'resnet50.onnx',
     ),
     (
-        'mmclassification/configs/resnet/resnet18_8xb32_in1k.py',
+        'mmpretrain/configs/resnet/resnet18_8xb32_in1k.py',
         'https://download.openmmlab.com/mmclassification/v0/resnet/resnet18_8xb32_in1k_20210831-fbbb1da6.pth',  # noqa: E501
         'resnet18.onnx',
         'https://media.githubusercontent.com/media/tpoisonooo/mmdeploy-onnx2ncnn-testdata/main/resnet18.onnx',  # noqa: E501
     ),
     (
-        'mmclassification/configs/mobilenet_v2/mobilenet-v2_8xb32_in1k.py',
+        'mmpretrain/configs/mobilenet_v2/mobilenet-v2_8xb32_in1k.py',
         'https://download.openmmlab.com/mmclassification/v0/mobilenet_v2/mobilenet_v2_batch256_imagenet_20200708-3b2dc3af.pth',  # noqa: E501
         'mobilenet-v2.onnx',
         'https://media.githubusercontent.com/media/tpoisonooo/mmdeploy-onnx2ncnn-testdata/main/mobilenet-v2.onnx',  # noqa: E501
@@ -36,7 +36,7 @@ def parse_args():
     parser.add_argument(
         '--run', type=bool, help='Execute mmdeploy_onnx2ncnn bin.')
     parser.add_argument(
-        '--repo-dir', type=str, default='~/', help='mmcls directory.')
+        '--repo-dir', type=str, default='~/', help='mmpretrain directory.')
     parser.add_argument(
         '--out',
         type=str,
@@ -56,7 +56,7 @@ def generate_onnx(args):
         model = conf[1]
         convert_cmd = [
             'python', 'tools/deploy.py',
-            'configs/mmcls/classification_ncnn_static.py', config, model,
+            'configs/mmpretrain/classification_ncnn_static.py', config, model,
             'cat-dog.png', '--work-dir', 'work_dir', '--device', 'cpu'
         ]
         print(subprocess.call(convert_cmd))

.github/scripts/windows/test_full_pipeline.ps1

@@ -23,16 +23,16 @@ $work_dir="work_dir"
 New-Item -Path $work_dir, .\data -ItemType Directory -Force
 $model_cfg="$work_dir\resnet18_8xb32_in1k.py"
 $checkpoint="$work_dir\resnet18_8xb32_in1k_20210831-fbbb1da6.pth"
-$sdk_cfg="configs\mmcls\classification_sdk_dynamic.py"
+$sdk_cfg="configs\mmpretrain\classification_sdk_dynamic.py"
 $input_img="tests\data\tiger.jpeg"
 
-python -m mim download mmcls --config resnet18_8xb32_in1k --dest $work_dir
+python -m mim download mmpretrain --config resnet18_8xb32_in1k --dest $work_dir
 
 if ($Backend -eq "ort") {
-    $deploy_cfg="configs\mmcls\classification_onnxruntime_dynamic.py"
+    $deploy_cfg="configs\mmpretrain\classification_onnxruntime_dynamic.py"
     $model="$work_dir\end2end.onnx"
 } elseif ($Backend -eq "trt") {
-    $deploy_cfg="configs\mmcls\classification_tensorrt-fp16_dynamic-224x224-224x224.py"
+    $deploy_cfg="configs\mmpretrain\classification_tensorrt-fp16_dynamic-224x224-224x224.py"
     $model="$work_dir\end2end.engine"
 } else {
     Write-Host "Unsupported Backend=$Backend"

.github/workflows/backend-ort.yml

@@ -37,9 +37,9 @@ jobs:
           python -m pip install mmcv-lite openmim
           python tools/scripts/build_ubuntu_x64_ort.py 8
           python -c 'import mmdeploy.apis.onnxruntime as ort_api; assert ort_api.is_available(with_custom_ops=True)'
-      - name: test mmcls full pipeline
+      - name: test mmpretrain full pipeline
         run: |
-          python -m mim install $(cat requirements/codebases.txt | grep mmcls)
+          python -m mim install $(cat requirements/codebases.txt | grep mmpretrain)
           export MMDEPLOY_DIR=$(pwd)
           export ONNXRUNTIME_DIR=$MMDEPLOY_DIR/../mmdeploy-dep/onnxruntime-linux-x64-1.8.1
           export LD_LIBRARY_PATH=$ONNXRUNTIME_DIR/lib:$MMDEPLOY_DIR/build/install/lib:$LD_LIBRARY_PATH

.github/workflows/backend-pplnn.yml

@@ -33,7 +33,7 @@ jobs:
           python3 -V
           python3 -m pip install openmim numpy
           python3 -m pip install -r requirements.txt
-          python3 -m mim install $(cat requirements/codebases.txt | grep mmcls)
+          python3 -m mim install $(cat requirements/codebases.txt | grep mmpretrain)
           python3 -m pip list
       - name: Build SDK
         run: |

.github/workflows/build.yml

@@ -233,7 +233,7 @@ jobs:
           python3 -m pip install opencv-python==4.5.4.60 opencv-python-headless==4.5.4.60 opencv-contrib-python==4.5.4.60
           python3 -m pip install openmim numpy pycuda clip transformers
           python3 -m pip install -r requirements.txt
-          python3 -m mim install $(cat requirements/codebases.txt | grep mmcls)
+          python3 -m mim install $(cat requirements/codebases.txt | grep mmpretrain)
           python3 -m pip list
       - name: Build SDK
         run: |
@@ -281,7 +281,7 @@ jobs:
           python -m pip install openmim
           python -m pip install -r requirements.txt
           python -m pip install -r requirements/backends.txt
-          python -m mim install "mmcls>=1.0.0rc2"
+          python -m mim install "mmpretrain>=1.0.0rc7"
           python -m pip list
       - name: Build mmdeploy
         run: |

.github/workflows/quantize.yml

@@ -45,12 +45,12 @@ jobs:
           python -m pip install -r requirements.txt
           python -m pip install -U numpy
 
-      - name: Install mmcls
+      - name: Install mmpretrain
         run: |
           python -m pip install openmim
-          python -m mim install $(cat requirements/codebases.txt | grep mmcls)
-          git clone --depth 1 --branch 1.x https://github.com/open-mmlab/mmclassification.git ~/mmclassification
-          cd ~/mmclassification && python -m pip install . && cd -
+          python -m mim install $(cat requirements/codebases.txt | grep mmpretrain)
+          git clone --depth 1 --branch 1.x https://github.com/open-mmlab/mmpretrain.git ~/mmpretrain
+          cd ~/mmpretrain && python -m pip install . && cd -
       - name: Install ppq
         run: |
           python -m pip install protobuf==3.20.0

README.md

@@ -67,7 +67,7 @@ MMDeploy is an open-source deep learning model deployment toolset. It is a part
 
 The currently supported codebases and models are as follows, and more will be included in the future
 
-- [mmcls](docs/en/04-supported-codebases/mmcls.md)
+- [mmpretrain](docs/en/04-supported-codebases/mmpretrain.md)
 - [mmdet](docs/en/04-supported-codebases/mmdet.md)
 - [mmseg](docs/en/04-supported-codebases/mmseg.md)
 - [mmedit](docs/en/04-supported-codebases/mmedit.md)

README_zh-CN.md

@@ -33,10 +33,10 @@
 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       |
+| mmdeploy | mmengine |   mmcv   |  mmdet   | mmpretrain 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
 
@@ -52,7 +52,7 @@ MMDeploy 是 [OpenMMLab](https://openmmlab.com/) 模型部署工具箱，**为
 
 ### 支持超多 OpenMMLab 算法库
 
-- [mmcls](docs/zh_cn/04-supported-codebases/mmcls.md)
+- [mmpretrain](docs/zh_cn/04-supported-codebases/mmpretrain.md)
 - [mmdet](docs/zh_cn/04-supported-codebases/mmdet.md)
 - [mmseg](docs/zh_cn/04-supported-codebases/mmseg.md)
 - [mmedit](docs/zh_cn/04-supported-codebases/mmedit.md)

configs/mmcls/classification_coreml_dynamic-224x224-224x224.py → configs/mmpretrain/classification_coreml_dynamic-224x224-224x224.py

@@ -1,6 +1,6 @@
 _base_ = ['../_base_/torchscript_config.py', '../_base_/backends/coreml.py']
 
-codebase_config = dict(type='mmcls', task='Classification')
+codebase_config = dict(type='mmpretrain', task='Classification')
 
 backend_config = dict(model_inputs=[
     dict(

configs/mmcls/classification_sdk_dynamic.py → configs/mmpretrain/classification_sdk_dynamic.py

@@ -4,4 +4,4 @@
 
 backend_config = dict(
     pipeline=[dict(type='LoadImageFromFile'),
-              dict(type='PackClsInputs')])
+              dict(type='PackInputs')])

configs/mmpretrain/classification_static.py

@@ -0,0 +1,3 @@
+_base_ = ['../_base_/onnx_config.py']
+
+codebase_config = dict(type='mmpretrain', task='Classification')

configs/mmcls/classification_torchscript.py → configs/mmpretrain/classification_torchscript.py

@@ -3,4 +3,4 @@
 ]
 
 ir_config = dict(input_shape=None)
-codebase_config = dict(type='mmcls', task='Classification')
+codebase_config = dict(type='mmpretrain', task='Classification')

csrc/mmdeploy/codebase/CMakeLists.txt

@@ -18,5 +18,10 @@ endif ()
 
 foreach (codebase IN LISTS CODEBASES)
     message(STATUS "build codebase: ${codebase}")
-    add_subdirectory(${codebase})
+    if (codebase STREQUAL "mmpretrain")
+        set(subdir_name "mmcls")
+    else()
+	set(subdir_name ${codebase})
+    endif()
+    add_subdirectory(${subdir_name})
 endforeach ()

docs/en/01-how-to-build/cmake_option.md

@@ -121,9 +121,9 @@
   </tr>
   <tr>
     <td>MMDEPLOY_CODEBASES</td>
-    <td>{"mmcls", "mmdet", "mmseg", "mmedit", "mmocr", "all"}</td>
+    <td>{"mmpretrain", "mmdet", "mmseg", "mmedit", "mmocr", "all"}</td>
     <td>all</td>
-    <td>Enable codebase's postprocess modules. You can provide a semicolon separated list of codebase names to enable them, e.g., <code>-DMMDEPLOY_CODEBASES="mmcls;mmdet"</code>. Or you can pass <code>all</code> to enable them all, i.e., <code>-DMMDEPLOY_CODEBASES=all</code></td>
+    <td>Enable codebase's postprocess modules. You can provide a semicolon separated list of codebase names to enable them, e.g., <code>-DMMDEPLOY_CODEBASES="mmpretrain;mmdet"</code>. Or you can pass <code>all</code> to enable them all, i.e., <code>-DMMDEPLOY_CODEBASES=all</code></td>
   </tr>
 
 </tbody>

docs/en/01-how-to-build/cross_build_ncnn_aarch64.md

@@ -13,16 +13,16 @@ Device
 
 ## 1. Model Convert on Host
 
-Refer to the doc to install [mmdeploy](../01-how-to-build/) and [mmcls](https://github.com/open-mmlab/mmclassification), and convert resnet18 for model package
+Refer to the doc to install [mmdeploy](../01-how-to-build/) and [mmpretrain](https://github.com/open-mmlab/mmpretrain), and convert resnet18 for model package
 
 ```bash
-export MODEL_CONFIG=/path/to/mmclassification/configs/resnet/resnet18_8xb32_in1k.py
+export MODEL_CONFIG=/path/to/mmpretrain/configs/resnet/resnet18_8xb32_in1k.py
 export MODEL_PATH=https://download.openmmlab.com/mmclassification/v0/resnet/resnet18_8xb32_in1k_20210831-fbbb1da6.pth
 
 # Convert resnet18
 cd /path/to/mmdeploy
 python tools/deploy.py \
-  configs/mmcls/classification_ncnn_static.py \
+  configs/mmpretrain/classification_ncnn_static.py \
   $MODEL_CONFIG \
   $MODEL_PATH \
   tests/data/tiger.jpeg \

docs/en/01-how-to-build/riscv.md

@@ -15,13 +15,13 @@ You can refer to [Build document](./linux-x86_64.md) to install ncnn inference e
 b) Convert model
 
 ```bash
-export MODEL_CONFIG=/path/to/mmclassification/configs/resnet/resnet18_8xb32_in1k.py
+export MODEL_CONFIG=/path/to/mmpretrain/configs/resnet/resnet18_8xb32_in1k.py
 export MODEL_PATH=https://download.openmmlab.com/mmclassification/v0/resnet/resnet18_8xb32_in1k_20210831-fbbb1da6.pth
 
 # Convert the model
 cd /path/to/mmdeploy
 python tools/deploy.py \
-  configs/mmcls/classification_ncnn_static.py \
+  configs/mmpretrain/classification_ncnn_static.py \
   $MODEL_CONFIG \
   $MODEL_PATH \
   tests/data/tiger.jpeg \

docs/en/01-how-to-build/rockchip.md

@@ -42,25 +42,25 @@ It is recommended to create a virtual environment for the project.
 conda install pytorch==1.8.0 torchvision==0.9.0 cudatoolkit=11.1 -c pytorch -c conda-forge
 ```
 
-To work with models from [MMClassification](https://mmclassification.readthedocs.io/en/1.x/get_started.html), you may need to install it additionally.
+To work with models from [MMPretrain](https://mmpretrain.readthedocs.io/en/latest/get_started.html), you may need to install it additionally.
 
 ## Usage
 
 Example:
 
 ```bash
 python tools/deploy.py \
-    configs/mmcls/classification_rknn_static.py \
-    /mmclassification_dir/configs/resnet/resnet50_8xb32_in1k.py \
+    configs/mmpretrain/classification_rknn_static.py \
+    /mmpretrain_dir/configs/resnet/resnet50_8xb32_in1k.py \
     https://download.openmmlab.com/mmclassification/v0/resnet/resnet50_batch256_imagenet_20200708-cfb998bf.pth \
-    /mmclassification_dir/demo/demo.JPEG \
+    /mmpretrain_dir/demo/demo.JPEG \
     --work-dir ../resnet50 \
     --device cpu
 ```
 
 ## Deployment config
 
-With the deployment config, you can modify the `backend_config` for your preference. An example `backend_config` of mmclassification is shown as below:
+With the deployment config, you can modify the `backend_config` for your preference. An example `backend_config` of mmpretrain is shown as below:
 
 ```python
 backend_config = dict(
@@ -119,7 +119,7 @@ First make sure that`--dump-info`is used during convert model, so that the worki
 ```bash
 cd /path/to/mmdeploy
 adb push resnet50  /data/local/tmp/resnet50
-adb push /mmclassification_dir/demo/demo.JPEG /data/local/tmp/resnet50/demo.JPEG
+adb push /mmpretrain_dir/demo/demo.JPEG /data/local/tmp/resnet50/demo.JPEG
 cd build
 adb push lib /data/local/tmp/lib
 adb push bin/image_classification /data/local/tmp/image_classification

docs/en/01-how-to-build/snpe.md

@@ -71,18 +71,18 @@ $ python3 -m pip install -e .
 
 ## 3. Test the model
 
-Take Resnet-18 as an example. First refer to [documentation to install mmcls](https://github.com/open-mmlab/mmclassification/tree/1.x)  and  use `tools/deploy.py` to convert the model.
+Take Resnet-18 as an example. First refer to [documentation to install mmpretrain](https://github.com/open-mmlab/mmpretrain/tree/main)  and  use `tools/deploy.py` to convert the model.
 
 ```bash
-$ export MODEL_CONFIG=/path/to/mmclassification/configs/resnet/resnet18_8xb16_cifar10.py
+$ export MODEL_CONFIG=/path/to/mmpretrain/configs/resnet/resnet18_8xb16_cifar10.py
 $ export MODEL_PATH=https://download.openmmlab.com/mmclassification/v0/resnet/resnet18_b16x8_cifar10_20210528-bd6371c8.pth
 
 # Convert the model
 $ cd /path/to/mmdeploy
-$ python3 tools/deploy.py  configs/mmcls/classification_snpe_static.py $MODEL_CONFIG  $MODEL_PATH   /path/to/test.png   --work-dir resnet18   --device cpu  --uri 10.0.0.1\:60000  --dump-info
+$ python3 tools/deploy.py  configs/mmpretrain/classification_snpe_static.py $MODEL_CONFIG  $MODEL_PATH   /path/to/test.png   --work-dir resnet18   --device cpu  --uri 10.0.0.1\:60000  --dump-info
 
 # Test
-$ python3 tools/test.py configs/mmcls/classification_snpe_static.py   $MODEL_CONFIG    --model reset18/end2end.dlc   --metrics accuracy precision f1_score recall  --uri 10.0.0.1\:60000
+$ python3 tools/test.py configs/mmpretrain/classification_snpe_static.py   $MODEL_CONFIG    --model reset18/end2end.dlc   --metrics accuracy precision f1_score recall  --uri 10.0.0.1\:60000
 ```
 
 Note that `--uri` is required to specify the ip and port of the snpe inference service, ipv4 and ipv6 addresses can be used.
@@ -163,7 +163,7 @@ $ adb push bin/image_classification /data/local/tmp/resnet18/
 Set up environment variable and execute the sample.
 
 ```bash
-$ adb push /path/to/mmcls/demo/demo.JPEG /data/local/tmp
+$ adb push /path/to/mmpretrain/demo/demo.JPEG /data/local/tmp
 $ adb shell
 venus:/ $ cd /data/local/tmp/resnet18
 venus:/data/local/tmp/resnet18 $ export LD_LIBRARY_PATH=${LD_LIBRARY_PATH}:/data/local/tmp/lib