Skip to content

Latest commit

 

History

History
136 lines (121 loc) · 2.66 KB

model-compatibility.md

File metadata and controls

136 lines (121 loc) · 2.66 KB

Model compatibility

AWS Panorama uses Amazon SageMaker Neo to compile your machine learning models for AWS Panorama Appliances. Here is a list of commonly used computer vision models that have been tested with Amazon SageMaker Neo for compilation. You can use these models for learning and demo purposes or for developing your machine learning applications.

Darknet

  • resnet50
  • yolov2
  • yolov2_tiny
  • yolov3_416
  • yolov3_tiny

GluonCV (MXnet)

  • DenseNet201
  • GoogLeNet
  • InceptionV3
  • MobileNet0.75
  • MobileNet1.0
  • MobileNetV2_0.5
  • MobileNetV2_1.0
  • MobileNetV3_Large
  • MobileNetV3_Small
  • ResNeSt50
  • ResNet18_v1
  • ResNet18_v2
  • ResNet50_v1
  • ResNet50_v2
  • SENet_154
  • SE_ResNext50_32x4d
  • SqueezeNet1.0
  • SqueezeNet1.1
  • Xception
  • darknet53
  • resnet18_v1b_0.89
  • resnet50_v1d_0.11
  • resnet50_v1d_0.86
  • ssd_512_mobilenet1.0_coco
  • ssd_512_mobilenet1.0_voc
  • ssd_512_resnet50_v1_coco
  • yolo3_darknet53_coco
  • yolo3_mobilenet1.0_coco

Keras

  • DenseNet121
  • DenseNet201
  • mobilenet_v1
  • mobilenet_v2
  • resnet50_v1
  • resnet50_v2

ONNX

  • mobilenetv2-1.0
  • resnet152v1
  • resnet18v1
  • resnet18v2
  • resnet50v1
  • resnet50v2
  • squeezenet1.1

PyTorch

  • densenet121
  • resnet152
  • resnet18
  • resnet50
  • squeezenet1.0
  • squeezenet1.1
  • vgg16_bn
  • yolov4
  • faster r-cnn resnet-50 fpn
  • mask r-cnn resnet-50 fpn

TensorFlow

  • DenseNet121
  • DenseNet201
  • MobileNet
  • MobileNetV2
  • NASNetLarge
  • NASNetMobile
  • ResNet101
  • ResNet101V2
  • ResNet152
  • ResNet152V2
  • ResNet50
  • ResNet50V2
  • Xception
  • mobilenet100_v1
  • mobilenet100_v2.0
  • mobilenet130_v2
  • mobilenet140_v2
  • resnet50_v1.5
  • resnet50_v2
  • squeezenet
  • faster_rcnn_inception_resnet_v2_atrous_coco
  • faster_rcnn_inception_resnet_v2_atrous_lowproposals_coco
  • faster_rcnn_inception_v2_coco
  • faster_rcnn_nas
  • faster_rcnn_nas_lowproposals_coco
  • faster_rcnn_resnet101_coco
  • faster_rcnn_resnet101_lowproposals_coco
  • faster_rcnn_resnet50_coco
  • faster_rcnn_resnet50_lowproposals_coco
  • mask_rcnn_inception_resnet_v2_atrous_coco
  • mask_rcnn_resnet101_atrous_coco
  • mask_rcnn_resnet50_atrous_coco
  • rfcn_resnet101_coco
  • ssd_mobilenet_v1_0.75_depth_coco
  • ssd_mobilenet_v1_0.75_depth_quantized_coco
  • ssd_mobilenet_v1_coco
  • ssd_mobilenet_v1_fpn_coco
  • ssd_mobilenet_v1_ppn_coco
  • ssd_mobilenet_v2_coco

TensorFlow-Lite

  • densenet_2018_04_27
  • inception_resnet_v2_2018_04_27
  • inception_v3_2018_04_27
  • mnasnet_0.5_224_09_07_2018
  • mnasnet_1.0_224_09_07_2018
  • mnasnet_1.3_224_09_07_2018
  • mobilenet_v1_0.25_128
  • mobilenet_v1_0.25_224
  • mobilenet_v1_0.5_128
  • mobilenet_v1_0.5_224
  • mobilenet_v1_0.75_128
  • mobilenet_v1_0.75_224
  • mobilenet_v1_1.0_128
  • mobilenet_v1_1.0_192
  • mobilenet_v2_1.0_224
  • resnet_v2_101
  • squeezenet_2018_04_27