Model Zoo

January 18, 2018 · View on GitHub

This page lists model archives that are pre-trained and pre-packaged, ready to be served for inference with MMS. To propose a model for inclusion, please submit a pull request.

Special thanks to the Apache MXNet community whose Model Zoo and Model Examples were used in generating these model archives.

Model File	Type	Dataset	Source	Size	Download
AlexNet	Image Classification	ImageNet	ONNX	233 MB	.model
CaffeNet	Image Classification	ImageNet	MXNet	216 MB	.model
Inception v1	Image Classification	ImageNet	ONNX	27 MB	.model
Inception v3 w/BatchNorm	Image Classification	ImageNet	MXNet	45 MB	.model
LSTM PTB	Language Modeling	PennTreeBank	MXNet	16 MB	.model
Network in Network (NiN)	Image Classification	ImageNet	MXNet	30 MB	.model
ResNet-152	Image Classification	ImageNet	MXNet	241 MB	.model
ResNet-18	Image Classification	ImageNet	MXNet	43 MB	.model
ResNet50-SSD	SSD (Single Shot MultiBox Detector)	ImageNet	MXNet	124 MB	.model
ResNext101-64x4d	Image Classification	ImageNet	MXNet	334 MB	.model
SqueezeNet	Image Classification	ImageNet	ONNX	5 MB	.model
SqueezeNet v1.1	Image Classification	ImageNet	MXNet	5 MB	.model
VGG16	Image Classification	ImageNet	MXNet	490 MB	.model
VGG19	Image Classification	ImageNet	MXNet	509 MB	.model
VGG19	Image Classification	ImageNet	ONNX	548 MB	.model

Details on Each Model

Each model below comes with a basic description, and where available, a link to a scholarly article about the model.

Many of these models use a kitten image to test inference. Use the following to get one that will work:

curl -O https://s3.amazonaws.com/model-server/inputs/kitten.jpg

AlexNet

Type: Image classification trained on ImageNet
Reference: Krizhevsky, et al.
Model Service: mxnet_vision_service.py
Start Server:

mxnet-model-server --models alexnet=https://s3.amazonaws.com/model-server/models/onnx-alexnet/alexnet.model

Run Prediction:

curl -X POST http://127.0.0.1:8080/alexnet/predict -F "input_0=@kitten.jpeg"

CaffeNet

Type: Image classification trained on ImageNet
Reference: Krizhevsky, et al.
Model Service: mxnet_vision_service.py
Start Server:

mxnet-model-server --models caffenet=https://s3.amazonaws.com/model-server/models/caffenet/caffenet.model

Run Prediction:

curl -X POST http://127.0.0.1:8080/caffenet/predict -F "data=@kitten.jpeg"

Inception v1

Type: Image classification trained on ImageNet
Reference: Szegedy, et al., Google
Model Service: mxnet_vision_service.py
Start Server:

mxnet-model-server --models inception-v1=https://s3.amazonaws.com/model-server/models/onnx-inception_v1/inception_v1.model

Run Prediction:

curl -X POST http://127.0.0.1:8080/inception-v1/predict -F "input_0=@kitten.jpeg"

Inception v3

Type: Image classification trained on ImageNet
Reference: Szegedy, et al., Google
Model Service: mxnet_vision_service.py
Start Server:

mxnet-model-server --models inception-bn=https://s3.amazonaws.com/model-server/models/inception-bn/Inception-BN.model

Run Prediction:

curl -X POST http://127.0.0.1:8080/inception-bn/predict -F "data=@kitten.jpeg"

LSTM PTB

Long short-term memory network trained on the PennTreeBank dataset.

Reference: Hochreiter, et al.
Model Service: lstm_ptb_service.py
Start Server:

mxnet-model-server --models lstm_ptb=https://s3.amazonaws.com/model-server/models/lstm_ptb/lstm_ptb.model

Run Prediction:

curl -X POST http://127.0.0.1:8080/lstm_ptb/predict -F "data=[{'input_sentence': 'on the exchange floor as soon as ual stopped trading we <unk> for a panic said one top floor trader'}]"

Network in Network

Type: Image classification trained on ImageNet
Reference: Lin, et al.
Model Service: mxnet_vision_service.py
Start Server:

mxnet-model-server --models nin=https://s3.amazonaws.com/model-server/models/nin/nin.model

Run Prediction:

curl -X POST http://127.0.0.1:8080/nin/predict -F "data=@kitten.jpeg"

ResNet-152

Type: Image classification trained on ImageNet
Reference: Lin, et al.
Model Service: mxnet_vision_service.py
Start Server:

mxnet-model-server --models resnet-152=https://s3.amazonaws.com/model-server/models/resnet-152/resnet-152.model

Run Prediction:

curl -X POST http://127.0.0.1:8080/resnet-152/predict -F "data=@kitten.jpeg"

ResNet-18

Type: Image classification trained on ImageNet
Reference: He, et al.
Model Service: mxnet_vision_service.py
Start Server:

mxnet-model-server --models resnet-18=https://s3.amazonaws.com/model-server/models/resnet-18/resnet-18.model

Run Prediction:

curl -X POST http://127.0.0.1:8080/resnet-18/predict -F "data=@kitten.jpeg"

ResNet50-SSD

Type: Image classification trained on ImageNet
Reference: Liu, et al.
Model Service: ssd_service.py
Start Server:

mxnet-model-server --models SSD=https://s3.amazonaws.com/model-server/models/resnet50_ssd/resnet50_ssd_model.model

Run Prediction:

wget https://www.dphotographer.co.uk/users/21963/thm1024/1337890426_Img_8133.jpg
curl -X POST http://127.0.0.1:8080/SSD/predict -F "data=@1337890426_Img_8133.jpg"

ResNext101-64x4d

Type: Image classification trained on ImageNet
Reference: Xie, et al.
Model Service: mxnet_vision_service.py
Start Server:

mxnet-model-server --models resnext101=https://s3.amazonaws.com/model-server/models/resnext-101-64x4d/resnext-101-64x4d.model

Run Prediction:

curl -X POST http://127.0.0.1:8080/resnext101/predict -F "data=@kitten.jpeg"

SqueezeNet

Type: Image classification trained on ImageNet
Reference: Iandola, et al.
Model Service: mxnet_vision_service.py
Start Server:

mxnet-model-server --models squeezenet=https://s3.amazonaws.com/model-server/models/onnx-squeezenet/squeezenet.model

Run Prediction:

curl -X POST http://127.0.0.1:8080/squeezenet/predict -F "input_0=@kitten.jpeg"

SqueezeNet v1.1

Type: Image classification trained on ImageNet
Reference: Iandola, et al.
Model Service: mxnet_vision_service.py
Start Server:

mxnet-model-server --models squeezenet_v1.1=https://s3.amazonaws.com/model-server/models/squeezenet_v1.1/squeezenet_v1.1.model

Run Prediction:

curl -X POST http://127.0.0.1:8080/squeezenet_v1.1/predict -F "data=@kitten.jpeg"

VGG16

Type: Image classification trained on ImageNet
Reference: Simonyan, et al.
Model Service: mxnet_vision_service.py
Start Server:

mxnet-model-server --models vgg16=https://s3.amazonaws.com/model-server/models/vgg16/vgg16.model

Run Prediction:

curl -X POST http://127.0.0.1:8080/vgg16/predict -F "data=@kitten.jpeg"

VGG19

Type: Image classification trained on ImageNet
Reference: Simonyan, et al.
Model Service: mxnet_vision_service.py
Start Server:

mxnet-model-server --models vgg19=https://s3.amazonaws.com/model-server/models/vgg19/vgg19.model

Run Prediction:

curl -X POST http://127.0.0.1:8080/vgg19/predict -F "data=@kitten.jpeg"

VGG19

Type: Image classification trained on ImageNet (imported from ONNX)
Reference: Simonyan, et al.
Model Service: mxnet_vision_service.py
Start Server:

mxnet-model-server --models vgg19=https://s3.amazonaws.com/model-server/models/onnx-vgg19/vgg19.model

Run Prediction:

curl -X POST http://127.0.0.1:8080/vgg19/predict -F "input_0=@kitten.jpeg"