问题描述
在“PWCDCNet.py”中，我定义了以下人工神经网络：
class Conv2D(tfk.layers.Layer):
def __init__(self,filters,kernel_size,strides,name=None,padding=1,dilation_rate=1):
super(Conv2D,self).__init__(name=name)
self.conv_out = tfk.layers.Conv2D(filters=filters,kernel_size=kernel_size,strides=strides,padding='same',kernel_initializer='he_normal',dilation_rate=dilation_rate,activation=tfk.layers.LeakyReLU(0.1))
def call(self,inputs):
return self.conv_out(inputs)
class DeConv2D(tfk.layers.Layer):
def __init__(self,kernel_size=4,strides=2,name=None):
super(DeConv2D,self).__init__(name=name)
self.deconv_out = tfk.layers.Conv2DTranspose(filters=filters,name=name)
def call(self,inputs):
return self.deconv_out(inputs)
def CostVolumn(c1,warp,search_range,name='cost_volumn'):
padded_lvl = tf.pad(warp,[[0,0],[search_range,search_range],[0,0]])
_,h,w,_ = tf.unstack(tf.shape(c1))
max_offset = search_range * 2 + 1
cost_vol = []
for y in range(0,max_offset):
for x in range(0,max_offset):
slice = tf.slice(padded_lvl,y,x,[-1,-1])
cost = tf.reduce_mean(c1 * slice,axis=3,keepdims=True)
cost_vol.append(cost)
cost_vol = tf.concat(cost_vol,axis=3)
cost_vol = tf.nn.leaky_relu(cost_vol,alpha=0.1,name=name)
return cost_vol
class PredictFlow(tfk.layers.Layer):
def __init__(self,name=None):
super(PredictFlow,self).__init__()
self.conv_out = tfk.layers.Conv2D(filters=2,kernel_size=3,strides=1,name=name,padding='same')
def call(self,inputs):
return self.conv_out(inputs)
class PWCDCNet(tf.keras.Model):
def __init__(self,max_displacement=4):
super(PWCDCNet,self).__init__()
self.conv1a = Conv2D( 16,name='conv1a')
self.conv1aa = Conv2D( 16,name='conv1aa')
self.conv1b = Conv2D( 16,name='conv1b')
self.conv2a = Conv2D( 32,name='conv2a')
self.conv2aa = Conv2D( 32,name='conv2aa')
self.conv2b = Conv2D( 32,name='conv2b')
self.conv3a = Conv2D( 64,name='conv3a')
self.conv3aa = Conv2D( 64,name='conv3aa')
self.conv3b = Conv2D( 64,name='conv3b')
self.conv4a = Conv2D( 96,name='conv4a')
self.conv4aa = Conv2D( 96,name='conv4aa')
self.conv4b = Conv2D( 96,name='conv4b')
self.conv5a = Conv2D(128,name='conv5a')
self.conv5aa = Conv2D(128,name='conv5aa')
self.conv5b = Conv2D(128,name='conv5b')
self.conv6aa = Conv2D(196,name='conv6aa')
self.conv6a = Conv2D(196,name='conv6a')
self.conv6b = Conv2D(196,name='conv6b')
self.LeakyReLU = tfk.layers.LeakyReLU(0.1,name='LeakyReLU')
self.conv6_0 = Conv2D(128,name='conv6_0')
self.conv6_1 = Conv2D(128,name='conv6_1')
self.conv6_2 = Conv2D(96,name='conv6_2')
self.conv6_3 = Conv2D(64,name='conv6_3')
self.conv6_4 = Conv2D(32,name='conv6_4')
self.deconv6 = DeConv2D(2,name='deconv_6')
self.upfeat6 = DeConv2D(2,name='upfeat_6')
self.conv5_0 = Conv2D(128,name='conv5_0')
self.conv5_1 = Conv2D(128,name='conv5_1')
self.conv5_2 = Conv2D(96,name='conv5_2')
self.conv5_3 = Conv2D(64,name='conv5_3')
self.conv5_4 = Conv2D(32,name='conv5_4')
self.deconv5 = DeConv2D(2,name='deconv_5')
self.upfeat5 = DeConv2D(2,name='upfeat_5')
self.conv4_0 = Conv2D(128,name='conv4_0')
self.conv4_1 = Conv2D(128,name='conv4_1')
self.conv4_2 = Conv2D(96,name='conv4_2')
self.conv4_3 = Conv2D(64,name='conv4_3')
self.conv4_4 = Conv2D(32,name='conv4_4')
self.deconv4 = DeConv2D(2,name='deconv4')
self.upfeat4 = DeConv2D(2,name='upfeat4')
self.conv3_0 = Conv2D(128,name='conv3_0')
self.conv3_1 = Conv2D(128,name='conv3_1')
self.conv3_2 = Conv2D(96,name='conv3_2')
self.conv3_3 = Conv2D(64,name='conv3_3')
self.conv3_4 = Conv2D(32,name='conv3_4')
self.deconv3 = DeConv2D(2,name='deconv3')
self.upfeat3 = DeConv2D(2,name='upfeat3')
self.conv2_0 = Conv2D(128,name='conv2_0')
self.conv2_1 = Conv2D(128,name='conv2_1')
self.conv2_2 = Conv2D(96,name='conv2_2')
self.conv2_3 = Conv2D(64,name='conv2_3')
self.conv2_4 = Conv2D(32,name='conv2_4')
self.deconv2 = DeConv2D(2,name='deconv2')
self.dc_conv1 = Conv2D(128,dilation_rate=1,name='dc_conv1')
self.dc_conv2 = Conv2D(128,padding=2,dilation_rate=2,name='dc_conv2')
self.dc_conv3 = Conv2D(128,padding=4,dilation_rate=4,name='dc_conv3')
self.dc_conv4 = Conv2D(96,padding=8,dilation_rate=8,name='dc_conv4')
self.dc_conv5 = Conv2D(64,padding=16,dilation_rate=16,name='dc_conv5')
self.dc_conv6 = Conv2D(32,name='dc_conv6')
self.predict_flow6 = PredictFlow(name='predict_flow6')
self.predict_flow5 = PredictFlow(name='predict_flow5')
self.predict_flow4 = PredictFlow(name='predict_flow4')
self.predict_flow3 = PredictFlow(name='predict_flow3')
self.predict_flow2 = PredictFlow(name='predict_flow2')
self.dc_conv7 = PredictFlow(name='dc_conv7')
def call(self,inputs):
im1 = inputs[:,:,:3]
im2 = inputs[:,3:]
c11 = self.conv1b(self.conv1aa(self.conv1a(im1)))
c21 = self.conv1b(self.conv1aa(self.conv1a(im2)))
c12 = self.conv2b(self.conv2aa(self.conv2a(c11)))
c22 = self.conv2b(self.conv2aa(self.conv2a(c21)))
c13 = self.conv3b(self.conv3aa(self.conv3a(c12)))
c23 = self.conv3b(self.conv3aa(self.conv3a(c22)))
c14 = self.conv4b(self.conv4aa(self.conv4a(c13)))
c24 = self.conv4b(self.conv4aa(self.conv4a(c23)))
c15 = self.conv5b(self.conv5aa(self.conv5a(c14)))
c25 = self.conv5b(self.conv5aa(self.conv5a(c24)))
c16 = self.conv6b(self.conv6a(self.conv6aa(c15)))
c26 = self.conv6b(self.conv6a(self.conv6aa(c25)))
### 6th flow
corr6 = CostVolumn(c1=c16,warp=c26,search_range=4)
x = tf.concat([self.conv6_0(corr6),corr6],3)
x = tf.concat([self.conv6_1(x),x],3)
x = tf.concat([self.conv6_2(x),3)
x = tf.concat([self.conv6_3(x),3)
x = tf.concat([self.conv6_4(x),3)
flow6 = self.predict_flow6(x)
up_flow6 = self.deconv6(flow6)
up_feat6 = self.upfeat6(x)
### 5th flow
warp5 = bilinear_warp(c25,up_flow6*0.625)
corr5 = CostVolumn(c1=c15,warp=warp5,search_range=4)
x = tf.concat([corr5,c15,up_flow6,up_feat6],3)
x = tf.concat([self.conv5_0(x),3)
x = tf.concat([self.conv5_1(x),3)
x = tf.concat([self.conv5_2(x),3)
x = tf.concat([self.conv5_3(x),3)
x = tf.concat([self.conv5_4(x),3)
flow5 = self.predict_flow5(x)
up_flow5 = self.deconv5(flow5)
up_feat5 = self.upfeat5(x)
### 4th flow
warp4 = bilinear_warp(c24,up_flow5*1.25)
corr4 = CostVolumn(c1=c14,warp=warp4,search_range=4)
x = tf.concat([corr4,c14,up_flow5,up_feat5],3)
x = tf.concat([self.conv4_0(x),3)
x = tf.concat([self.conv4_1(x),3)
x = tf.concat([self.conv4_2(x),3)
x = tf.concat([self.conv4_3(x),3)
x = tf.concat([self.conv4_4(x),3)
flow4 = self.predict_flow4(x)
up_flow4 = self.deconv4(flow4)
up_feat4 = self.upfeat4(x)
### 3rd flow
warp3 = bilinear_warp(c23,up_flow4*2.5)
corr3 = CostVolumn(c1=c13,warp=warp3,search_range=4)
x = tf.concat([corr3,c13,up_flow4,up_feat4],3)
x = tf.concat([self.conv3_0(x),3)
x = tf.concat([self.conv3_1(x),3)
x = tf.concat([self.conv3_2(x),3)
x = tf.concat([self.conv3_3(x),3)
x = tf.concat([self.conv3_4(x),3)
flow3 = self.predict_flow3(x)
up_flow3 = self.deconv3(flow3)
up_feat3 = self.upfeat3(x)
# 2nd flow
warp2 = bilinear_warp(c22,up_flow3*5.0)
corr2 = CostVolumn(c1=c12,warp=warp2,search_range=4)
x = tf.concat([corr2,c12,up_flow3,up_feat3],3)
x = tf.concat([self.conv2_0(x),3)
x = tf.concat([self.conv2_1(x),3)
x = tf.concat([self.conv2_2(x),3)
x = tf.concat([self.conv2_3(x),3)
x = tf.concat([self.conv2_4(x),3)
flow2 = self.predict_flow2(x)
x = self.dc_conv4(self.dc_conv3(self.dc_conv2(self.dc_conv1(x))))
flow2 = flow2 + self.dc_conv7(self.dc_conv6(self.dc_conv5(x)))
return flow2
from PWCNet_tf2.PWCDCNet import PWCDCNet
PWC_model = tf.keras.Model(PWCDCNet())
checkpoint = tf.train.Checkpoint(PWC_model)
save_path = checkpoint.save('/some_directories/PWCNet_tf2/checkpoints/0001/tf_ckpt')
checkpoint.restore(save_path)
但如果我尝试：
output = PWC_model(input)
我收到错误：
NotImplementedError: When subclassing the `Model` class,you should implement a `call` method.
我想知道，因为每个层都实现了调用方法，而且整个 ANN 也实现了调用方法。是我导入模型的方式导致的错误吗？
先谢谢大家！
解决方法
错误是我调用的：
PWC_model = tf.keras.Model(PWCDCNet())
代替：
PWC_model = PWCDCNet()