def inception_v3(pretrained=False, model_root=None, **kwargs):
if pretrained:
if 'transform_input' not in kwargs:
kwargs['transform_input'] = True
model = Inception3(**kwargs)
misc.load_state_dict(model, model_urls['inception_v3_google'], model_root)
return model
return Inception3(**kwargs)
def squeezenet1_0(pretrained=False, model_root=None, **kwargs):
r"""SqueezeNet model architecture from the `"SqueezeNet: AlexNet-level
accuracy with 50x fewer parameters and <0.5MB model size"
<https://arxiv.org/abs/1602.07360>`_ paper.
"""
model = SqueezeNet(version=1.0, **kwargs)
if pretrained:
misc.load_state_dict(model, model_urls['squeezenet1_0'], model_root)
return model
def squeezenet1_0(pretrained=False, model_root=None, **kwargs):
r"""SqueezeNet model architecture from the `"SqueezeNet: AlexNet-level
accuracy with 50x fewer parameters and <0.5MB model size"
<https://arxiv.org/abs/1602.07360>`_ paper.
"""
model = SqueezeNet(version=1.0, **kwargs)
if pretrained:
misc.load_state_dict(model, model_urls['squeezenet1_0'], model_root)
return model
def inception_v3(pretrained=False, model_root=None, **kwargs):
if pretrained:
if 'transform_input' not in kwargs:
kwargs['transform_input'] = True
model = Inception3(**kwargs)
misc.load_state_dict(model, model_urls['inception_v3_google'], model_root)
return model
return Inception3(**kwargs)
def squeezenet1_1(pretrained=False, model_root=None, **kwargs):
r"""SqueezeNet 1.1 model from the `official SqueezeNet repo
<https://github.com/DeepScale/SqueezeNet/tree/master/SqueezeNet_v1.1>`_.
SqueezeNet 1.1 has 2.4x less computation and slightly fewer parameters
than SqueezeNet 1.0, without sacrificing accuracy.
"""
model = SqueezeNet(version=1.1, **kwargs)
if pretrained:
misc.load_state_dict(model, model_urls['squeezenet1_1'], model_root)
return model
def squeezenet1_1(pretrained=False, model_root=None, **kwargs):
r"""SqueezeNet 1.1 model from the `official SqueezeNet repo
<https://github.com/DeepScale/SqueezeNet/tree/master/SqueezeNet_v1.1>`_.
SqueezeNet 1.1 has 2.4x less computation and slightly fewer parameters
than SqueezeNet 1.0, without sacrificing accuracy.
"""
model = SqueezeNet(version=1.1, **kwargs)
if pretrained:
misc.load_state_dict(model, model_urls['squeezenet1_1'], model_root)
return model
def resnet152(pretrained=False, model_root=None, **kwargs):
model = ResNet(Bottleneck, [3, 8, 36, 3], **kwargs)
if pretrained:
misc.load_state_dict(model, model_urls['resnet152'], model_root)
return model
def resnet34(pretrained=False, model_root=None, **kwargs):
model = ResNet(BasicBlock, [3, 4, 6, 3], **kwargs)
if pretrained:
misc.load_state_dict(model, model_urls['resnet34'], model_root)
return model
def resnet152(pretrained=False, model_root=None, **kwargs):
model = ResNet(Bottleneck, [3, 8, 36, 3], **kwargs)
if pretrained:
misc.load_state_dict(model, model_urls['resnet152'], model_root)
return model
def resnet34(pretrained=False, model_root=None, **kwargs):
model = ResNet(BasicBlock, [3, 4, 6, 3], **kwargs)
if pretrained:
misc.load_state_dict(model, model_urls['resnet34'], model_root)
return model
