Add openvino interpreter samples (#159)

* Add OpenVINO interpreter samples * Move openvino_interpreter_samples into openvino/samples and rename openvino_launcher.py into openvino/launcher.py.
openvinotoolkit · Mar 17, 2021 · a698cac · a698cac
1 parent 944e9f9
commit a698cac
Show file tree

Hide file tree

Showing 14 changed files with 1,659 additions and 15 deletions.
diff --git a/datumaro/plugins/openvino/README.md b/datumaro/plugins/openvino/README.md
@@ -0,0 +1,90 @@
+# OpenVINO™ Inference Interpreter 
+Interpreter samples to parse OpenVINO™ inference outputs.
+
+## Models supported from interpreter samples
+There are detection and image classification examples. 
+
+- Detection (SSD-based)
+  - Intel Pre-trained Models > Object Detection  
+    - [face-detection-0200](https://docs.openvinotoolkit.org/latest/omz_models_intel_face_detection_0200_description_face_detection_0200.html)
+    - [face-detection-0202](https://docs.openvinotoolkit.org/latest/omz_models_intel_face_detection_0202_description_face_detection_0202.html)
+    - [face-detection-0204](https://docs.openvinotoolkit.org/latest/omz_models_intel_face_detection_0204_description_face_detection_0204.html)
+    - [person-detection-0200](https://docs.openvinotoolkit.org/latest/omz_models_intel_person_detection_0200_description_person_detection_0200.html)
+    - [person-detection-0201](https://docs.openvinotoolkit.org/latest/omz_models_intel_person_detection_0201_description_person_detection_0201.html)
+    - [person-detection-0202](https://docs.openvinotoolkit.org/latest/omz_models_intel_person_detection_0202_description_person_detection_0202.html) 
+    - [person-vehicle-bike-detection-2000](https://docs.openvinotoolkit.org/latest/omz_models_intel_person_vehicle_bike_detection_2000_description_person_vehicle_bike_detection_2000.html)
+    - [person-vehicle-bike-detection-2001](https://docs.openvinotoolkit.org/latest/omz_models_intel_person_vehicle_bike_detection_2001_description_person_vehicle_bike_detection_2001.html)
+    - [person-vehicle-bike-detection-2002](https://docs.openvinotoolkit.org/latest/omz_models_intel_person_vehicle_bike_detection_2002_description_person_vehicle_bike_detection_2002.html)
+    - [vehicle-detection-0200](https://docs.openvinotoolkit.org/latest/omz_models_intel_vehicle_detection_0200_description_vehicle_detection_0200.html)
+    - [vehicle-detection-0201](https://docs.openvinotoolkit.org/latest/omz_models_intel_vehicle_detection_0201_description_vehicle_detection_0201.html)
+    - [vehicle-detection-0202](https://docs.openvinotoolkit.org/latest/omz_models_intel_vehicle_detection_0202_description_vehicle_detection_0202.html)
+
+  - Public Pre-Trained Models(OMZ) > Object Detection
+    - [ssd_mobilenet_v1_coco](https://docs.openvinotoolkit.org/latest/omz_models_public_ssd_mobilenet_v1_coco_ssd_mobilenet_v1_coco.html)
+    - [ssd_mobilenet_v2_coco](https://docs.openvinotoolkit.org/latest/omz_models_public_ssd_mobilenet_v2_coco_ssd_mobilenet_v2_coco.html)
+
+- Image Classification
+  - Public Pre-Trained Models(OMZ) > Classification
+    - [mobilenet-v2-pytorch](https://docs.openvinotoolkit.org/latest/omz_models_public_mobilenet_v2_pytorch_mobilenet_v2_pytorch.html)
+
+You can find more OpenVINO™ Trained Models [here](https://docs.openvinotoolkit.org/latest/omz_models_intel_index.html)
+To run the inference with OpenVINO™, the model format should be Intermediate Representation(IR). 
+For the Caffe/TensorFlow/MXNet/Kaldi/ONNX models, please see the [Model Conversion Instruction](https://docs.openvinotoolkit.org/latest/openvino_docs_MO_DG_prepare_model_convert_model_Converting_Model.html)
+
+You need to implement your own interpreter samples to support the other OpenVINO™ Trained Models.
+
+## Model download
+- Prerequisites
+    - OpenVINO™ (To install OpenVINO™, please see the [OpenVINO™ Installation Instruction](https://docs.openvinotoolkit.org/latest/openvino_docs_install_guides_installing_openvino_linux.html))
+    - OpenVINO™ models (To download OpenVINO™ models, please see the [Model Downloader Instruction](https://docs.openvinotoolkit.org/latest/omz_tools_downloader_README.html))
+    - PASCAL VOC 2012 dataset (To download VOC 2012 dataset, please go [VOC2012 download](http://host.robots.ox.ac.uk/pascal/VOC/voc2012/#devkit))
+
+  ```bash
+  # cd <openvino_dir>/deployment_tools/open_model_zoo/tools/downloader
+  # ./downloader.py --name <model_name>
+  #
+  # Examples
+  cd /opt/intel/openvino/deployment_tools/open_model_zoo/tools/downloader
+  ./downloader.py --name face-detection-0200
+  ```
+
+## Model inference
+- Prerequisites: 
+    - OpenVINO™ (To install OpenVINO™, please see the [OpenVINO™ Installation Instruction](https://docs.openvinotoolkit.org/latest/openvino_docs_install_guides_installing_openvino_linux.html))
+    - Datumaro (To install Datumaro, please see the [User Manual](docs/user_manual.md))
+    - OpenVINO™ models (To download OpenVINO™ models, please see the [Model Downloader Instruction](https://docs.openvinotoolkit.org/latest/omz_tools_downloader_README.html))
+    - PASCAL VOC 2012 dataset (To download VOC 2012 dataset, please go [VOC2012 download](http://host.robots.ox.ac.uk/pascal/VOC/voc2012/#devkit))
+
+- To run the inference with OpenVINO™ models and the interpreter samples, please follow the instructions below.
+
+  ```bash
+  # source <openvino_dir>/bin/setupvars.sh
+  # datum create -o <proj_dir>
+  # datum model add -l <launcher> -p <proj_dir> --copy -- -d <path_to_xml> -w <path_to_bin> -i <path_to_interpreter_script>
+  # datum add path -p <proj_dir> -f <format> <path_to_dataset>
+  # datum model run -p <proj_dir> -m model-0
+  #
+  # Examples
+  # Detection> ssd_mobilenet_v2_coco
+  source /opt/intel/openvino/bin/setupvars.sh
+  cd datumaro/plugins/openvino
+  datum create -o proj_ssd_mobilenet_v2_coco_detection
+  datum model add -l openvino -p proj_ssd_mobilenet_v2_coco_detection --copy -- \
+      --output-layers=do_ExpandDims_conf/sigmoid \
+      -d model/ssd_mobilenet_v2_coco.xml \
+      -w model/ssd_mobilenet_v2_coco.bin \
+      -i samples/ssd_mobilenet_coco_detection_interp.py
+  datum add path -p proj_ssd_mobilenet_v2_coco_detection -f voc VOCdevkit/
+  datum model run -p proj_ssd_mobilenet_v2_coco_detection -m model-0
+
+  # Classification> mobilenet-v2-pytorch
+  source /opt/intel/openvino/bin/setupvars.sh
+  cd datumaro/plugins/openvino
+  datum create -o proj_mobilenet_v2_classification
+  datum model add -l openvino -p proj_mobilenet_v2_classification --copy -- \
+      -d model/mobilenet-v2-pytorch.xml \
+      -w model/mobilenet-v2-pytorch.bin \
+      -i samples/mobilenet_v2_pytorch_interp.py
+  datum add path -p proj_mobilenet_v2_classification -f voc VOCdevkit/
+  datum model run -p proj_mobilenet_v2_classification -m model-0
+  ```
diff --git a/datumaro/plugins/openvino/__init__.py b/datumaro/plugins/openvino/__init__.py
diff --git a/datumaro/plugins/openvino_launcher.py → datumaro/plugins/openvino/launcher.py b/datumaro/plugins/openvino_launcher.py → datumaro/plugins/openvino/launcher.py
diff --git a/datumaro/plugins/openvino/samples/coco.class b/datumaro/plugins/openvino/samples/coco.class
@@ -0,0 +1,91 @@
+person
+bicycle
+car
+motorcycle
+airplane
+bus
+train
+truck
+boat
+trafficlight
+firehydrant
+streetsign
+stopsign
+parkingmeter
+bench
+bird
+cat
+dog
+horse
+sheep
+cow
+elephant
+bear
+zebra
+giraffe
+hat
+backpack
+umbrella
+shoe
+eyeglasses
+handbag
+tie
+suitcase
+frisbee
+skis
+snowboard
+sportsball
+kite
+baseballbat
+baseballglove
+skateboard
+surfboard
+tennisracket
+bottle
+plate
+wineglass
+cup
+fork
+knife
+spoon
+bowl
+banana
+apple
+sandwich
+orange
+broccoli
+carrot
+hotdog
+pizza
+donut
+cake
+chair
+couch
+pottedplant
+bed
+mirror
+diningtable
+window
+desk
+toilet
+door
+tv
+laptop
+mouse
+remote
+keyboard
+cellphone
+microwave
+oven
+toaster
+sink
+refrigerator
+blender
+book
+clock
+vase
+scissors
+teddybear
+hairdrier
+toothbrush
+hairbrush