预训练模型下载

Keras Applications(tensorflow.keras.applications) 提供了预训练好的深度学习模型, 这些模型可以用于预测、特征提取等

当初始化一个模型时就会自动下载, 默认下载的路径是: ~/.keras/models/

图像分类模型

模型

在 ImageNet 数据上预训练过的用于图像分类的模型

• Xception
• VGG16
• VGG19
• ResNet, ResNetV2, ResNeXt
• InceptionV3
• InceptionResNet2
• MobileNet
• MobileNetV2
• DenseNet
• NASNet

API

from tensorflow.keras.applications.xception import Xception

# channels_last only; 299x299
xception_model = Xception(
    include_top = True,
    weights = "imagenet",
    input_tensor = None, 
    input_shape = None,
    pooling = None,
    classes = 1000,
)

from tensorflow.keras.applications.vgg16 import VGG16

# channels_first and channels_last; 224x224
vgg16_model = VGG16(
    include_top = True,
    weights = "imagenet",
    input_tensor = None, 
    input_shape = None,
    pooling = None,
    classes = 1000,
)

from tensorflow.keras.applications.vgg19 import VGG19

vgg19_model = VGG19(
    include_top = True, 
    weights = 'imagenet',
    input_tensor = None, 
    input_shape = None, 
    pooling = None, 
    classes = 1000,
)

from tensorflow.keras.applications.resnet50 import ResNet50

resnet50_model = ResNet50(
    include_top = True, 
    weights = 'imagenet', 
    input_tensor = None, 
    input_shape = None, 
    pooling = None, 
    classes = 1000,
)

from tensorflow.keras.applications.inception_v3 import InceptionV3

inception_v3_model = InceptionV3(
    include_top = True, 
    weights = 'imagenet', 
    input_tensor = None, 
    input_shape = None, 
    pooling = None, 
    classes = 1000,
)

from tensorflow.keras.applications.inception_resnet_v2 import InceptionResNetV2

inception_resnet_v2_model = InceptionResNetV2(
    include_top = True, 
    weights = 'imagenet', 
    input_tensor = None, 
    input_shape = None, 
    pooling = None, 
    classes = 1000,
)

from tensorflow.keras.applications.mobilenet import MobileNet

mobilenet_model = MobileNet(
    input_shape = None, 
    alpha = 1.0, 
    depth_multiplier = 1, 
    dropout = 1e-3, 
    include_top = True, 
    weights = 'imagenet', 
    input_tensor = None, 
    pooling = None, 
    classes = 1000,
)

from tensorflow.keras.applications.densenet import DenseNet121

densenet_model = DenseNet121(
    include_top = True, 
    weights = 'imagenet', 
    input_tensor = None, 
    input_shape = None, 
    pooling = None, 
    classes = 1000,
)

from tensorflow.keras.applications.densenet import DenseNet169

densenet_model = DenseNet169(
    include_top = True, 
    weights = 'imagenet', 
    input_tensor = None, 
    input_shape = None, 
    pooling = None, 
    classes = 1000,
)

from tensorflow.keras.applications.densenet import DenseNet201

densenet_model = DenseNet201(
    include_top = True, 
    weights = 'imagenet', 
    input_tensor = None, 
    input_shape = None, 
    pooling = None, 
    classes = 1000,
)

from tensorflow.keras.applications.nasnet import NASNetLarge

nasnet_model = NASNetLarge(
    input_shape = None, 
    include_top = True, 
    weights = 'imagenet', 
    input_tensor = None, 
    pooling = None, 
    classes = 1000,
)

from tensorflow.keras.applications.nasnet import NASNetMobile

nasnet_model = NASNetMobile(
    input_shape = None, 
    include_top = True, 
    weights = 'imagenet', 
    input_tensor = None, 
    pooling = None, 
    classes = 1000,
)

from tensorflow.keras.applications.mobilenet_v2 import MobileNetV2

mobilenet_v2_model = MobileNetV2(
    input_shape = None, 
    alpha = 1.0, 
    depth_multiplier = 1, 
    include_top = True, 
    weights = 'imagenet', 
    input_tensor = None, 
    pooling = None, 
    classes = 1000,
)

示例

图像分类模型使用示例

from keras.preprocessing import image
from keras.applications.resnet50 import ResNet50
from keras.applications.resnet50 import preprocess_input, decode_prediction
import numpy as np

# Load model
model = ResNet50(weights = "imagenet")

# Image data
img_path = "elephant.jpg"
img = image.load_img(img_path, target_size = (224, 224))
x = image.img_to_array(img)
x = np.expand_dims(x, axis = 0)
x = preprocess_input(x)

preds = model.predict(x)
print("Predicted:", decode_prediction(preds, top = 3)[0])