Model zoo performances

This page lets you discover all of Akida model zoo machine learning models with their respective performances.

Note

The download links provided point towards standard Tensorflow Keras models that must be converted to Akida model using cnn2snn.convert.

image_icon_ref Image domain

Classification

Architecture

Resolution

Dataset

Quantization

Top-1 accuracy

Example

Size (KB)

NPs

Download

MobileNetV1 0.25

160

ImageNet

8/4/4

40.86%

mb_ex

365.4

23

mb_160_25_dl

MobileNetV1 0.5

160

ImageNet

8/4/4

55.94%

mb_ex

1017.1

30

mb_160_50_dl

MobileNetV1

160

ImageNet

8/4/4

66.40%

mb_ex

3554.5

78

mb_160_dl

MobileNetV1 0.25

224

ImageNet

8/4/4

45.12%

mb_ex

366.9

25

mb_224_25_dl

MobileNetV1 0.5

224

ImageNet

8/4/4

59.76%

mb_ex

1075.4

38

mb_224_50_dl

MobileNetV1

224

ImageNet

8/4/4

69.53%

mb_ex

5251.8

123

mb_224_dl

MobileNetV1

160

Cats vs dogs

8/4/4

98.11%

mb_cvd_ex

2767.9

65

mb_cvd_dl

MobileNetV1 0.25

224

Imagenette

8/4/4

91.03%

172.7

22

mb_ite_25_dl

MobileNetV1 0.5

224

Imagenette

8/4/4

95.16%

678.6

32

mb_ite_50_dl

MobileNetV1

224

Imagenette

8/4/4

97.15%

4473.4

110

mb_ite_dl

MobileNetV1 0.5 edge

160

ImageNet

8/4/4

49.69%

mbe_ex

1935.1

38

mbe_160_dl

MobileNetV1 0.5 edge

224

ImageNet

8/4/4

51.83%

mbe_ex

1993.4

46

mbe_224_dl

DS-CNN

224

CIFAR10

4/4/4

93.04%

ds_ex

2493.9

50

ds_dl

VGG-like

224

CIFAR10

2/2/2

90.67%

6443.5

58

vgg_c10_dl

VGG11

224

ImageNet

8/4/4

51.09%

34825.2

21

vgg11_dl

VGG11

224

SIIM-ISIC Melanoma Classification

8/4/4

98.31% - AUROC 0.8020

31804.8

21

vgg_mel_dl

VGG11

224

ODIR-5K Ocular disease recognition

8/4/4

90.53% - AUROC 0.9473

3131.3

21

vgg_odir_dl

VGG11

224

Retinal OCT ocular disease recognition

8/4/4

80.60% - AUROC 0.9768

3131.3

21

vgg_oct_dl

GXNOR

28

MNIST

2/2/1

99.24%

gx_ex

420.8

4

gx_dl

Object detection

Architecture

Resolution

Dataset

Quantization

mAP

Example

Size (KB)

NPs

Download

YOLOv2

224

PASCAL-VOC 2007 - person and car classes

8/4/4

29.39%

yl_voc_ex

2924.0

71

yl_voc_dl

YOLOv2

224

WIDER FACE

8/4/4

71.44%

2915.8

71

yl_wf_dl

Regression

Architecture

Resolution

Dataset

Quantization

MAE

Example

Size (KB)

NPs

Download

VGG-like

32

UTKFace (age estimation)

8/2/2

6.1791

reg_ex

139.8

6

reg_dl

Face recognition

Architecture

Resolution

Dataset

Quantization

Accuracy

Size (KB)

NPs

Download

MobileNetV1 0.5

112x96

CASIA Webface face identification

8/4/4

69.17%

1882.0

21

fid_dl

MobileNetV1 0.5 edge

112x96

CASIA Webface face identification

8/4/4

71.40%

6932.1

35

fide_dl

MobileNetV1 0.5

112x96

LFW face verification

8/4/4

97.27%

643.2

20

fver_dl

audio_icon_ref Audio domain

Keyword spotting

Architecture

Dataset

Quantization

Top-1 accuracy

Example

Size (KB)

NPs

Download

DS-CNN

Google speech command

8/4/4

91.34%

kws_ex

22.8

5

kws_dl

time_icon_ref Time domain

Fault detection

Architecture

Dataset

Quantization

Accuracy

Size (KB)

NPs

Download

Convtiny

CWRU Electric Motor Ball Bearing Fault Diagnosis

8/2/4

98.9%

56.9

4

cwru_dl

Classification

Architecture

Resolution

Dataset

Quantization

Accuracy

Size (KB)

NPs

Download

VGG11

160

Physionet2017 ECG classification

8/4/4

73.89% - AUROC 0.8149

3131.3

21

ecg_dl

pointcloud_icon_ref Point cloud

Classification

Architecture

Dataset

Quantization

Accuracy

Input scaling

Size (KB)

NPs

Download

PointNet++

ModelNet40 3D Point Cloud

8/4/4

84.76%

(127, 127)

528.5

17

p++_dl