自学教程：Python layers.GlobalAveragePooling2D方法代码示例

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def get_model(session):    # create the base pre-trained model    base_model = Xception(weights=None, include_top=False, input_shape=(270, 480, 3))    # add a global spatial average pooling layer    x = base_model.output    x = GlobalAveragePooling2D()(x)    # add a fully-connected layer    x = Dense(1024, activation='relu')(x)    # putput layer    predictions = Dense(session.training_dataset_info['number_of_labels'], activation='softmax')(x)    # model    model = Model(inputs=base_model.input, outputs=predictions)    learning_rate = 0.001    opt = keras.optimizers.adam(lr=learning_rate, decay=1e-5)    model.compile(loss='categorical_crossentropy',                  optimizer=opt,                  metrics=['accuracy'])    return model 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def resnet_pseudo(self,dim=224,freeze_layers=10,full_freeze='N'):		model = ResNet50(weights='imagenet',include_top=False)		x = model.output		x = GlobalAveragePooling2D()(x)		x = Dense(512, activation='relu')(x)		x = Dropout(0.5)(x)		x = Dense(512, activation='relu')(x)		x = Dropout(0.5)(x)		out = Dense(5,activation='softmax')(x)		model_final = Model(input = model.input,outputs=out)		if full_freeze != 'N':			for layer in model.layers[0:freeze_layers]:				layer.trainable = False		return model_final	# VGG16 Model for transfer Learning 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def resnet_pseudo(self,dim=224,freeze_layers=10,full_freeze='N'):		model = ResNet50(weights='imagenet',include_top=False)		x = model.output		x = GlobalAveragePooling2D()(x)		x = Dense(512, activation='relu')(x)		x = Dropout(0.5)(x)		x = Dense(512, activation='relu')(x)		x = Dropout(0.5)(x)		out = Dense(1)(x)		model_final = Model(input = model.input,outputs=out)		if full_freeze != 'N':			for layer in model.layers[0:freeze_layers]:				layer.trainable = False		return model_final	# VGG16 Model for transfer Learning 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def resnet_pseudo(self,dim=224,freeze_layers=10,full_freeze='N'):        model = ResNet50(weights='imagenet',include_top=False)        x = model.output        x = GlobalAveragePooling2D()(x)        x = Dense(512, activation='relu')(x)        x = Dropout(0.5)(x)        x = Dense(512, activation='relu')(x)        x = Dropout(0.5)(x)        out = Dense(5,activation='softmax')(x)        model_final = Model(input = model.input,outputs=out)        if full_freeze != 'N':            for layer in model.layers[0:freeze_layers]:                layer.trainable = False        return model_final# VGG16 Model for transfer Learning 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def classifier_layers(x, input_shape, trainable=False):    # compile times on theano tend to be very high, so we use smaller ROI pooling regions to workaround    # (hence a smaller stride in the region that follows the ROI pool)    x = TimeDistributed(SeparableConv2D(1536, (3, 3),                                        padding='same',                                        use_bias=False),                        name='block14_sepconv1')(x)    x = TimeDistributed(BatchNormalization(), name='block14_sepconv1_bn')(x)    x = Activation('relu', name='block14_sepconv1_act')(x)    x = TimeDistributed(SeparableConv2D(2048, (3, 3),                                        padding='same',                                        use_bias=False),                        name='block14_sepconv2')(x)    x = TimeDistributed(BatchNormalization(), name='block14_sepconv2_bn')(x)    x = Activation('relu', name='block14_sepconv2_act')(x)    TimeDistributed(GlobalAveragePooling2D(), name='avg_pool')(x)    return x 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def _squeeze(self, inputs):        """Squeeze and Excitation.        This function defines a squeeze structure.        # Arguments            inputs: Tensor, input tensor of conv layer.        """        input_channels = int(inputs.shape[-1])        x = GlobalAveragePooling2D()(inputs)        x = Dense(input_channels, activation='relu')(x)        x = Dense(input_channels, activation='hard_sigmoid')(x)        x = Reshape((1, 1, input_channels))(x)        x = Multiply()([inputs, x])        return x 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def global_pool2d():    def compile_fn(di, dh):        layer = layers.GlobalAveragePooling2D()        def fn(di):            return {'out': layer(di['in'])}        return fn    return siso_keras_module('GlobalAveragePool', compile_fn, {}) 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def learn():    (train_x, train_y, sample_weight), (test_x, test_y) = load_data()    datagen = ImageDataGenerator(horizontal_flip=True,                                 vertical_flip=True)    train_generator = datagen.flow(train_x, train_y, sample_weight=sample_weight)    base = VGG16(weights='imagenet', include_top=False, input_shape=(None, None, 3))    for layer in base.layers[:-4]:        layer.trainable = False    model = models.Sequential([        base,        layers.BatchNormalization(),        layers.Conv2D(64, (3, 3), activation='relu', padding='same'),        layers.GlobalAveragePooling2D(),        layers.BatchNormalization(),        layers.Dense(64, activation='relu'),        layers.BatchNormalization(),        layers.Dropout(0.20),        layers.Dense(80, activation='softmax')    ])    model.compile(optimizer=optimizers.RMSprop(lr=1e-5),                  loss='sparse_categorical_crossentropy',                  metrics=['accuracy'])    model.summary()    reduce_lr = ReduceLROnPlateau(verbose=1)    model.fit_generator(train_generator, epochs=400,                        steps_per_epoch=100,                        validation_data=(test_x[:800], test_y[:800]),                        callbacks=[reduce_lr])    result = model.evaluate(test_x, test_y)    print(result)    model.save('12306.image.model.h5', include_optimizer=False) 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def _squeeze_excite_block(input, filters, k=1, name=None):    init = input    se_shape = (1, 1, filters * k) if K.image_data_format() == 'channels_last' else (filters * k, 1, 1)    se = GlobalAveragePooling2D()(init)    se = Reshape(se_shape)(se)    se = Dense((filters * k) // 16, activation='relu', kernel_initializer='he_normal', use_bias=False,name=name+'_fc1')(se)    se = Dense(filters * k, activation='sigmoid', kernel_initializer='he_normal', use_bias=False,name=name+'_fc2')(se)    return se# pyramid pooling function 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def squeeze_excite_block(input, ratio=16):    ''' Create a channel-wise squeeze-excite block    Args:        input: input tensor        filters: number of output filters    Returns: a keras tensor    References    -   [Squeeze and Excitation Networks](https://arxiv.org/abs/1709.01507)    '''    init = input    channel_axis = 1 if K.image_data_format() == "channels_first" else -1    filters = init._keras_shape[channel_axis]    se_shape = (1, 1, filters)    se = GlobalAveragePooling2D()(init)    se = Reshape(se_shape)(se)    se = Dense(filters // ratio, activation='relu', kernel_initializer='he_normal', use_bias=False)(se)    se = Dense(filters, activation='sigmoid', kernel_initializer='he_normal', use_bias=False)(se)    if K.image_data_format() == 'channels_first':        se = Permute((3, 1, 2))(se)    x = multiply([init, se])    return x 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def squeeze_excite_block(input_tensor, ratio=16):    """ Create a channel-wise squeeze-excite block    Args:        input_tensor: input Keras tensor        ratio: number of output filters    Returns: a Keras tensor    References    -   [Squeeze and Excitation Networks](https://arxiv.org/abs/1709.01507)    """    init = input_tensor    channel_axis = 1 if K.image_data_format() == "channels_first" else -1    filters = _tensor_shape(init)[channel_axis]    se_shape = (1, 1, filters)    se = GlobalAveragePooling2D()(init)    se = Reshape(se_shape)(se)    se = Dense(filters // ratio, activation='relu', kernel_initializer='he_normal', use_bias=False)(se)    se = Dense(filters, activation='sigmoid', kernel_initializer='he_normal', use_bias=False)(se)    if K.image_data_format() == 'channels_first':        se = Permute((3, 1, 2))(se)    x = multiply([init, se])    return x 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def model_fn(actions):    # unpack the actions from the list    kernel_1, filters_1, kernel_2, filters_2, kernel_3, filters_3, kernel_4, filters_4 = actions    ip = Input(shape=(32, 32, 3))    x = Conv2D(filters_1, (kernel_1, kernel_1), strides=(2, 2), padding='same', activation='relu')(ip)    x = Conv2D(filters_2, (kernel_2, kernel_2), strides=(1, 1), padding='same', activation='relu')(x)    x = Conv2D(filters_3, (kernel_3, kernel_3), strides=(2, 2), padding='same', activation='relu')(x)    x = Conv2D(filters_4, (kernel_4, kernel_4), strides=(1, 1), padding='same', activation='relu')(x)    x = GlobalAveragePooling2D()(x)    x = Dense(10, activation='softmax')(x)    model = Model(ip, x)    return model 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def VGG16_pseudo(self,dim=224,freeze_layers=10,full_freeze='N'):		model = VGG16(weights='imagenet',include_top=False)		x = model.output		x = GlobalAveragePooling2D()(x)		x = Dense(512, activation='relu')(x)		x = Dropout(0.5)(x)		x = Dense(512, activation='relu')(x)		x = Dropout(0.5)(x)		out = Dense(5,activation='softmax')(x)		model_final = Model(input = model.input,outputs=out)		if full_freeze != 'N':			for layer in model.layers[0:freeze_layers]:				layer.trainable = False		return model_final 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def VGG16_pseudo(self,dim=224,freeze_layers=10,full_freeze='N'):        model = VGG16(weights='imagenet',include_top=False)        x = model.output        x = GlobalAveragePooling2D()(x)        x = Dense(512, activation='relu')(x)        x = Dropout(0.5)(x)        x = Dense(512, activation='relu')(x)        x = Dropout(0.5)(x)        out = Dense(5,activation='softmax')(x)        model_final = Model(input = model.input,outputs=out)        if full_freeze != 'N':            for layer in model.layers[0:freeze_layers]:                layer.trainable = False        return model_final 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def compile_sesemi(network, input_shape, nb_classes,                   lrate, in_network_dropout, super_dropout):    weight_decay = 0.0005    initer = initializers.glorot_uniform()    fc_params = dict(            use_bias=True,            activation='softmax',            kernel_initializer=initer,            kernel_regularizer=l2(weight_decay),        )    cnn_trunk = network.create_network(input_shape, in_network_dropout)        super_in = Input(shape=input_shape, name='super_data')    self_in  = Input(shape=input_shape, name='self_data')    super_out = cnn_trunk(super_in)    self_out  = cnn_trunk(self_in)        super_out = GlobalAveragePooling2D(name='super_gap')(super_out)    self_out  = GlobalAveragePooling2D(name='self_gap')(self_out)    if super_dropout > 0.0:        super_out = Dropout(super_dropout, name='super_dropout')(super_out)        super_out = Dense(nb_classes, name='super_clf', **fc_params)(super_out)    self_out  = Dense(proxy_labels, name='self_clf', **fc_params)(self_out)    sesemi_model = Model(inputs=[self_in, super_in],                         outputs=[self_out, super_out])    inference_model = Model(inputs=[super_in], outputs=[super_out])    sgd = optimizers.SGD(lr=lrate, momentum=0.9, nesterov=True)    sesemi_model.compile(optimizer=sgd,                         loss={'super_clf': 'categorical_crossentropy',                               'self_clf' : 'categorical_crossentropy'},                         loss_weights={'super_clf': 1.0, 'self_clf': 1.0},                         metrics=None)    return sesemi_model, inference_model 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def squared_root_normalization(x):    """    Squared root normalization for convolution layers` output    first apply global average pooling followed by squared root on all elements    then l2 normalize the vector    :param x: input tensor, output of convolution layer    :return:    """    x = GlobalAveragePooling2D()(x)    #output shape = (None, nc)   # x = K.sqrt(x)    #x = K.l2_normalize(x, axis=0)    return x 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def model1(weights_path=None):    '''    Basic ResNet-FT for baseline comparisions.    Creates a model by for all aesthetic attributes along    with overall aesthetic score, by finetuning resnet50    :param weights_path: path of the weight file    :return: Keras model instance    '''    _input = Input(shape=(299, 299, 3))    resnet = ResNet50(include_top=False, weights='imagenet', input_tensor=_input)    last_layer_output = GlobalAveragePooling2D()(resnet.get_layer('activation_49').output)    # output of model    outputs = []    attrs = ['BalacingElements', 'ColorHarmony', 'Content', 'DoF',             'Light', 'MotionBlur', 'Object', 'RuleOfThirds', 'VividColor']    for attribute in attrs:        outputs.append(Dense(1, init='glorot_uniform', activation='tanh', name=attribute)(last_layer_output))    non_negative_attrs = ['Repetition', 'Symmetry', 'score']    for attribute in non_negative_attrs:        outputs.append(Dense(1, init='glorot_uniform', activation='sigmoid', name=attribute)(last_layer_output))    model = Model(input=_input, output=outputs)    if weights_path:        model.load_weights(weights_path)    return model 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def classifier_layers(x, input_shape, trainable=False):    # compile times on theano tend to be very high, so we use smaller ROI pooling regions to workaround    # (hence a smaller stride in the region that follows the ROI pool)    channel_axis = 1 if K.image_data_format() == 'channels_first' else 4    # Mixed 7a (Reduction-B block): 8 x 8 x 2080    branch_0 = conv2d_bn_td(x, 256, 1, name='Reduction_B_block' + '_conv1')    branch_0 = conv2d_bn_td(branch_0, 384, 3, strides=2, padding='valid', name='Reduction_B_block' + '_conv2')    branch_1 = conv2d_bn_td(x, 256, 1, name='Reduction_B_block' + '_conv3')    branch_1 = conv2d_bn_td(branch_1, 288, 3, strides=2, padding='valid', name='Reduction_B_block' + '_conv4')    branch_2 = conv2d_bn_td(x, 256, 1, name='Reduction_B_block' + '_conv5')    branch_2 = conv2d_bn_td(branch_2, 288, 3, name='Reduction_B_block' + '_conv6')    branch_2 = conv2d_bn_td(branch_2, 320, 3, strides=2, padding='valid', name='Reduction_B_block' + '_conv7')    branch_pool = TimeDistributed(MaxPooling2D(3, strides=2, padding='valid'))(x)    branches = [branch_0, branch_1, branch_2, branch_pool]    x = Concatenate(axis=channel_axis, name='mixed_7a')(branches)    # 10x block8 (Inception-ResNet-C block): 8 x 8 x 2080    for block_idx in range(1, 10):        x = inception_resnet_block_td(x,                                      scale=0.2,                                      block_type='block8',                                      block_idx=block_idx)    x = inception_resnet_block_td(x,                                  scale=1.,                                  activation=None,                                  block_type='block8',                                  block_idx=10)    # Final convolution block: 8 x 8 x 1536    x = conv2d_bn_td(x, 1536, 1, name='conv_7b')    TimeDistributed(GlobalAveragePooling2D(), name='avg_pool')(x)    return x 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def build_model(base_model, num_classes):    x = base_model.output    x = GlobalAveragePooling2D(name='Avg_pool_1')(x)    x = Dense(1024, activation='relu', name='dense_one')(x)    x = Dense(1024, activation='relu', name='dense_two')(x)    x = Dense(512, activation='relu', name='dense_three')(x)    x = Dense(num_classes, activation='softmax', name='main_output')(x)    return x 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def get_model_tocompiled(base_model, num_class):    x = base_model.output    x = GlobalAveragePooling2D()(x)    x = Dense(256, activation='relu')(x)    x = Dropout(0.5)(x)    predictions = Dense(num_class, activation='softmax', name='predictions')(x)    clf = Model(base_model.input, predictions)    return clf 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def se_block(input_feature, ratio=8):	"""Contains the implementation of Squeeze-and-Excitation(SE) block.	As described in https://arxiv.org/abs/1709.01507.	"""		channel_axis = 1 if K.image_data_format() == "channels_first" else -1	channel = input_feature._keras_shape[channel_axis]	se_feature = GlobalAveragePooling2D()(input_feature)	se_feature = Reshape((1, 1, channel))(se_feature)	assert se_feature._keras_shape[1:] == (1,1,channel)	se_feature = Dense(channel // ratio,					   activation='relu',					   kernel_initializer='he_normal',					   use_bias=True,					   bias_initializer='zeros')(se_feature)	assert se_feature._keras_shape[1:] == (1,1,channel//ratio)	se_feature = Dense(channel,					   activation='sigmoid',					   kernel_initializer='he_normal',					   use_bias=True,					   bias_initializer='zeros')(se_feature)	assert se_feature._keras_shape[1:] == (1,1,channel)	if K.image_data_format() == 'channels_first':		se_feature = Permute((3, 1, 2))(se_feature)	se_feature = multiply([input_feature, se_feature])	return se_feature 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def channel_attention(input_feature, ratio=8):		channel_axis = 1 if K.image_data_format() == "channels_first" else -1	channel = input_feature._keras_shape[channel_axis]		shared_layer_one = Dense(channel//ratio,							 activation='relu',							 kernel_initializer='he_normal',							 use_bias=True,							 bias_initializer='zeros')	shared_layer_two = Dense(channel,							 kernel_initializer='he_normal',							 use_bias=True,							 bias_initializer='zeros')		avg_pool = GlobalAveragePooling2D()(input_feature)    	avg_pool = Reshape((1,1,channel))(avg_pool)	assert avg_pool._keras_shape[1:] == (1,1,channel)	avg_pool = shared_layer_one(avg_pool)	assert avg_pool._keras_shape[1:] == (1,1,channel//ratio)	avg_pool = shared_layer_two(avg_pool)	assert avg_pool._keras_shape[1:] == (1,1,channel)		max_pool = GlobalMaxPooling2D()(input_feature)	max_pool = Reshape((1,1,channel))(max_pool)	assert max_pool._keras_shape[1:] == (1,1,channel)	max_pool = shared_layer_one(max_pool)	assert max_pool._keras_shape[1:] == (1,1,channel//ratio)	max_pool = shared_layer_two(max_pool)	assert max_pool._keras_shape[1:] == (1,1,channel)		cbam_feature = Add()([avg_pool,max_pool])	cbam_feature = Activation('sigmoid')(cbam_feature)		if K.image_data_format() == "channels_first":		cbam_feature = Permute((3, 1, 2))(cbam_feature)		return multiply([input_feature, cbam_feature]) 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def get_model(base_model,               layer,               lr=1e-3,               input_shape=(224,224,1),               classes=2,              activation="softmax",              dropout=None,               pooling="avg",               weights=None,              pretrained="imagenet"):     base = base_model(input_shape=input_shape,                      include_top=False,                      weights=pretrained,                       channels="gray")     if pooling == "avg":         x = GlobalAveragePooling2D()(base.output)     elif pooling == "max":         x = GlobalMaxPooling2D()(base.output)     elif pooling is None:         x = Flatten()(base.output)     if dropout is not None:         x = Dropout(dropout)(x)     x = Dense(classes, activation=activation)(x)     model = Model(inputs=base.input, outputs=x)     if weights is not None:         model.load_weights(weights)     for l in model.layers[:layer]:        l.trainable = False     model.compile(loss="binary_crossentropy", metrics=["accuracy"],                   optimizer=optimizers.Adam(lr))     return model############ DATA ############# == PREPROCESSING == # 

# 需要导入模块: from keras import layers [as 别名]# 或者: from keras.layers import GlobalAveragePooling2D [as 别名]def get_model(base_model,               layer,               lr=1e-3,               input_shape=(224,224,1),               classes=2,              activation="softmax",              dropout=None,               pooling="avg",               weights=None,              pretrained=None):     base = base_model(input_shape=input_shape,                      include_top=False,                      weights=pretrained,                       channels="gray")     if pooling == "avg":         x = GlobalAveragePooling2D()(base.output)     elif pooling == "max":         x = GlobalMaxPooling2D()(base.output)     elif pooling is None:         x = Flatten()(base.output)     if dropout is not None:         x = Dropout(dropout)(x)     x = Dense(classes, activation=activation)(x)     model = Model(inputs=base.input, outputs=x)     if weights is not None:         model.load_weights(weights)     for l in model.layers[:layer]:        l.trainable = False     model.compile(loss="binary_crossentropy", metrics=["accuracy"],                   optimizer=optimizers.Adam(lr))     return model############ DATA ############# == PREPROCESSING == #