Overview of OpenVINO™ Toolkit Intel’s Pre-Trained Models

OpenVINO toolkit provides a set of Intel’s pre-trained models that you can use for learning and demo purposes or for developing deep learning software. Most recent version is available in the repo on GitHub. The table Intel’s Pre-Trained Models Device Support summarizes devices supported by each model.

The models can be downloaded via Model Downloader.

Object Detection Models

Model Name

Framework

Accuracy

GFlops

mParams

faster-rcnn-resnet101-coco-sparse-60-0001

TensorFlow

38.74%

849.91

52.79

face-detection-adas-0001

Caffe

94.10%

2.835

1.053

face-detection-retail-0004

Caffe

83.00%

1.067

0.588

face-detection-retail-0005

PyTorch

84.52%

0.982

1.021

face-detection-0200

PyTorch

86.74%

0.786

1.828

face-detection-0202

PyTorch

91.94%

1.767

1.842

face-detection-0204

PyTorch

92.89%

2.406

1.851

face-detection-0205

PyTorch

93.57%

2.853

2.392

face-detection-0206

PyTorch

94.27%

339.60

69.92

person-detection-retail-0002

Caffe

80.14%

12.427

3.244

person-detection-retail-0013

Caffe

88.62%

2.300

0.723

person-detection-action-recognition-0005

Caffe

83.8%

7.140

1.951

person-detection-action-recognition-0006

TensorFlow

80.74%

8.225

2.001

person-detection-action-recognition-teacher-0002

Caffe

72.4%

7.140

1.951

person-detection-raisinghand-recognition-0001

Caffe

90.5%

7.138

1.951

person-detection-0200

PyTorch

0.244

0.786

1.817

person-detection-0201

PyTorch

0.299

1.768

1.817

person-detection-0202

PyTorch

0.328

3.143

1.817

person-detection-0203

PyTorch

0.408

6.519

2.394

person-detection-0106

PyTorch

0.442

404.264

71.565

pedestrian-detection-adas-0002

Caffe

88%

2.836

1.165

pedestrian-and-vehicle-detector-adas-0001

Caffe

90%

3.974

1.650

vehicle-detection-adas-0002

Caffe

90.6%

2.798

1.079

vehicle-detection-0200

PyTorch

0.254

0.786

1.817

vehicle-detection-0201

PyTorch

0.323

1.768

1.817

vehicle-detection-0202

PyTorch

0.36

3.143

1.817

person-vehicle-bike-detection-crossroad-0078

Caffe

65.12%

3.964

1.178

person-vehicle-bike-detection-crossroad-1016

PyTorch

62.55%

3.560

2.887

person-vehicle-bike-detection-crossroad-yolov3-1020

Keras

48.89%

65.98

61.92

person-vehicle-bike-detection-2000

PyTorch

0.165

0.787

1.821

person-vehicle-bike-detection-2001

PyTorch

0.226

1.770

1.821

person-vehicle-bike-detection-2002

PyTorch

0.248

3.146

1.821

person-vehicle-bike-detection-2003

PyTorch

0.336

6.550

2.416

person-vehicle-bike-detection-2004

PyTorch

0.274

1.811

2.327

vehicle-license-plate-detection-barrier-0106

TensorFlow

99.65%

0.349

0.634

product-detection-0001

PyTorch

0.715

3.598

3.212

person-detection-asl-0001

PyTorch

80.0%

0.986

1.338

yolo-v2-ava-0001

TensorFlow

63.9%

29.421

50.6451

yolo-v2-ava-sparse-35-0001

TensorFlow

63.71%

29.421

50.6451

yolo-v2-ava-sparse-70-0001

TensorFlow

62.9%

29.421

50.6451

yolo-v2-tiny-ava-0001

TensorFlow

35.37%

6.9949

15.8587

yolo-v2-tiny-ava-sparse-30-0001

TensorFlow

36.33%

6.9949

15.8587

yolo-v2-tiny-ava-sparse-60-0001

TensorFlow

35.32%

6.9949

15.8587

yolo-v2-tiny-vehicle-detection-0001

Keras

88.64%

5.424

11.229

Object Recognition Models

Model Name

Framework

Accuracy

GFlops

mParams

age-gender-recognition-retail-0013

Caffe

95.80%

0.094

2.138

head-pose-estimation-adas-0001

Caffe

0.105

1.911

license-plate-recognition-barrier-0001

Caffe

88.58%

0.328

1.218

vehicle-attributes-recognition-barrier-0039

Caffe

81.15%

0.126

0.626

vehicle-attributes-recognition-barrier-0042

PyTorch

82.71%

0.462

11.177

emotions-recognition-retail-0003

Caffe

70.20%

0.126

2.483

landmarks-regression-retail-0009

PyTorch

0.0705

0.021

0.191

facial-landmarks-35-adas-0002

Caffe

0.042

4.595

person-attributes-recognition-crossroad-0230

PyTorch

0.174

0.735

person-attributes-recognition-crossroad-0234

PyTorch

2.167

23.510

person-attributes-recognition-crossroad-0238

PyTorch

1.034

21.797

gaze-estimation-adas-0002

PyTorch

0.139

1.882

Reidentification Models

Model Name

Framework

Accuracy

GFlops

mParams

face-reidentification-retail-0095

PyTorch

0.9947

0.588

1.107

person-reidentification-retail-0288

PyTorch

86.1%

0.174

0.183

person-reidentification-retail-0287

PyTorch

92.9%

0.564

0.595

person-reidentification-retail-0286

PyTorch

94.8 %

1.170

1.234

person-reidentification-retail-0277

PyTorch

96.2 %

1.993

2.103

Semantic Segmentation Models

Model Name

Framework

Accuracy

GFlops

mParams

road-segmentation-adas-0001

PyTorch

0.901

4.770

0.184

semantic-segmentation-adas-0001

Caffe

0.6907

58.572

6.686

unet-camvid-onnx-0001

PyTorch

71.95%

260.1

31.03

icnet-camvid-ava-0001

PyTorch

75.42%

75.818

26.704

icnet-camvid-ava-sparse-30-0001

PyTorch

75.87%

75.818

26.704

icnet-camvid-ava-sparse-60-0001

PyTorch

75.79%

75.818

26.704

Instance Segmentation Models

Model Name

Framework

Accuracy

GFlops

mParams

instance-segmentation-security-0002

PyTorch

40.8%

423.08

48.373

instance-segmentation-security-0091

PyTorch

45.8%

828.63

101.24

instance-segmentation-security-0228

PyTorch

38.85%

147.24

49.833

instance-segmentation-security-1039

PyTorch

32.9%

13.97

10.567

instance-segmentation-security-1040

PyTorch

35.0%

29.33

13.567

Human Pose Estimation Models

Model Name

Framework

Accuracy

GFlops

mParams

human-pose-estimation-0001

Caffe

42.8%

15.435

4.099

human-pose-estimation-0005

PyTorch

45.6%

5.9206

8.1504

human-pose-estimation-0006

PyTorch

51.1%

8.844

8.1504

human-pose-estimation-0007

PyTorch

54.3%

14.3253

8.1504

Image Processing

Model Name

Framework

Accuracy

GFlops

mParams

single-image-super-resolution-1032

PyTorch

11.654

0.030

single-image-super-resolution-1033

PyTorch

16.062

0.030

text-image-super-resolution-0001

PyTorch

1.379

0.030

Text Detection

Model Name

Framework

Accuracy

GFlops

mParams

text-detection-0003

TensorFlow

82.12%

51.256

6.747

text-detection-0004

TensorFlow

79.43%

23.305

4.328

horizontal-text-detection-0001

PyTorch

88.45%

7.718

2.259

Text Recognition

Model Name

Framework

Accuracy

GFlops

mParams

text-recognition-0012

TensorFlow

0.8818

1.485

5.568

text-recognition-0014

PyTorch

0.8887

0.2726

1.4187

text-recognition-0015 - encoder

PyTorch

12.4

398

text-recognition-0015 - decoder

PyTorch

0.03

4.33

handwritten-score-recognition-0003

TensorFlow

98.83%

0.792

5.555

handwritten-japanese-recognition-0001

PyTorch

117.136

117.136

15.31

handwritten-simplified-chinese-recognition-0001

PyTorch

75.31%

134.513

17.270

formula-recognition-medium-scan-0001 - encoder

PyTorch

81.5%

16.56

1.86

formula-recognition-medium-scan-0001 - decoder

PyTorch

81.5%

1.69

2.56

formula-recognition-polynomials-handwritten-0001 - encoder

PyTorch

70.5%

12.8447

0.2017

formula-recognition-polynomials-handwritten-0001 - decoder

PyTorch

70.5%

0.2017

2.5449

Text Spotting

Model Name

Framework

Accuracy

GFlops

mParams

text-spotting-0005 - detector

PyTorch

71.29%

184.495

27.010

text-spotting-0005 - recognizer-encoder

PyTorch

71.29%

2.082

1.328

text-spotting-0005 - recognizer-decoder

PyTorch

71.29%

0.106

0.283

Action Recognition Models

Model Name

Framework

Accuracy

GFlops

mParams

driver-action-recognition-adas-0002 - encoder

PyTorch

0.676

2.863

driver-action-recognition-adas-0002 - decoder

PyTorch

0.147

4.205

action-recognition-0001 - encoder

PyTorch

7.340

21.276

action-recognition-0001 - decoder

PyTorch

0.147

4.405

asl-recognition-0004

PyTorch

0.847

6.660

4.133

common-sign-language-0002

PyTorch

98.00%

4.227

4.113

weld-porosity-detection-0001

PyTorch

97.14%

3.636

11.173

Image Retrieval

Model Name

Framework

Accuracy

GFlops

mParams

image-retrieval-0001

TensorFlow

0.834

0.613

2.535

Compressed models

Model Name

Framework

Accuracy

GFlops

mParams

resnet50-binary-0001

PyTorch

70.69%

1.002

7.446

resnet18-xnor-binary-onnx-0001

PyTorch

61.71%

Question Answering

Model Name

Framework

Accuracy

GFlops

mParams

bert-large-uncased-whole-word-masking-squad-0001

PyTorch

87.20%

246.93

333.96

bert-large-uncased-whole-word-masking-squad-int8-0001

PyTorch

86.36%

246.93

333.96

bert-large-uncased-whole-word-masking-squad-emb-0001

PyTorch

90.5%

246.93

333.96

bert-small-uncased-whole-word-masking-squad-0001

PyTorch

85.04%

23.9

57.94

bert-small-uncased-whole-word-masking-squad-0002

PyTorch

85.4%

23.9

41.1

bert-small-uncased-whole-word-masking-squad-int8-0002

PyTorch

84.4%

23.9

41.1

bert-small-uncased-whole-word-masking-squad-emb-int8-0001

PyTorch

87.6%

23.9

41.1

Machine Translation

Model Name

Framework

Accuracy

GFlops

mParams

machine-translation-nar-en-ru-0001

PyTorch

21.6%

23.17

69.29

machine-translation-nar-ru-en-0001

PyTorch

22.8%

23.17

69.29

machine-translation-nar-en-de-0002

PyTorch

17.7%

23.19

77.47

machine-translation-nar-de-en-0002

PyTorch

21.4%

23.19

77.47

Text To Speech

Model Name

Framework

Accuracy

GFlops

mParams

text-to-speech-en-0001 - duration-prediction

PyTorch

15.84

13.569

text-to-speech-en-0001 - regression

PyTorch

7.65

4.96

text-to-speech-en-0001 - generation

PyTorch

48.38

12.77

text-to-speech-en-multi-0001 - duration-prediction

PyTorch

28.75

26.18

text-to-speech-en-multi-0001 - regression

PyTorch

7.81

5.12

text-to-speech-en-multi-0001 - generation

PyTorch

48.38

12.77

Noise suppression

Model Name

Framework

Accuracy

GFlops

mParams

noise-suppression-poconetlike-0001

PyTorch

1.2

7.22

Time Series Forecasting

Model Name

Framework

Accuracy

GFlops

mParams

time-series-forecasting-electricity-0001

PyTorch

0.40

2.26