Models

synopsis

Global models available. All pre-trained version of the following models have been trained using ImageNet. Input image expected (3, H, W) where H and W are at least 224. Images expected to have a range between 0 and 1 and normalized using mean = [0.485, 0.456, 0.406] and std = [0.229, 0.224, 0.225]. Output vector of 1000 classes.

npu.vision.models.alexnet = <npu.core.Model.Model object>
Alexnet

AlexNet model. Access pre-trained version alexnet(pretrained=True)

npu.vision.models.densenet161 = <npu.core.Model.Model object>
Densenet161

Densenet-161 model. Access pre-trained version densenet161(pretrained=True)

npu.vision.models.googlenet = <npu.core.Model.Model object>
Googlenet

GoogleNet model. Access pre-trained version googlenet(pretrained=True)

npu.vision.models.inception_v3 = <npu.core.Model.Model object>
Inception_v3

Inception v3 model. Access pretrained version inception_v3(pretrained=True)

npu.vision.models.mnasnet1_0 = <npu.core.Model.Model object>
Mnasnet1_0

MNASNet model. Access pre-trained version mnasnet1_0(pretrained=True)

npu.vision.models.mobilenet_v2 = <npu.core.Model.Model object>
Mobilenet_v2

MobileNet v2 model. Access pretrained version mobilenet_v2(pretrained=True)

npu.vision.models.resnet101 = <npu.core.Model.Model object>
Resnet101

ResNet-101 model. Access pretrained version resnet101(pretrained=True)

npu.vision.models.resnet152 = <npu.core.Model.Model object>
Resnet152

ResNet-152 model. Access pretrained version resnet152(pretrained=True)

npu.vision.models.resnet18 = <npu.core.Model.Model object>
Resnet18

ResNet-18 model. Access pre-trained version resnet18(pretrained=True)

npu.vision.models.resnet34 = <npu.core.Model.Model object>
Resnet34

ResNet-34 model. Access pretrained version resnet34(pretrained=True)

npu.vision.models.resnet50 = <npu.core.Model.Model object>
Resnet50

ResNet-50 model. Access pretrained version resnet50(pretrained=True)

npu.vision.models.resnext101_32x8d = <npu.core.Model.Model object>
Resnext101_32x8d

ResNext-101 model. Access pretrained version resnext101_32x8d(pretrained=True)

npu.vision.models.resnext50_32x4d = <npu.core.Model.Model object>
Resnext50_32x4d

ResNext-50 model. Access pretrained version resnext50_32x4d(pretrained=True)

npu.vision.models.shufflenet_v2_x1_0 = <npu.core.Model.Model object>
Shufflenet_v2_x1_0

ShuffleNetV2 model with 1.0x output channels.

Access pre-trained version shufflenet_v2_x1_0(pretrained=True)

npu.vision.models.vgg16 = <npu.core.Model.Model object>
Vgg16

VGG-16 model with batchnorm layers. Access pre-trained version vgg16(pretrained=True)

npu.vision.models.vgg19 = <npu.core.Model.Model object>
Vgg19

VGG-19 model with batchnorm layers. Access pre-trained version vgg19(pretrained=True)

npu.vision.models.wide_resnet101_2 = <npu.core.Model.Model object>
Wide_resnet101_2

Wide ResNet-101-2 model. Access pretrained version wide_resnet101_2(pretrained=True)

npu.vision.models.wide_resnet50_2 = <npu.core.Model.Model object>
Wide_resnet50_2

Wide ResNet-50-2 model. Access pretrained version wide_resnet50_2(pretrained=True)