Skip to content

P
PaddleClas

PaddlePaddle / PaddleClas
大约 2 年 前同步成功

通知 118
Star 4999
Fork 1114
P
PaddleClas
提交 0102ad56 编写于 作者: L lyuwenyu
浏览文件
操作

Skip __init__ to release unnecessary dependents

上级 67a7e585
隐藏空白更改
内联 并排
Showing with 795 addition and 816 deletion
hubconf.py
浏览文件 @ 0102ad56
...@@ -12,821 +12,800 @@ ...@@ -12,821 +12,800 @@
# See the License for the specific language governing permissions and # See the License for the specific language governing permissions and
# limitations under the License. # limitations under the License.
dependencies = ['paddle', 'numpy'] dependencies = ['paddle']
import paddle import paddle
from ppcls.modeling import architectures import os
import sys
def _load_pretrained_parameters(model, name):
url = 'https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/{}_pretrained.pdparams'.format( class _SysPathG(object):
name) def __enter__(self, ):
path = paddle.utils.download.get_weights_path_from_url(url) sys.path.insert(0,
model.set_state_dict(paddle.load(path)) os.path.join(
return model os.path.dirname(os.path.abspath(__file__)),
'ppcls', 'modeling'))
def alexnet(pretrained=False, **kwargs): def __exit__(self, type, value, traceback):
""" sys.path.pop(0)
AlexNet
Args:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. with _SysPathG():
kwargs: import architectures
class_dim: int=1000. Output dim of last fc layer.
Returns: def _load_pretrained_parameters(model, name):
model: nn.Layer. Specific `AlexNet` model depends on args. url = 'https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/{}_pretrained.pdparams'.format(
""" name)
model = architectures.AlexNet(**kwargs) path = paddle.utils.download.get_weights_path_from_url(url)
if pretrained: model.set_state_dict(paddle.load(path))
model = _load_pretrained_parameters(model, 'AlexNet') return model
return model def alexnet(pretrained=False, **kwargs):
"""
AlexNet
def vgg11(pretrained=False, **kwargs): Args:
""" pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
VGG11 kwargs:
Args: class_dim: int=1000. Output dim of last fc layer.
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. Returns:
kwargs: model: nn.Layer. Specific `AlexNet` model depends on args.
class_dim: int=1000. Output dim of last fc layer. """
stop_grad_layers: int=0. The parameters in blocks which index larger than `stop_grad_layers`, will be set `param.trainable=False` model = architectures.AlexNet(**kwargs)
Returns: if pretrained:
model: nn.Layer. Specific `VGG11` model depends on args. model = _load_pretrained_parameters(model, 'AlexNet')
"""
model = architectures.VGG11(**kwargs) return model
if pretrained:
model = _load_pretrained_parameters(model, 'VGG11') def vgg11(pretrained=False, **kwargs):
"""
return model VGG11
Args:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
def vgg13(pretrained=False, **kwargs): kwargs:
""" class_dim: int=1000. Output dim of last fc layer.
VGG13 stop_grad_layers: int=0. The parameters in blocks which index larger than `stop_grad_layers`, will be set `param.trainable=False`
Args: Returns:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. model: nn.Layer. Specific `VGG11` model depends on args.
kwargs: """
class_dim: int=1000. Output dim of last fc layer. model = architectures.VGG11(**kwargs)
stop_grad_layers: int=0. The parameters in blocks which index larger than `stop_grad_layers`, will be set `param.trainable=False` if pretrained:
Returns: model = _load_pretrained_parameters(model, 'VGG11')
model: nn.Layer. Specific `VGG13` model depends on args.
""" return model
model = architectures.VGG13(**kwargs)
if pretrained: def vgg13(pretrained=False, **kwargs):
model = _load_pretrained_parameters(model, 'VGG13') """
VGG13
return model Args:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
kwargs:
def vgg16(pretrained=False, **kwargs): class_dim: int=1000. Output dim of last fc layer.
""" stop_grad_layers: int=0. The parameters in blocks which index larger than `stop_grad_layers`, will be set `param.trainable=False`
VGG16 Returns:
Args: model: nn.Layer. Specific `VGG13` model depends on args.
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. """
kwargs: model = architectures.VGG13(**kwargs)
class_dim: int=1000. Output dim of last fc layer. if pretrained:
stop_grad_layers: int=0. The parameters in blocks which index larger than `stop_grad_layers`, will be set `param.trainable=False` model = _load_pretrained_parameters(model, 'VGG13')
Returns:
model: nn.Layer. Specific `VGG16` model depends on args. return model
"""
model = architectures.VGG16(**kwargs) def vgg16(pretrained=False, **kwargs):
if pretrained: """
model = _load_pretrained_parameters(model, 'VGG16') VGG16
Args:
return model pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
kwargs:
class_dim: int=1000. Output dim of last fc layer.
def vgg19(pretrained=False, **kwargs): stop_grad_layers: int=0. The parameters in blocks which index larger than `stop_grad_layers`, will be set `param.trainable=False`
""" Returns:
VGG19 model: nn.Layer. Specific `VGG16` model depends on args.
Args: """
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. model = architectures.VGG16(**kwargs)
kwargs: if pretrained:
class_dim: int=1000. Output dim of last fc layer. model = _load_pretrained_parameters(model, 'VGG16')
stop_grad_layers: int=0. The parameters in blocks which index larger than `stop_grad_layers`, will be set `param.trainable=False`
Returns: return model
model: nn.Layer. Specific `VGG19` model depends on args.
""" def vgg19(pretrained=False, **kwargs):
model = architectures.VGG19(**kwargs) """
if pretrained: VGG19
model = _load_pretrained_parameters(model, 'VGG19') Args:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
return model kwargs:
class_dim: int=1000. Output dim of last fc layer.
stop_grad_layers: int=0. The parameters in blocks which index larger than `stop_grad_layers`, will be set `param.trainable=False`
def resnet18(pretrained=False, **kwargs): Returns:
""" model: nn.Layer. Specific `VGG19` model depends on args.
ResNet18 """
Args: model = architectures.VGG19(**kwargs)
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. if pretrained:
kwargs: model = _load_pretrained_parameters(model, 'VGG19')
class_dim: int=1000. Output dim of last fc layer.
input_image_channel: int=3. The number of input image channels return model
data_format: str='NCHW'. The data format of batch input images, should in ('NCHW', 'NHWC')
Returns: def resnet18(pretrained=False, **kwargs):
model: nn.Layer. Specific `ResNet18` model depends on args. """
""" ResNet18
model = architectures.ResNet18(**kwargs) Args:
if pretrained: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
model = _load_pretrained_parameters(model, 'ResNet18') kwargs:
class_dim: int=1000. Output dim of last fc layer.
return model input_image_channel: int=3. The number of input image channels
data_format: str='NCHW'. The data format of batch input images, should in ('NCHW', 'NHWC')
Returns:
def resnet34(pretrained=False, **kwargs): model: nn.Layer. Specific `ResNet18` model depends on args.
""" """
ResNet34 model = architectures.ResNet18(**kwargs)
Args: if pretrained:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. model = _load_pretrained_parameters(model, 'ResNet18')
kwargs:
class_dim: int=1000. Output dim of last fc layer. return model
input_image_channel: int=3. The number of input image channels
data_format: str='NCHW'. The data format of batch input images, should in ('NCHW', 'NHWC') def resnet34(pretrained=False, **kwargs):
Returns: """
model: nn.Layer. Specific `ResNet34` model depends on args. ResNet34
""" Args:
model = architectures.ResNet34(**kwargs) pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
if pretrained: kwargs:
model = _load_pretrained_parameters(model, 'ResNet34') class_dim: int=1000. Output dim of last fc layer.
input_image_channel: int=3. The number of input image channels
return model data_format: str='NCHW'. The data format of batch input images, should in ('NCHW', 'NHWC')
Returns:
model: nn.Layer. Specific `ResNet34` model depends on args.
def resnet50(pretrained=False, **kwargs): """
""" model = architectures.ResNet34(**kwargs)
ResNet50 if pretrained:
Args: model = _load_pretrained_parameters(model, 'ResNet34')
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
kwargs: return model
class_dim: int=1000. Output dim of last fc layer.
input_image_channel: int=3. The number of input image channels def resnet50(pretrained=False, **kwargs):
data_format: str='NCHW'. The data format of batch input images, should in ('NCHW', 'NHWC') """
Returns: ResNet50
model: nn.Layer. Specific `ResNet50` model depends on args. Args:
""" pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
model = architectures.ResNet50(**kwargs) kwargs:
if pretrained: class_dim: int=1000. Output dim of last fc layer.
model = _load_pretrained_parameters(model, 'ResNet50') input_image_channel: int=3. The number of input image channels
data_format: str='NCHW'. The data format of batch input images, should in ('NCHW', 'NHWC')
return model Returns:
model: nn.Layer. Specific `ResNet50` model depends on args.
"""
def resnet101(pretrained=False, **kwargs): model = architectures.ResNet50(**kwargs)
""" if pretrained:
ResNet101 model = _load_pretrained_parameters(model, 'ResNet50')
Args:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. return model
kwargs:
class_dim: int=1000. Output dim of last fc layer. def resnet101(pretrained=False, **kwargs):
input_image_channel: int=3. The number of input image channels """
data_format: str='NCHW'. The data format of batch input images, should in ('NCHW', 'NHWC') ResNet101
Returns: Args:
model: nn.Layer. Specific `ResNet101` model depends on args. pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
""" kwargs:
model = architectures.ResNet101(**kwargs) class_dim: int=1000. Output dim of last fc layer.
if pretrained: input_image_channel: int=3. The number of input image channels
model = _load_pretrained_parameters(model, 'ResNet101') data_format: str='NCHW'. The data format of batch input images, should in ('NCHW', 'NHWC')
Returns:
return model model: nn.Layer. Specific `ResNet101` model depends on args.
"""
model = architectures.ResNet101(**kwargs)
def resnet152(pretrained=False, **kwargs): if pretrained:
""" model = _load_pretrained_parameters(model, 'ResNet101')
ResNet152
Args: return model
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
kwargs: def resnet152(pretrained=False, **kwargs):
class_dim: int=1000. Output dim of last fc layer. """
input_image_channel: int=3. The number of input image channels ResNet152
data_format: str='NCHW'. The data format of batch input images, should in ('NCHW', 'NHWC') Args:
Returns: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
model: nn.Layer. Specific `ResNet152` model depends on args. kwargs:
""" class_dim: int=1000. Output dim of last fc layer.
model = architectures.ResNet152(**kwargs) input_image_channel: int=3. The number of input image channels
if pretrained: data_format: str='NCHW'. The data format of batch input images, should in ('NCHW', 'NHWC')
model = _load_pretrained_parameters(model, 'ResNet152') Returns:
model: nn.Layer. Specific `ResNet152` model depends on args.
return model """
model = architectures.ResNet152(**kwargs)
if pretrained:
def squeezenet1_0(pretrained=False, **kwargs): model = _load_pretrained_parameters(model, 'ResNet152')
"""
SqueezeNet1_0 return model
Args:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. def squeezenet1_0(pretrained=False, **kwargs):
kwargs: """
class_dim: int=1000. Output dim of last fc layer. SqueezeNet1_0
Returns: Args:
model: nn.Layer. Specific `SqueezeNet1_0` model depends on args. pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
""" kwargs:
model = architectures.SqueezeNet1_0(**kwargs) class_dim: int=1000. Output dim of last fc layer.
if pretrained: Returns:
model = _load_pretrained_parameters(model, 'SqueezeNet1_0') model: nn.Layer. Specific `SqueezeNet1_0` model depends on args.
"""
return model model = architectures.SqueezeNet1_0(**kwargs)
if pretrained:
model = _load_pretrained_parameters(model, 'SqueezeNet1_0')
def squeezenet1_1(pretrained=False, **kwargs):
""" return model
SqueezeNet1_1
Args: def squeezenet1_1(pretrained=False, **kwargs):
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. """
kwargs: SqueezeNet1_1
class_dim: int=1000. Output dim of last fc layer. Args:
Returns: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
model: nn.Layer. Specific `SqueezeNet1_1` model depends on args. kwargs:
""" class_dim: int=1000. Output dim of last fc layer.
model = architectures.SqueezeNet1_1(**kwargs) Returns:
if pretrained: model: nn.Layer. Specific `SqueezeNet1_1` model depends on args.
model = _load_pretrained_parameters(model, 'SqueezeNet1_1') """
model = architectures.SqueezeNet1_1(**kwargs)
return model if pretrained:
model = _load_pretrained_parameters(model, 'SqueezeNet1_1')
def densenet121(pretrained=False, **kwargs): return model
"""
DenseNet121 def densenet121(pretrained=False, **kwargs):
Args: """
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. DenseNet121
kwargs: Args:
class_dim: int=1000. Output dim of last fc layer. pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
dropout: float=0. Probability of setting units to zero. kwargs:
bn_size: int=4. The number of channals per group class_dim: int=1000. Output dim of last fc layer.
Returns: dropout: float=0. Probability of setting units to zero.
model: nn.Layer. Specific `DenseNet121` model depends on args. bn_size: int=4. The number of channals per group
""" Returns:
model = architectures.DenseNet121(**kwargs) model: nn.Layer. Specific `DenseNet121` model depends on args.
if pretrained: """
model = _load_pretrained_parameters(model, 'DenseNet121') model = architectures.DenseNet121(**kwargs)
if pretrained:
return model model = _load_pretrained_parameters(model, 'DenseNet121')
return model
def densenet161(pretrained=False, **kwargs):
""" def densenet161(pretrained=False, **kwargs):
DenseNet161 """
Args: DenseNet161
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. Args:
kwargs: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
class_dim: int=1000. Output dim of last fc layer. kwargs:
dropout: float=0. Probability of setting units to zero. class_dim: int=1000. Output dim of last fc layer.
bn_size: int=4. The number of channals per group dropout: float=0. Probability of setting units to zero.
Returns: bn_size: int=4. The number of channals per group
model: nn.Layer. Specific `DenseNet161` model depends on args. Returns:
""" model: nn.Layer. Specific `DenseNet161` model depends on args.
model = architectures.DenseNet161(**kwargs) """
if pretrained: model = architectures.DenseNet161(**kwargs)
model = _load_pretrained_parameters(model, 'DenseNet161') if pretrained:
model = _load_pretrained_parameters(model, 'DenseNet161')
return model
return model
def densenet169(pretrained=False, **kwargs): def densenet169(pretrained=False, **kwargs):
""" """
DenseNet169 DenseNet169
Args: Args:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
kwargs: kwargs:
class_dim: int=1000. Output dim of last fc layer. class_dim: int=1000. Output dim of last fc layer.
dropout: float=0. Probability of setting units to zero. dropout: float=0. Probability of setting units to zero.
bn_size: int=4. The number of channals per group bn_size: int=4. The number of channals per group
Returns: Returns:
model: nn.Layer. Specific `DenseNet169` model depends on args. model: nn.Layer. Specific `DenseNet169` model depends on args.
""" """
model = architectures.DenseNet169(**kwargs) model = architectures.DenseNet169(**kwargs)
if pretrained: if pretrained:
model = _load_pretrained_parameters(model, 'DenseNet169') model = _load_pretrained_parameters(model, 'DenseNet169')
return model return model
def densenet201(pretrained=False, **kwargs):
def densenet201(pretrained=False, **kwargs): """
""" DenseNet201
DenseNet201 Args:
Args: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. kwargs:
kwargs: class_dim: int=1000. Output dim of last fc layer.
class_dim: int=1000. Output dim of last fc layer. dropout: float=0. Probability of setting units to zero.
dropout: float=0. Probability of setting units to zero. bn_size: int=4. The number of channals per group
bn_size: int=4. The number of channals per group Returns:
Returns: model: nn.Layer. Specific `DenseNet201` model depends on args.
model: nn.Layer. Specific `DenseNet201` model depends on args. """
""" model = architectures.DenseNet201(**kwargs)
model = architectures.DenseNet201(**kwargs) if pretrained:
if pretrained: model = _load_pretrained_parameters(model, 'DenseNet201')
model = _load_pretrained_parameters(model, 'DenseNet201')
return model
return model
def densenet264(pretrained=False, **kwargs):
"""
def densenet264(pretrained=False, **kwargs): DenseNet264
""" Args:
DenseNet264 pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
Args: kwargs:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. class_dim: int=1000. Output dim of last fc layer.
kwargs: dropout: float=0. Probability of setting units to zero.
class_dim: int=1000. Output dim of last fc layer. bn_size: int=4. The number of channals per group
dropout: float=0. Probability of setting units to zero. Returns:
bn_size: int=4. The number of channals per group model: nn.Layer. Specific `DenseNet264` model depends on args.
Returns: """
model: nn.Layer. Specific `DenseNet264` model depends on args. model = architectures.DenseNet264(**kwargs)
""" if pretrained:
model = architectures.DenseNet264(**kwargs) model = _load_pretrained_parameters(model, 'DenseNet264')
if pretrained:
model = _load_pretrained_parameters(model, 'DenseNet264') return model
return model def inceptionv3(pretrained=False, **kwargs):
"""
InceptionV3
def inceptionv3(pretrained=False, **kwargs): Args:
""" pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
InceptionV3 kwargs:
Args: class_dim: int=1000. Output dim of last fc layer.
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. Returns:
kwargs: model: nn.Layer. Specific `InceptionV3` model depends on args.
class_dim: int=1000. Output dim of last fc layer. """
Returns: model = architectures.InceptionV3(**kwargs)
model: nn.Layer. Specific `InceptionV3` model depends on args. if pretrained:
""" model = _load_pretrained_parameters(model, 'InceptionV3')
model = architectures.InceptionV3(**kwargs)
if pretrained: return model
model = _load_pretrained_parameters(model, 'InceptionV3')
def inceptionv4(pretrained=False, **kwargs):
return model """
InceptionV4
Args:
def inceptionv4(pretrained=False, **kwargs): pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
""" kwargs:
InceptionV4 class_dim: int=1000. Output dim of last fc layer.
Args: Returns:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. model: nn.Layer. Specific `InceptionV4` model depends on args.
kwargs: """
class_dim: int=1000. Output dim of last fc layer. model = architectures.InceptionV4(**kwargs)
Returns: if pretrained:
model: nn.Layer. Specific `InceptionV4` model depends on args. model = _load_pretrained_parameters(model, 'InceptionV4')
"""
model = architectures.InceptionV4(**kwargs) return model
if pretrained:
model = _load_pretrained_parameters(model, 'InceptionV4') def googlenet(pretrained=False, **kwargs):
"""
return model GoogLeNet
Args:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
def googlenet(pretrained=False, **kwargs): kwargs:
""" class_dim: int=1000. Output dim of last fc layer.
GoogLeNet Returns:
Args: model: nn.Layer. Specific `GoogLeNet` model depends on args.
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. """
kwargs: model = architectures.GoogLeNet(**kwargs)
class_dim: int=1000. Output dim of last fc layer. if pretrained:
Returns: model = _load_pretrained_parameters(model, 'GoogLeNet')
model: nn.Layer. Specific `GoogLeNet` model depends on args.
""" return model
model = architectures.GoogLeNet(**kwargs)
if pretrained: def shufflenetv2_x0_25(pretrained=False, **kwargs):
model = _load_pretrained_parameters(model, 'GoogLeNet') """
ShuffleNetV2_x0_25
return model Args:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
kwargs:
def shufflenetv2_x0_25(pretrained=False, **kwargs): class_dim: int=1000. Output dim of last fc layer.
""" Returns:
ShuffleNetV2_x0_25 model: nn.Layer. Specific `ShuffleNetV2_x0_25` model depends on args.
Args: """
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. model = architectures.ShuffleNetV2_x0_25(**kwargs)
kwargs: if pretrained:
class_dim: int=1000. Output dim of last fc layer. model = _load_pretrained_parameters(model, 'ShuffleNetV2_x0_25')
Returns:
model: nn.Layer. Specific `ShuffleNetV2_x0_25` model depends on args. return model
"""
model = architectures.ShuffleNetV2_x0_25(**kwargs) def mobilenetv1(pretrained=False, **kwargs):
if pretrained: """
model = _load_pretrained_parameters(model, 'ShuffleNetV2_x0_25') MobileNetV1
Args:
return model pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
kwargs:
class_dim: int=1000. Output dim of last fc layer.
def mobilenetv1(pretrained=False, **kwargs): Returns:
""" model: nn.Layer. Specific `MobileNetV1` model depends on args.
MobileNetV1 """
Args: model = architectures.MobileNetV1(**kwargs)
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. if pretrained:
kwargs: model = _load_pretrained_parameters(model, 'MobileNetV1')
class_dim: int=1000. Output dim of last fc layer.
Returns: return model
model: nn.Layer. Specific `MobileNetV1` model depends on args.
""" def mobilenetv1_x0_25(pretrained=False, **kwargs):
model = architectures.MobileNetV1(**kwargs) """
if pretrained: MobileNetV1_x0_25
model = _load_pretrained_parameters(model, 'MobileNetV1') Args:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
return model kwargs:
class_dim: int=1000. Output dim of last fc layer.
Returns:
def mobilenetv1_x0_25(pretrained=False, **kwargs): model: nn.Layer. Specific `MobileNetV1_x0_25` model depends on args.
""" """
MobileNetV1_x0_25 model = architectures.MobileNetV1_x0_25(**kwargs)
Args: if pretrained:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. model = _load_pretrained_parameters(model, 'MobileNetV1_x0_25')
kwargs:
class_dim: int=1000. Output dim of last fc layer. return model
Returns:
model: nn.Layer. Specific `MobileNetV1_x0_25` model depends on args. def mobilenetv1_x0_5(pretrained=False, **kwargs):
""" """
model = architectures.MobileNetV1_x0_25(**kwargs) MobileNetV1_x0_5
if pretrained: Args:
model = _load_pretrained_parameters(model, 'MobileNetV1_x0_25') pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
kwargs:
return model class_dim: int=1000. Output dim of last fc layer.
Returns:
model: nn.Layer. Specific `MobileNetV1_x0_5` model depends on args.
def mobilenetv1_x0_5(pretrained=False, **kwargs): """
""" model = architectures.MobileNetV1_x0_5(**kwargs)
MobileNetV1_x0_5 if pretrained:
Args: model = _load_pretrained_parameters(model, 'MobileNetV1_x0_5')
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
kwargs: return model
class_dim: int=1000. Output dim of last fc layer.
Returns: def mobilenetv1_x0_75(pretrained=False, **kwargs):
model: nn.Layer. Specific `MobileNetV1_x0_5` model depends on args. """
""" MobileNetV1_x0_75
model = architectures.MobileNetV1_x0_5(**kwargs) Args:
if pretrained: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
model = _load_pretrained_parameters(model, 'MobileNetV1_x0_5') kwargs:
class_dim: int=1000. Output dim of last fc layer.
return model Returns:
model: nn.Layer. Specific `MobileNetV1_x0_75` model depends on args.
"""
def mobilenetv1_x0_75(pretrained=False, **kwargs): model = architectures.MobileNetV1_x0_75(**kwargs)
""" if pretrained:
MobileNetV1_x0_75 model = _load_pretrained_parameters(model, 'MobileNetV1_x0_75')
Args:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. return model
kwargs:
class_dim: int=1000. Output dim of last fc layer. def mobilenetv2_x0_25(pretrained=False, **kwargs):
Returns: """
model: nn.Layer. Specific `MobileNetV1_x0_75` model depends on args. MobileNetV2_x0_25
""" Args:
model = architectures.MobileNetV1_x0_75(**kwargs) pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
if pretrained: kwargs:
model = _load_pretrained_parameters(model, 'MobileNetV1_x0_75') class_dim: int=1000. Output dim of last fc layer.
Returns:
return model model: nn.Layer. Specific `MobileNetV2_x0_25` model depends on args.
"""
model = architectures.MobileNetV2_x0_25(**kwargs)
def mobilenetv2_x0_25(pretrained=False, **kwargs): if pretrained:
""" model = _load_pretrained_parameters(model, 'MobileNetV2_x0_25')
MobileNetV2_x0_25
Args: return model
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
kwargs: def mobilenetv2_x0_5(pretrained=False, **kwargs):
class_dim: int=1000. Output dim of last fc layer. """
Returns: MobileNetV2_x0_5
model: nn.Layer. Specific `MobileNetV2_x0_25` model depends on args. Args:
""" pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
model = architectures.MobileNetV2_x0_25(**kwargs) kwargs:
if pretrained: class_dim: int=1000. Output dim of last fc layer.
model = _load_pretrained_parameters(model, 'MobileNetV2_x0_25') Returns:
model: nn.Layer. Specific `MobileNetV2_x0_5` model depends on args.
return model """
model = architectures.MobileNetV2_x0_5(**kwargs)
if pretrained:
def mobilenetv2_x0_5(pretrained=False, **kwargs): model = _load_pretrained_parameters(model, 'MobileNetV2_x0_5')
"""
MobileNetV2_x0_5 return model
Args:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. def mobilenetv2_x0_75(pretrained=False, **kwargs):
kwargs: """
class_dim: int=1000. Output dim of last fc layer. MobileNetV2_x0_75
Returns: Args:
model: nn.Layer. Specific `MobileNetV2_x0_5` model depends on args. pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
""" kwargs:
model = architectures.MobileNetV2_x0_5(**kwargs) class_dim: int=1000. Output dim of last fc layer.
if pretrained: Returns:
model = _load_pretrained_parameters(model, 'MobileNetV2_x0_5') model: nn.Layer. Specific `MobileNetV2_x0_75` model depends on args.
"""
return model model = architectures.MobileNetV2_x0_75(**kwargs)
if pretrained:
model = _load_pretrained_parameters(model, 'MobileNetV2_x0_75')
def mobilenetv2_x0_75(pretrained=False, **kwargs):
""" return model
MobileNetV2_x0_75
Args: def mobilenetv2_x1_5(pretrained=False, **kwargs):
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. """
kwargs: MobileNetV2_x1_5
class_dim: int=1000. Output dim of last fc layer. Args:
Returns: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
model: nn.Layer. Specific `MobileNetV2_x0_75` model depends on args. kwargs:
""" class_dim: int=1000. Output dim of last fc layer.
model = architectures.MobileNetV2_x0_75(**kwargs) Returns:
if pretrained: model: nn.Layer. Specific `MobileNetV2_x1_5` model depends on args.
model = _load_pretrained_parameters(model, 'MobileNetV2_x0_75') """
model = architectures.MobileNetV2_x1_5(**kwargs)
return model if pretrained:
model = _load_pretrained_parameters(model, 'MobileNetV2_x1_5')
def mobilenetv2_x1_5(pretrained=False, **kwargs): return model
"""
MobileNetV2_x1_5 def mobilenetv2_x2_0(pretrained=False, **kwargs):
Args: """
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. MobileNetV2_x2_0
kwargs: Args:
class_dim: int=1000. Output dim of last fc layer. pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
Returns: kwargs:
model: nn.Layer. Specific `MobileNetV2_x1_5` model depends on args. class_dim: int=1000. Output dim of last fc layer.
""" Returns:
model = architectures.MobileNetV2_x1_5(**kwargs) model: nn.Layer. Specific `MobileNetV2_x2_0` model depends on args.
if pretrained: """
model = _load_pretrained_parameters(model, 'MobileNetV2_x1_5') model = architectures.MobileNetV2_x2_0(**kwargs)
if pretrained:
return model model = _load_pretrained_parameters(model, 'MobileNetV2_x2_0')
return model
def mobilenetv2_x2_0(pretrained=False, **kwargs):
""" def mobilenetv3_large_x0_35(pretrained=False, **kwargs):
MobileNetV2_x2_0 """
Args: MobileNetV3_large_x0_35
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. Args:
kwargs: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
class_dim: int=1000. Output dim of last fc layer. kwargs:
Returns: class_dim: int=1000. Output dim of last fc layer.
model: nn.Layer. Specific `MobileNetV2_x2_0` model depends on args. Returns:
""" model: nn.Layer. Specific `MobileNetV3_large_x0_35` model depends on args.
model = architectures.MobileNetV2_x2_0(**kwargs) """
if pretrained: model = architectures.MobileNetV3_large_x0_35(**kwargs)
model = _load_pretrained_parameters(model, 'MobileNetV2_x2_0') if pretrained:
model = _load_pretrained_parameters(model,
return model 'MobileNetV3_large_x0_35')
return model
def mobilenetv3_large_x0_35(pretrained=False, **kwargs):
""" def mobilenetv3_large_x0_5(pretrained=False, **kwargs):
MobileNetV3_large_x0_35 """
Args: MobileNetV3_large_x0_5
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. Args:
kwargs: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
class_dim: int=1000. Output dim of last fc layer. kwargs:
Returns: class_dim: int=1000. Output dim of last fc layer.
model: nn.Layer. Specific `MobileNetV3_large_x0_35` model depends on args. Returns:
""" model: nn.Layer. Specific `MobileNetV3_large_x0_5` model depends on args.
model = architectures.MobileNetV3_large_x0_35(**kwargs) """
if pretrained: model = architectures.MobileNetV3_large_x0_5(**kwargs)
model = _load_pretrained_parameters(model, 'MobileNetV3_large_x0_35') if pretrained:
model = _load_pretrained_parameters(model,
return model 'MobileNetV3_large_x0_5')
return model
def mobilenetv3_large_x0_5(pretrained=False, **kwargs):
""" def mobilenetv3_large_x0_75(pretrained=False, **kwargs):
MobileNetV3_large_x0_5 """
Args: MobileNetV3_large_x0_75
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. Args:
kwargs: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
class_dim: int=1000. Output dim of last fc layer. kwargs:
Returns: class_dim: int=1000. Output dim of last fc layer.
model: nn.Layer. Specific `MobileNetV3_large_x0_5` model depends on args. Returns:
""" model: nn.Layer. Specific `MobileNetV3_large_x0_75` model depends on args.
model = architectures.MobileNetV3_large_x0_5(**kwargs) """
if pretrained: model = architectures.MobileNetV3_large_x0_75(**kwargs)
model = _load_pretrained_parameters(model, 'MobileNetV3_large_x0_5') if pretrained:
model = _load_pretrained_parameters(model,
return model 'MobileNetV3_large_x0_75')
return model
def mobilenetv3_large_x0_75(pretrained=False, **kwargs):
""" def mobilenetv3_large_x1_0(pretrained=False, **kwargs):
MobileNetV3_large_x0_75 """
Args: MobileNetV3_large_x1_0
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. Args:
kwargs: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
class_dim: int=1000. Output dim of last fc layer. kwargs:
Returns: class_dim: int=1000. Output dim of last fc layer.
model: nn.Layer. Specific `MobileNetV3_large_x0_75` model depends on args. Returns:
""" model: nn.Layer. Specific `MobileNetV3_large_x1_0` model depends on args.
model = architectures.MobileNetV3_large_x0_75(**kwargs) """
if pretrained: model = architectures.MobileNetV3_large_x1_0(**kwargs)
model = _load_pretrained_parameters(model, 'MobileNetV3_large_x0_75') if pretrained:
model = _load_pretrained_parameters(model,
return model 'MobileNetV3_large_x1_0')
return model
def mobilenetv3_large_x1_0(pretrained=False, **kwargs):
""" def mobilenetv3_large_x1_25(pretrained=False, **kwargs):
MobileNetV3_large_x1_0 """
Args: MobileNetV3_large_x1_25
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. Args:
kwargs: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
class_dim: int=1000. Output dim of last fc layer. kwargs:
Returns: class_dim: int=1000. Output dim of last fc layer.
model: nn.Layer. Specific `MobileNetV3_large_x1_0` model depends on args. Returns:
""" model: nn.Layer. Specific `MobileNetV3_large_x1_25` model depends on args.
model = architectures.MobileNetV3_large_x1_0(**kwargs) """
if pretrained: model = architectures.MobileNetV3_large_x1_25(**kwargs)
model = _load_pretrained_parameters(model, 'MobileNetV3_large_x1_0') if pretrained:
model = _load_pretrained_parameters(model,
return model 'MobileNetV3_large_x1_25')
return model
def mobilenetv3_large_x1_25(pretrained=False, **kwargs):
""" def mobilenetv3_small_x0_35(pretrained=False, **kwargs):
MobileNetV3_large_x1_25 """
Args: MobileNetV3_small_x0_35
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. Args:
kwargs: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
class_dim: int=1000. Output dim of last fc layer. kwargs:
Returns: class_dim: int=1000. Output dim of last fc layer.
model: nn.Layer. Specific `MobileNetV3_large_x1_25` model depends on args. Returns:
""" model: nn.Layer. Specific `MobileNetV3_small_x0_35` model depends on args.
model = architectures.MobileNetV3_large_x1_25(**kwargs) """
if pretrained: model = architectures.MobileNetV3_small_x0_35(**kwargs)
model = _load_pretrained_parameters(model, 'MobileNetV3_large_x1_25') if pretrained:
model = _load_pretrained_parameters(model,
return model 'MobileNetV3_small_x0_35')
return model
def mobilenetv3_small_x0_35(pretrained=False, **kwargs):
""" def mobilenetv3_small_x0_5(pretrained=False, **kwargs):
MobileNetV3_small_x0_35 """
Args: MobileNetV3_small_x0_5
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. Args:
kwargs: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
class_dim: int=1000. Output dim of last fc layer. kwargs:
Returns: class_dim: int=1000. Output dim of last fc layer.
model: nn.Layer. Specific `MobileNetV3_small_x0_35` model depends on args. Returns:
""" model: nn.Layer. Specific `MobileNetV3_small_x0_5` model depends on args.
model = architectures.MobileNetV3_small_x0_35(**kwargs) """
if pretrained: model = architectures.MobileNetV3_small_x0_5(**kwargs)
model = _load_pretrained_parameters(model, 'MobileNetV3_small_x0_35') if pretrained:
model = _load_pretrained_parameters(model,
return model 'MobileNetV3_small_x0_5')
return model
def mobilenetv3_small_x0_5(pretrained=False, **kwargs):
""" def mobilenetv3_small_x0_75(pretrained=False, **kwargs):
MobileNetV3_small_x0_5 """
Args: MobileNetV3_small_x0_75
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. Args:
kwargs: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
class_dim: int=1000. Output dim of last fc layer. kwargs:
Returns: class_dim: int=1000. Output dim of last fc layer.
model: nn.Layer. Specific `MobileNetV3_small_x0_5` model depends on args. Returns:
""" model: nn.Layer. Specific `MobileNetV3_small_x0_75` model depends on args.
model = architectures.MobileNetV3_small_x0_5(**kwargs) """
if pretrained: model = architectures.MobileNetV3_small_x0_75(**kwargs)
model = _load_pretrained_parameters(model, 'MobileNetV3_small_x0_5') if pretrained:
model = _load_pretrained_parameters(model,
return model 'MobileNetV3_small_x0_75')
return model
def mobilenetv3_small_x0_75(pretrained=False, **kwargs):
""" def mobilenetv3_small_x1_0(pretrained=False, **kwargs):
MobileNetV3_small_x0_75 """
Args: MobileNetV3_small_x1_0
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. Args:
kwargs: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
class_dim: int=1000. Output dim of last fc layer. kwargs:
Returns: class_dim: int=1000. Output dim of last fc layer.
model: nn.Layer. Specific `MobileNetV3_small_x0_75` model depends on args. Returns:
""" model: nn.Layer. Specific `MobileNetV3_small_x1_0` model depends on args.
model = architectures.MobileNetV3_small_x0_75(**kwargs) """
if pretrained: model = architectures.MobileNetV3_small_x1_0(**kwargs)
model = _load_pretrained_parameters(model, 'MobileNetV3_small_x0_75') if pretrained:
model = _load_pretrained_parameters(model,
return model 'MobileNetV3_small_x1_0')
return model
def mobilenetv3_small_x1_0(pretrained=False, **kwargs):
""" def mobilenetv3_small_x1_25(pretrained=False, **kwargs):
MobileNetV3_small_x1_0 """
Args: MobileNetV3_small_x1_25
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. Args:
kwargs: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
class_dim: int=1000. Output dim of last fc layer. kwargs:
Returns: class_dim: int=1000. Output dim of last fc layer.
model: nn.Layer. Specific `MobileNetV3_small_x1_0` model depends on args. Returns:
""" model: nn.Layer. Specific `MobileNetV3_small_x1_25` model depends on args.
model = architectures.MobileNetV3_small_x1_0(**kwargs) """
if pretrained: model = architectures.MobileNetV3_small_x1_25(**kwargs)
model = _load_pretrained_parameters(model, 'MobileNetV3_small_x1_0') if pretrained:
model = _load_pretrained_parameters(model,
return model 'MobileNetV3_small_x1_25')
return model
def mobilenetv3_small_x1_25(pretrained=False, **kwargs):
""" def resnext101_32x4d(pretrained=False, **kwargs):
MobileNetV3_small_x1_25 """
Args: ResNeXt101_32x4d
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. Args:
kwargs: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
class_dim: int=1000. Output dim of last fc layer. kwargs:
Returns: class_dim: int=1000. Output dim of last fc layer.
model: nn.Layer. Specific `MobileNetV3_small_x1_25` model depends on args. Returns:
""" model: nn.Layer. Specific `ResNeXt101_32x4d` model depends on args.
model = architectures.MobileNetV3_small_x1_25(**kwargs) """
if pretrained: model = architectures.ResNeXt101_32x4d(**kwargs)
model = _load_pretrained_parameters(model, 'MobileNetV3_small_x1_25') if pretrained:
model = _load_pretrained_parameters(model, 'ResNeXt101_32x4d')
return model
return model
def resnext101_32x4d(pretrained=False, **kwargs): def resnext101_64x4d(pretrained=False, **kwargs):
""" """
ResNeXt101_32x4d ResNeXt101_64x4d
Args: Args:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
kwargs: kwargs:
class_dim: int=1000. Output dim of last fc layer. class_dim: int=1000. Output dim of last fc layer.
Returns: Returns:
model: nn.Layer. Specific `ResNeXt101_32x4d` model depends on args. model: nn.Layer. Specific `ResNeXt101_64x4d` model depends on args.
""" """
model = architectures.ResNeXt101_32x4d(**kwargs) model = architectures.ResNeXt101_64x4d(**kwargs)
if pretrained: if pretrained:
model = _load_pretrained_parameters(model, 'ResNeXt101_32x4d') model = _load_pretrained_parameters(model, 'ResNeXt101_64x4d')
return model return model
def resnext152_32x4d(pretrained=False, **kwargs):
def resnext101_64x4d(pretrained=False, **kwargs): """
""" ResNeXt152_32x4d
ResNeXt101_64x4d Args:
Args: pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. kwargs:
kwargs: class_dim: int=1000. Output dim of last fc layer.
class_dim: int=1000. Output dim of last fc layer. Returns:
Returns: model: nn.Layer. Specific `ResNeXt152_32x4d` model depends on args.
model: nn.Layer. Specific `ResNeXt101_64x4d` model depends on args. """
""" model = architectures.ResNeXt152_32x4d(**kwargs)
model = architectures.ResNeXt101_64x4d(**kwargs) if pretrained:
if pretrained: model = _load_pretrained_parameters(model, 'ResNeXt152_32x4d')
model = _load_pretrained_parameters(model, 'ResNeXt101_64x4d')
return model
return model
def resnext152_64x4d(pretrained=False, **kwargs):
"""
def resnext152_32x4d(pretrained=False, **kwargs): ResNeXt152_64x4d
""" Args:
ResNeXt152_32x4d pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
Args: kwargs:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. class_dim: int=1000. Output dim of last fc layer.
kwargs: Returns:
class_dim: int=1000. Output dim of last fc layer. model: nn.Layer. Specific `ResNeXt152_64x4d` model depends on args.
Returns: """
model: nn.Layer. Specific `ResNeXt152_32x4d` model depends on args. model = architectures.ResNeXt152_64x4d(**kwargs)
""" if pretrained:
model = architectures.ResNeXt152_32x4d(**kwargs) model = _load_pretrained_parameters(model, 'ResNeXt152_64x4d')
if pretrained:
model = _load_pretrained_parameters(model, 'ResNeXt152_32x4d') return model
return model def resnext50_32x4d(pretrained=False, **kwargs):
"""
ResNeXt50_32x4d
def resnext152_64x4d(pretrained=False, **kwargs): Args:
""" pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
ResNeXt152_64x4d kwargs:
Args: class_dim: int=1000. Output dim of last fc layer.
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. Returns:
kwargs: model: nn.Layer. Specific `ResNeXt50_32x4d` model depends on args.
class_dim: int=1000. Output dim of last fc layer. """
Returns: model = architectures.ResNeXt50_32x4d(**kwargs)
model: nn.Layer. Specific `ResNeXt152_64x4d` model depends on args. if pretrained:
""" model = _load_pretrained_parameters(model, 'ResNeXt50_32x4d')
model = architectures.ResNeXt152_64x4d(**kwargs)
if pretrained: return model
model = _load_pretrained_parameters(model, 'ResNeXt152_64x4d')
def resnext50_64x4d(pretrained=False, **kwargs):
return model """
ResNeXt50_64x4d
Args:
def resnext50_32x4d(pretrained=False, **kwargs): pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
""" kwargs:
ResNeXt50_32x4d class_dim: int=1000. Output dim of last fc layer.
Args: Returns:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise. model: nn.Layer. Specific `ResNeXt50_64x4d` model depends on args.
kwargs: """
class_dim: int=1000. Output dim of last fc layer. model = architectures.ResNeXt50_64x4d(**kwargs)
Returns: if pretrained:
model: nn.Layer. Specific `ResNeXt50_32x4d` model depends on args. model = _load_pretrained_parameters(model, 'ResNeXt50_64x4d')
"""
model = architectures.ResNeXt50_32x4d(**kwargs) return model
if pretrained:
model = _load_pretrained_parameters(model, 'ResNeXt50_32x4d')
return model
def resnext50_64x4d(pretrained=False, **kwargs):
"""
ResNeXt50_64x4d
Args:
pretrained: bool=False. If `True` load pretrained parameters, `False` otherwise.
kwargs:
class_dim: int=1000. Output dim of last fc layer.
Returns:
model: nn.Layer. Specific `ResNeXt50_64x4d` model depends on args.
"""
model = architectures.ResNeXt50_64x4d(**kwargs)
if pretrained:
model = _load_pretrained_parameters(model, 'ResNeXt50_64x4d')
return model
Markdown is supported
0% 或 .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑！
想要评论请 注册 或