首先，大家好；
我只想过滤二进制图像数据集（两种不同类型的图像具有两种不同类型的过滤矩阵），过滤后我只想添加两层回到正常的cnn
>>> import builtins
>>> builtins is __builtins__
True
我尽力而为，但是出现了这个错误：
import keras
import keras.backend as K
import numpy as np
from keras import Input, layers
from keras.models import Model
def my_filter1(shape, dtype=None):
kernelValues_cat = np.array([-0.03861977, -0.00681433, 0.19763936, 0.16233025, 0.01814837,
-0.10037766, 0.025743 , 0.07676292, 0.10076649, -0.01162791,
0.277762 , -0.03399856, 0.06740112, 0.08228938, -0.22894219,
-0.01097343, 0.10896538, -0.33455664, 0.01249053, -0.07131144,
-0.30083737, 0.0450361 , 0.02632423, 0.0326063 , 0.07632089,
0.08170559, 0.17779405, 0.14002198, -0.0534982 , 0.03667643,
0.00984352, -0.22772053, -0.11768475, -0.1448647 , 0.05544433,
-0.02038156, 0.07040382, 0.03736563, -0.04353458, -0.18037376,
-0.11258822, 0.24918582, -0.17927748, 0.04648639, 0.09226271,
0.14498603, 0.0091729 , -0.01569342, 0.12471238, -0.17353317])
return K.variable(kernelValues_cat , dtype='float32')
def my_filter2(shape, dtype=None):
kernelValues_dog = np.array([ 0.00562452, 0.07233931, 0.01056713, -0.00271974, 0.01258311,
0.01658106, -0.07546224, 0.00367415, 0.00498003, -0.06350716,
0.00421606, -0.02176326, 0.0160836 , 0.01281294, 0.05058465,
-0.03861704, 0.0773253 , -0.03759991, -0.06474894, -0.056652 ,
-0.06309445, 0.03082605, -0.06260598, 0.00689435, -0.04214211,
0.10993017, 0.03741956, 0.0137707 , -0.02764472, -0.05047289,
0.07559353, -0.0010061 , -0.06301045, -0.00553294, -0.01028815,
0.01860766, -0.02061985, -0.02880365, 0.0324572 , 0.03683988,
-0.04732571, -0.02691277, 0.01911413, 0.03938391, -0.02919731,
0.03790436, 0.06117813, -0.06713407, 0.07434538, 0.04775169])
return K.variable(kernelValues_dog , dtype='float32')
img_input1 = layers.Input(shape=(150, 150, 3))
x1 = layers.Conv2D(filters=1,
kernel_size = 3,
kernel_initializer=my_filter1,
strides=2,
padding='valid') (img_input1)
img_input2 = layers.Input(shape=(150, 150, 3))
x2 = layers.Conv2D(filters=1,
kernel_size = 3,
kernel_initializer=my_filter2,
strides=2,
padding='valid') (img_input2)
added = keras.layers.Add()([x1, x2])
x = layers.Conv2D(16, 3, activation='relu')(added)
x = layers.MaxPooling2D(2)(x)
x = layers.Conv2D(32, 3, activation='relu')(x)
x = layers.MaxPooling2D(2)(x)
x = layers.Conv2D(64, 3, activation='relu')(x)
x = layers.MaxPooling2D(2)(x)
在那之后，只想看看我的模型，对我来说任何帮助都将非常感谢
InvalidArgumentError Traceback (most recent call last)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/ops.py in _create_c_op(graph, node_def, inputs, control_inputs, op_def)
1653 try:
-> 1654 c_op = pywrap_tf_session.TF_FinishOperation(op_desc)
1655 except errors.InvalidArgumentError as e:
InvalidArgumentError: Shape must be rank 4 but is rank 1 for '{{node conv2d_8/convolution}} = Conv2D[T=DT_FLOAT, data_format="NHWC", dilations=[1, 1, 1, 1], explicit_paddings=[], padding="SAME", strides=[1, 2, 2, 1], use_cudnn_on_gpu=true](input_10, conv2d_8/convolution/ReadVariableOp)' with input shapes: [?,150,150,3], [50].
During handling of the above exception, another exception occurred:
ValueError Traceback (most recent call last)
12 frames
/usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/ops.py in _create_c_op(graph, node_def, inputs, control_inputs, op_def)
1655 except errors.InvalidArgumentError as e:
1656 # Convert to ValueError for backwards compatibility.
-> 1657 raise ValueError(str(e))
1658
1659 return c_op
ValueError: Shape must be rank 4 but is rank 1 for '{{node conv2d_8/convolution}} = Conv2D[T=DT_FLOAT, data_format="NHWC", dilations=[1, 1, 1, 1], explicit_paddings=[], padding="SAME", strides=[1, 2, 2, 1], use_cudnn_on_gpu=true](input_10, conv2d_8/convolution/ReadVariableOp)' with input shapes: [?,150,150,3], [50].