TypeError:__init __()缺少1个必需的位置参数:"units" [英] TypeError: __init__() missing 1 required positional argument: "units"'

问题描述

插入的keras库和软件包在下面，

import keras
from keras.models import Sequential
from keras.layers.convolutional import Conv2D
from keras.layers import Dense, Flatten, MaxPooling2D
from keras.utils import np_utils

初始化CNN模型

classifier = Sequential()

#1第一步:-卷积

classifier.add(Conv2D(32, 3, 3, input_shape=(64,64,3), activation='relu'))

第2步:-合并

classifier.add(MaxPooling2D(pool_size = (2,2) ))

#3第三步:-展平

classifier.add(Flatten())

#完全连接

classifier.add(Dense(output_dim = 128, activation='relu'))   #Here show the error
classifier.add(Dense(output_dim = 1, activation='sigmoid'))

#编译CNN

classifier.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])

错误下方

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-9-3df4f496ff43> in <module>()
      1 #full connection
----> 2 classifier.add(Dense(output_dim = 128, activation='relu'))
      3 classifier.add(Dense(output_dim = 1, activation='sigmoid'))
      4 
      5 #compiling the CNN

TypeError: __init__() missing 1 required positional argument: 'units'

注意:需要帮助，请帮助我.

Note: Need help, so please help me.

推荐答案

您传递了错误的参数，因为错误提示您可能想编写:

You are passing the wrong parameters, as the error states you probably wanted to write:

classifier.add(Dense(units = 128, activation='relu'))   
classifier.add(Dense(units = 1, activation='sigmoid'))

这篇关于TypeError:__init __()缺少1个必需的位置参数:"units"的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！