![](/img/trans.png)
[英]How to specify input_shape in Conv2D layer in Tensor Flow 2.0, keras
[英]Trouble figuring out how to define the input_shape in the Conv2D layer in Keras for my own dataset
定义输入形状时出现这些错误
ValueError: Error when checking input: expected conv2d_1_input to have 4 dimensions, but got array with shape (4000, 20, 20)
要么
ValueError: Input 0 is incompatible with layer conv2d_1: expected ndim=4, found ndim=5
我正在使用不同的Keras NN尝试对自己的数据集进行分类。
到目前为止,我的ANN取得了成功,但是CNN遇到了麻烦。
数据集由指定大小的矩阵组成,并填充有0,其中包含指定大小的子矩阵,并填充有1。 子矩阵是可选的,目标是训练NN以预测矩阵是否包含子矩阵。 为了使检测更加困难,我将各种噪声添加到矩阵中。
这是单个矩阵散乱的图片,黑色部分为0,白色部分为1。 图像的像素与矩阵中的条目之间存在1:1的对应关系。
我同时使用numpy savetxt和loadtxt将它们保存为文本。 然后看起来像这样:
#________________Array__Info:__(4000, 20, 20)__________
#________________Entry__Number__1________
0 0 1 1 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0
0 0 0 1 0 0 0 1 0 0 1 0 0 0 0 0 0 1 0 1
0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0
0 0 0 1 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0
0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 1 0 0 1
0 0 1 1 0 0 0 0 0 0 0 1 0 0 1 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 1 0 0 0 0
0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1
0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 1 1 1 0
0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0 0 0 0 1
0 0 0 1 0 0 0 1 0 1 0 0 0 0 0 1 0 0 0 0
0 0 0 0 0 0 0 0 0 1 1 0 0 1 0 0 0 1 1 1
0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 1
0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0
0 0 1 1 0 1 0 0 1 0 0 0 1 0 0 0 0 0 0 0
0 1 0 1 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 1
1 0 1 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0
0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0
#________________Entry__Number__2________
0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0
1 1 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 1
1 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 1 1 0 0
0 1 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0
0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0
0 0 1 0 0 0 1 0 0 1 0 0 0 0 0 1 0 0 0 1
0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0
1 0 1 0 0 1 0 1 0 1 0 0 0 0 1 1 1 0 0 1
0 0 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0
1 0 0 0 1 1 0 0 0 0 1 0 0 1 0 0 0 1 0 0
0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 1
0 0 0 0 0 1 1 0 0 0 0 1 0 1 0 0 0 0 0 0
0 0 1 1 0 0 0 0 0 0 0 1 1 1 1 1 0 1 0 0
0 0 0 0 0 0 0 1 1 0 1 1 1 1 1 1 0 0 0 1
0 1 0 0 0 0. . . . . . (and so on)
代码:(不包括进口商品)
# data
inputData = dsg.loadDataset("test_input.txt")
outputData = dsg.loadDataset("test_output.txt")
print("the size of the dataset is: ", inputData.shape, " of type: ", type(inputData))
# parameters
# CNN
cnn = Sequential()
cnn.add(Conv2D(32, (3, 3), input_shape = inputData.shape, activation = 'relu'))
cnn.add(MaxPooling2D(pool_size = (2, 2)))
cnn.add(Flatten())
cnn.add(Dense(units=64, activation='relu'))
cnn.add(Dense(units=1, activation='sigmoid'))
cnn.compile(optimizer = "adam", loss = 'binary_crossentropy', metrics = ['accuracy'])
cnn.summary()
cnn.fit(inputData,
outputData,
epochs=100,
validation_split=0.2)
我收到此输出错误消息
Using TensorFlow backend.
the size of the dataset is: (4000, 20, 20) of type: <class 'numpy.ndarray'>
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
conv2d_1 (Conv2D) (None, 3998, 18, 32) 5792
_________________________________________________________________
max_pooling2d_1 (MaxPooling2 (None, 1999, 9, 32) 0
_________________________________________________________________
flatten_1 (Flatten) (None, 575712) 0
_________________________________________________________________
dense_1 (Dense) (None, 64) 36845632
_________________________________________________________________
dense_2 (Dense) (None, 1) 65
=================================================================
Total params: 36,851,489
Trainable params: 36,851,489
Non-trainable params: 0
_________________________________________________________________
Traceback (most recent call last):
File "D:\GOOGLE DRIVE\School\sem-2-2018\BSP2\BiCS-BSP-2\CNN\matrixCNN.py", line 47, in <module>
validation_split=0.2)
File "C:\Code\Python\lib\site-packages\keras\models.py", line 963, in fit
validation_steps=validation_steps)
File "C:\Code\Python\lib\site-packages\keras\engine\training.py", line 1637, in fit
batch_size=batch_size)
File "C:\Code\Python\lib\site-packages\keras\engine\training.py", line 1483, in _standardize_user_data
exception_prefix='input')
File "C:\Code\Python\lib\site-packages\keras\engine\training.py", line 113, in _standardize_input_data
'with shape ' + str(data_shape))
ValueError: Error when checking input: expected conv2d_1_input to have 4 dimensions, but got array with shape (4000, 20, 20)
我真的不知道该如何解决。 我查看了Conv2D的文档,该文档说要以这样的形式放置:(批处理,高度,宽度,通道)。 在我的情况下(我认为):
input_shape=(4000, 20, 20, 1)
,因为我只有4000个20 * 20矩阵,且只有1和0
但是然后我收到此错误消息:
Using TensorFlow backend.
the size of the dataset is: (4000, 20, 20) of type: <class 'numpy.ndarray'>
Traceback (most recent call last):
File "D:\GOOGLE DRIVE\School\sem-2-2018\BSP2\BiCS-BSP-2\CNN\matrixCNN.py", line 30, in <module>
cnn.add(Conv2D(32, (3, 3), input_shape = (4000, 12, 12, 1), activation = 'relu'))
File "C:\Code\Python\lib\site-packages\keras\models.py", line 467, in add
layer(x)
File "C:\Code\Python\lib\site-packages\keras\engine\topology.py", line 573, in __call__
self.assert_input_compatibility(inputs)
File "C:\Code\Python\lib\site-packages\keras\engine\topology.py", line 472, in assert_input_compatibility
str(K.ndim(x)))
ValueError: Input 0 is incompatible with layer conv2d_1: expected ndim=4, found ndim=5
我应以哪种精确形状将数据传递到CNN?
所有文件都在这里可用谢谢您的宝贵时间。
您的CNN期望形状为(num_samples, 20, 20, 1)
,而数据的格式为(num_samples, 20, 20)
。
由于您只有1个通道,因此您可以将数据重塑为(4000, 20, 20, 1)
inputData = inputData.reshape(-1, 20, 20, 1)
如果要在模型内部进行重塑,只需添加一个Reshape
图层即可。 作为第一层:
model.add(Reshape(input_shape = (20, 20), target_shape=(20, 20, 1)))
多亏了Primusa的帮助和我发现的另一个线程 ,我才开始工作。 这是我添加的内容:
inputData = inputData.reshape(4000, 20, 20, 1)
outputData = outputData.reshape(4000, 1)
与conv2D层是
cnn.add(Conv2D(32, (3, 3), input_shape = (20, 20, 1), activation = 'relu'))
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.