python - AttributeError : 'Tensor' object has no attribute 'lower' when applying Global MaxPooling instead of GlobalAveragePooling [duplicate]

Question

This question already has an answer here:

'Tensor' object has no attribute 'lower'

(1个答案)


在4个月前关闭。




这是我的进口:

import tensorflow as tf
import keras
from keras.models import Sequential, Model
from keras.layers import Conv2D, Flatten, MaxPooling2D, Dense, Input, Reshape, Concatenate, GlobalAveragePooling2D, BatchNormalization, Dropout, Activation, GlobalMaxPooling2D
from keras.utils import Sequence
import efficientnet.keras as efn

然后，我创建了以下模型:

efnB0_model = efn.EfficientNetB0(include_top=False, weights="imagenet", input_shape=(224, 224, 3))
efnB0_model.trainable = False

def create_model(input_shape = (224, 224, 3)):
    input_img = Input(shape=input_shape)
    model = efnB0_model (input_img)
    #model = GlobalAveragePooling2D(name='avg_pool')(model)
    model = GlobalMaxPooling2D(model)
    model = Dropout(0.2)(model)
    backbone = model

    branches = []
    for i in range(7):
            branches.append(backbone)
            branches[i] = Dense(360, name="branch_"+str(i)+"_Dense_16000")(branches[i])
            branches[i] = BatchNormalization()(branches[i])
            branches[i] = Activation("relu") (branches[i])
            branches[i] = Dropout(0.2)(branches[i])           
            branches[i] = Dense(35, activation = "softmax", name="branch_"+str(i)+"_output")(branches[i])
        
    output = Concatenate(axis=1)(branches)
    output = Reshape((7, 35))(output)
    model = Model(input_img, output)

    return model

当我使用 GlobaleAveragePooling2D(在代码中被注释掉)并运行 model = create_model()时，一切正常。当我改为使用 GlobalMaxPooling2D并运行 model = create_model()时，出现此错误:

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-83-834f03506210> in <module>()
----> 1 model = create_model()

3 frames
/usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py in normalize_data_format(value)
    309     if value is None:
    310         value = image_data_format()
--> 311     data_format = value.lower()
    312     if data_format not in {'channels_first', 'channels_last'}:
    313         raise ValueError('The `data_format` argument must be one of '

AttributeError: 'Tensor' object has no attribute 'lower

我应该如何修改我的代码？
非常感谢!

最佳答案

括号()丢失。 model = GlobalMaxPooling2D() (model)对我有用。
因此，这是解决方案:

def create_model(input_shape = (224, 224, 3)):
    input_img = Input(shape=input_shape)
    model = efnB0_model (input_img)
    #model = GlobalAveragePooling2D(name='avg_pool')(model)
    model = GlobalMaxPooling2D()(model)
    model = Dropout(0.2)(model)
    backbone = model

    branches = []
    for i in range(7):
            branches.append(backbone)
            branches[i] = Dense(360, name="branch_"+str(i)+"_Dense_16000")(branches[i])
            branches[i] = BatchNormalization()(branches[i])
            branches[i] = Activation("relu") (branches[i])
            branches[i] = Dropout(0.2)(branches[i])           
            branches[i] = Dense(35, activation = "softmax", name="branch_"+str(i)+"_output")(branches[i])
        
    output = Concatenate(axis=1)(branches)
    output = Reshape((7, 35))(output)
    model = Model(input_img, output)

    return model

关于python - AttributeError : 'Tensor' object has no attribute 'lower' when applying Global MaxPooling instead of GlobalAveragePooling [duplicate]，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/63117269/

上一篇：java - 斯坦福大学CoreNLP中文模型培训

下一篇：error-handling - Apache Camel 中doFinally block 的错误处理

Answer 1

括号()丢失。 model = GlobalMaxPooling2D() (model)对我有用。
因此，这是解决方案:

def create_model(input_shape = (224, 224, 3)):
    input_img = Input(shape=input_shape)
    model = efnB0_model (input_img)
    #model = GlobalAveragePooling2D(name='avg_pool')(model)
    model = GlobalMaxPooling2D()(model)
    model = Dropout(0.2)(model)
    backbone = model

    branches = []
    for i in range(7):
            branches.append(backbone)
            branches[i] = Dense(360, name="branch_"+str(i)+"_Dense_16000")(branches[i])
            branches[i] = BatchNormalization()(branches[i])
            branches[i] = Activation("relu") (branches[i])
            branches[i] = Dropout(0.2)(branches[i])           
            branches[i] = Dense(35, activation = "softmax", name="branch_"+str(i)+"_output")(branches[i])
        
    output = Concatenate(axis=1)(branches)
    output = Reshape((7, 35))(output)
    model = Model(input_img, output)

    return model