Remove deprecated models (#3001)

* Remove deprecated models

* remove AC configs links and reference in datasets

* remove models usage in demos and tests

* remove missed files

* remove caffe2 to onnx script

* update ci requirements

* update known frameworks, restore some req

Author: eaidova

CONTRIBUTING.md

@@ -4,7 +4,6 @@ We appreciate your intention to contribute model to the OpenVINO™ Open Mod
 
 Frameworks supported by the Open Model Zoo:
 * Caffe\*
-* Caffe2\* (via conversion to ONNX\*)
 * TensorFlow\*
 * PyTorch\* (via conversion to ONNX\*)
 * MXNet\*
@@ -113,7 +112,7 @@ For replacement operation:
 - `replacement` — Replacement string
 - `count` (*optional*)  — Exact number of replacements (if number of `pattern` occurrences less then this number, downloading will be aborted)
 
-**`conversion_to_onnx_args`** (*only for Caffe2\*, PyTorch\* models*)
+**`conversion_to_onnx_args`** (*only for PyTorch\* models*)
 
 List of ONNX\* conversion parameters, see `model_optimizer_args` for details.
 

ci/requirements-ac-test.txt

@@ -20,8 +20,6 @@ decorator==4.4.2
     # via networkx
 defusedxml==0.7.1
     # via -r tools/accuracy_checker/requirements-core.in
-editdistance==0.5.3
-    # via -r tools/accuracy_checker/requirements.in
 fast-ctc-decode==0.3.0
     # via -r tools/accuracy_checker/requirements.in
 filelock==3.0.12

ci/requirements-ac.txt

@@ -16,8 +16,6 @@ decorator==4.4.2
     # via networkx
 defusedxml==0.7.1
     # via -r tools/accuracy_checker/requirements-core.in
-editdistance==0.5.3
-    # via -r tools/accuracy_checker/requirements.in
 fast-ctc-decode==0.3.0
     # via -r tools/accuracy_checker/requirements.in
 filelock==3.0.12

ci/requirements-conversion.txt

@@ -26,8 +26,6 @@ defusedxml==0.7.1
     #   -r ${INTEL_OPENVINO_DIR}/deployment_tools/model_optimizer/requirements_tf2.txt
 flatbuffers==1.12
     # via tensorflow
-future==0.18.2
-    # via -r tools/model_tools/requirements-caffe2.in
 gast==0.3.3
     # via tensorflow
 google-auth==1.35.0
@@ -84,7 +82,6 @@ oauthlib==3.1.1
 onnx==1.10.1
     # via
     #   -r ${INTEL_OPENVINO_DIR}/deployment_tools/model_optimizer/requirements_onnx.txt
-    #   -r tools/model_tools/requirements-caffe2.in
     #   -r tools/model_tools/requirements-pytorch.in
 opt-einsum==3.3.0
     # via tensorflow
@@ -126,7 +123,6 @@ six==1.15.0
     #   google-auth
     #   google-pasta
     #   grpcio
-    #   h5py
     #   keras-preprocessing
     #   onnx
     #   protobuf
@@ -147,7 +143,6 @@ termcolor==1.1.0
     # via tensorflow
 torch==1.8.1
     # via
-    #   -r tools/model_tools/requirements-caffe2.in
     #   -r tools/model_tools/requirements-pytorch.in
     #   torchvision
 torchvision==0.9.1

ci/requirements-quantization.txt

@@ -20,8 +20,6 @@ defusedxml==0.7.1
     # via
     #   -r ${INTEL_OPENVINO_DIR}/deployment_tools/model_optimizer/requirements_kaldi.txt
     #   -r tools/accuracy_checker/requirements-core.in
-editdistance==0.5.3
-    # via -r tools/accuracy_checker/requirements.in
 fast-ctc-decode==0.3.0
     # via -r tools/accuracy_checker/requirements.in
 filelock==3.0.12

ci/update-requirements.py

@@ -87,7 +87,7 @@ def pc(target, *sources):
     pc('ci/requirements-check-basics.txt',
        'ci/requirements-check-basics.in', 'ci/requirements-documentation.in')
     pc('ci/requirements-conversion.txt',
-        *(f'tools/model_tools/requirements-{suffix}.in' for suffix in ['caffe2', 'pytorch', 'tensorflow']),
+        *(f'tools/model_tools/requirements-{suffix}.in' for suffix in ['pytorch', 'tensorflow']),
         *(openvino_dir / f'deployment_tools/model_optimizer/requirements_{suffix}.txt'
             for suffix in ['caffe', 'mxnet', 'onnx', 'tf2']))
     pc('ci/requirements-demos.txt',

data/datasets.md

@@ -40,7 +40,7 @@ To use this dataset with OMZ tools, make sure `<DATASET_DIR>` contains the follo
 
 ### Datasets in dataset_definitions.yml
 * `imagenet_1000_classes` used for evaluation models trained on ILSVRC 2012 dataset with 1000 classes. (model examples: [`alexnet`](../models/public/alexnet/README.md), [`vgg16`](../models/public/vgg16/README.md))
-* `imagenet_1000_classes_2015` used for evaluation models trained on ILSVRC 2015 dataset with 1000 classes. (model examples: [`se-resnet-152`](../models/public/se-resnet-152/README.md), [`se-resnext-50`](../models/public/se-resnext-50/README.md))
+* `imagenet_1000_classes_2015` used for evaluation models trained on ILSVRC 2015 dataset with 1000 classes. (model examples: [`se-resnet-50`](../models/public/se-resnet-50/README.md), [`se-resnext-50`](../models/public/se-resnext-50/README.md))
 * `imagenet_1001_classes` used for evaluation models trained on ILSVRC 2012 dataset with 1001 classes (background label + original labels). (model examples: [`googlenet-v2-tf`](../models/public/googlenet-v2-tf/README.md), [`resnet-50-tf`](../models/public/resnet-50-tf/README.md))
 
 ## [Common Objects in Context (COCO)](https://cocodataset.org/#home)
@@ -62,9 +62,9 @@ To use this dataset with OMZ tools, make sure `<DATASET_DIR>` contains the follo
 
 ### Datasets in dataset_definitions.yml
 * `ms_coco_mask_rcnn` used for evaluation models trained on COCO dataset for object detection and instance segmentation tasks. Background label + label map with 80 public available object categories are used. Annotations are saved in order of ascending image ID.
-* `ms_coco_detection_91_classes` used for evaluation models trained on COCO dataset for object detection tasks. Background label + label map with 80 public available object categories are used (original indexing to 91 categories is preserved. You can find more information about object categories labels [here](https://tech.amikelive.com/node-718/what-object-categories-labels-are-in-coco-dataset/)). Annotations are saved in order of ascending image ID. (model examples: [`faster_rcnn_resnet50_coco`](../models/public/faster_rcnn_resnet50_coco/README.md), [`ssd_resnet50_v1_fpn_coco`](../models/public/ssd_resnet50_v1_fpn_coco/README.md))
+* `ms_coco_detection_91_classes` used for evaluation models trained on COCO dataset for object detection tasks. Background label + label map with 80 public available object categories are used (original indexing to 91 categories is preserved. You can find more information about object categories labels [here](https://tech.amikelive.com/node-718/what-object-categories-labels-are-in-coco-dataset/)). Annotations are saved in order of ascending image ID. (model examples: [`faster_rcnn_resnet50_coco`](../models/public/faster_rcnn_resnet50_coco/README.md), [`ssd_mobilenet_v1_coco`](../models/public/ssd_mobilenet_v1_coco/README.md))
 * `ms_coco_detection_80_class_with_background` used for evaluation models trained on COCO dataset for object detection tasks. Background label + label map with 80 public available object categories are used. Annotations are saved in order of ascending image ID. (model examples: [`faster-rcnn-resnet101-coco-sparse-60-0001`](../models/intel/faster-rcnn-resnet101-coco-sparse-60-0001/README.md), [`ssd-resnet34-1200-onnx`](../models/public/ssd-resnet34-1200-onnx/README.md))
-* `ms_coco_detection_80_class_without_background` used for evaluation models trained on COCO dataset for object detection tasks. Label map with 80 public available object categories is used. Annotations are saved in order of ascending image ID. (model examples: [`ctdet_coco_dlav0_384`](../models/public/ctdet_coco_dlav0_384/README.md), [`yolo-v3-tf`](../models/public/yolo-v3-tf/README.md))
+* `ms_coco_detection_80_class_without_background` used for evaluation models trained on COCO dataset for object detection tasks. Label map with 80 public available object categories is used. Annotations are saved in order of ascending image ID. (model examples: [`ctdet_coco_dlav0_512`](../models/public/ctdet_coco_dlav0_512/README.md), [`yolo-v3-tf`](../models/public/yolo-v3-tf/README.md))
 * `ms_coco_keypoints` used for evaluation models trained on COCO dataset for human pose estimation tasks. Each annotation stores multiple keypoints for one image. (model examples: [`human-pose-estimation-0001`](../models/intel/human-pose-estimation-0001/README.md))
 * `ms_coco_single_keypoints` used for evaluation models trained on COCO dataset for human pose estimation tasks. Each annotation stores single keypoints for image, so several annotation can be associated to one image. (model examples: [`single-human-pose-estimation-0001`](../models/public/single-human-pose-estimation-0001/README.md))
 

demos/classification_benchmark_demo/cpp/README.md

@@ -38,78 +38,50 @@ omz_converter --list models.lst
 * alexnet
 * caffenet
 * densenet-121
-* densenet-121-caffe2
 * densenet-121-tf
-* densenet-161
-* densenet-161-tf
-* densenet-169
-* densenet-169-tf
-* densenet-201
-* densenet-201-tf
 * dla-34
 * efficientnet-b0
-* efficientnet-b0_auto_aug
 * efficientnet-b0-pytorch
-* efficientnet-b5
-* efficientnet-b5-pytorch
-* efficientnet-b7_auto_aug
-* efficientnet-b7-pytorch
 * googlenet-v1
 * googlenet-v1-tf
 * googlenet-v2
 * googlenet-v3
 * googlenet-v3-pytorch
 * googlenet-v4-tf
 * hbonet-0.25
-* hbonet-0.5
 * hbonet-1.0
 * inception-resnet-v2-tf
 * mixnet-l
 * mobilenet-v1-0.25-128
-* mobilenet-v1-0.50-160
-* mobilenet-v1-0.50-224
 * mobilenet-v1-1.0-224
 * mobilenet-v1-1.0-224-tf
 * mobilenet-v2
 * mobilenet-v2-1.0-224
 * mobilenet-v2-1.4-224
 * mobilenet-v2-pytorch
 * nfnet-f0
-* octave-densenet-121-0.125
-* octave-resnet-101-0.125
-* octave-resnet-200-0.125
 * octave-resnet-26-0.25
-* octave-resnet-50-0.125
-* octave-resnext-101-0.25
-* octave-resnext-50-0.25
-* octave-se-resnet-50-0.125
 * regnetx-3.2gf
 * repvgg-a0
 * repvgg-b1
 * repvgg-b3
 * resnest-50-pytorch
 * resnet-18-pytorch
-* resnet-50-caffe2
 * resnet-50-pytorch
 * resnet-50-tf
 * resnet18-xnor-binary-onnx-0001
 * resnet50-binary-0001
 * rexnet-v1-x1.0
 * se-inception
-* se-resnet-101
-* se-resnet-152
 * se-resnet-50
-* se-resnext-101
 * se-resnext-50
 * shufflenet-v2-x0.5
 * shufflenet-v2-x1.0
 * squeezenet1.0
 * squeezenet1.1
-* squeezenet1.1-caffe2
 * swin-tiny-patch4-window7-224
 * vgg16
 * vgg19
-* vgg19-caffe2
 
 > **NOTE**: Refer to the tables [Intel's Pre-Trained Models Device Support](../../../models/intel/device_support.md) and [Public Pre-Trained Models Device Support](../../../models/public/device_support.md) for the details on models inference support at different devices.
 
@@ -136,10 +108,7 @@ Please note that you should use `<omz_dir>/data/dataset_classes/imagenet_2015.tx
 
 * googlenet-v2
 * se-inception
-* se-resnet-101
-* se-resnet-152
 * se-resnet-50
-* se-resnext-101
 * se-resnext-50
 
 and `<omz_dir>/data/dataset_classes/imagenet_2012.txt` labels file with all other models supported by the demo.

demos/classification_benchmark_demo/cpp/models.lst

@@ -2,75 +2,47 @@
 alexnet
 caffenet
 densenet-121
-densenet-121-caffe2
 densenet-121-tf
-densenet-161
-densenet-161-tf
-densenet-169
-densenet-169-tf
-densenet-201
-densenet-201-tf
 dla-34
 efficientnet-b0
-efficientnet-b0_auto_aug
 efficientnet-b0-pytorch
-efficientnet-b5
-efficientnet-b5-pytorch
-efficientnet-b7_auto_aug
-efficientnet-b7-pytorch
 googlenet-v1
 googlenet-v1-tf
 googlenet-v2
 googlenet-v3
 googlenet-v3-pytorch
 googlenet-v4-tf
 hbonet-0.25
-hbonet-0.5
 hbonet-1.0
 inception-resnet-v2-tf
 mixnet-l
 mobilenet-v1-0.25-128
-mobilenet-v1-0.50-160
-mobilenet-v1-0.50-224
 mobilenet-v1-1.0-224
 mobilenet-v1-1.0-224-tf
 mobilenet-v2
 mobilenet-v2-1.0-224
 mobilenet-v2-1.4-224
 mobilenet-v2-pytorch
 nfnet-f0
-octave-densenet-121-0.125
-octave-resnet-101-0.125
-octave-resnet-200-0.125
 octave-resnet-26-0.25
-octave-resnet-50-0.125
-octave-resnext-101-0.25
-octave-resnext-50-0.25
-octave-se-resnet-50-0.125
 regnetx-3.2gf
 repvgg-a0
 repvgg-b1
 repvgg-b3
 resnest-50-pytorch
 resnet-18-pytorch
-resnet-50-caffe2
 resnet-50-pytorch
 resnet-50-tf
 resnet18-xnor-binary-onnx-0001
 resnet50-binary-0001
 rexnet-v1-x1.0
 se-inception
-se-resnet-101
-se-resnet-152
 se-resnet-50
-se-resnext-101
 se-resnext-50
 shufflenet-v2-x0.5
 shufflenet-v2-x1.0
 squeezenet1.0
 squeezenet1.1
-squeezenet1.1-caffe2
 swin-tiny-patch4-window7-224
 vgg16
 vgg19
-vgg19-caffe2

demos/classification_demo/python/README.md

@@ -34,22 +34,10 @@ omz_converter --list models.lst
 * alexnet
 * caffenet
 * densenet-121
-* densenet-121-caffe2
 * densenet-121-tf
-* densenet-161
-* densenet-161-tf
-* densenet-169
-* densenet-169-tf
-* densenet-201
-* densenet-201-tf
 * dla-34
 * efficientnet-b0
-* efficientnet-b0_auto_aug
 * efficientnet-b0-pytorch
-* efficientnet-b5
-* efficientnet-b5-pytorch
-* efficientnet-b7_auto_aug
-* efficientnet-b7-pytorch
 * efficientnet-v2-b0
 * efficientnet-v2-s
 * googlenet-v1
@@ -59,55 +47,39 @@ omz_converter --list models.lst
 * googlenet-v3-pytorch
 * googlenet-v4-tf
 * hbonet-0.25
-* hbonet-0.5
 * hbonet-1.0
 * inception-resnet-v2-tf
 * mixnet-l
 * mobilenet-v1-0.25-128
-* mobilenet-v1-0.50-160
-* mobilenet-v1-0.50-224
 * mobilenet-v1-1.0-224
 * mobilenet-v1-1.0-224-tf
 * mobilenet-v2
 * mobilenet-v2-1.0-224
 * mobilenet-v2-1.4-224
 * mobilenet-v2-pytorch
 * nfnet-f0
-* octave-densenet-121-0.125
-* octave-resnet-101-0.125
-* octave-resnet-200-0.125
 * octave-resnet-26-0.25
-* octave-resnet-50-0.125
-* octave-resnext-101-0.25
-* octave-resnext-50-0.25
-* octave-se-resnet-50-0.125
 * regnetx-3.2gf
 * repvgg-a0
 * repvgg-b1
 * repvgg-b3
 * resnest-50-pytorch
 * resnet-18-pytorch
-* resnet-50-caffe2
 * resnet-50-pytorch
 * resnet-50-tf
 * resnet18-xnor-binary-onnx-0001
 * resnet50-binary-0001
 * rexnet-v1-x1.0
 * se-inception
-* se-resnet-101
-* se-resnet-152
 * se-resnet-50
-* se-resnext-101
 * se-resnext-50
 * shufflenet-v2-x0.5
 * shufflenet-v2-x1.0
 * squeezenet1.0
 * squeezenet1.1
-* squeezenet1.1-caffe2
 * swin-tiny-patch4-window7-224
 * vgg16
 * vgg19
-* vgg19-caffe2
 
 > **NOTE**: Refer to the tables [Intel's Pre-Trained Models Device Support](../../../models/intel/device_support.md) and [Public Pre-Trained Models Device Support](../../../models/public/device_support.md) for the details on models inference support at different devices.
 
@@ -119,10 +91,7 @@ Please note that you should use `<omz_dir>/data/dataset_classes/imagenet_2015.tx
 
 * googlenet-v2
 * se-inception
-* se-resnet-101
-* se-resnet-152
 * se-resnet-50
-* se-resnext-101
 * se-resnext-50
 
 and `<omz_dir>/data/dataset_classes/imagenet_2012.txt` labels file with all other models supported by the demo.