没用过tensorflow, 装的1.14.0，大量warning

[WenjiaWang0312]

我什么也没跑，只是改了下model.py，想看下网络结构
python3 model.py
在最后加入下面内容

if __name__ =='__main__':
    from IPython import embed
    embed()
    s=SRNet()

然后就这样

WARNING:tensorflow:Entity <bound method BatchNormalization.call of <tensorflow.python.layers.normalization.Ba$
chNormalization object at 0x7f3905978cf8>> could not be transformed and will be executed as-is. Please report
this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOS$
TY=10`) and attach the full output. Cause: converting <bound method BatchNormalization.call of <tensorflow.pyt
hon.layers.normalization.BatchNormalization object at 0x7f3905978cf8>>: AssertionError: Bad argument number fo
r Name: 3, expecting 4
WARNING:tensorflow:Entity <bound method Conv.call of <tensorflow.python.layers.convolutional.Conv2D object at
0x7f3905978cf8>> could not be transformed and will be executed as-is. Please report this to the AutgoGraph tea
m. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the ful
l output. Cause: converting <bound method Conv.call of <tensorflow.python.layers.convolutional.Conv2D object a
t 0x7f3905978cf8>>: AssertionError: Bad argument number for Name: 3, expecting 4
WARNING:tensorflow:Entity <bound method BatchNormalization.call of <tensorflow.python.layers.normalization.Bat
chNormalization object at 0x7f3905978cf8>> could not be transformed and will be executed as-is. Please report
this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSI
TY=10`) and attach the full output. Cause: converting <bound method BatchNormalization.call of <tensorflow.pyt
hon.layers.normalization.BatchNormalization object at 0x7f3905978cf8>>: AssertionError: Bad argument number fo
r Name: 3, expecting 4

这种错误报几百行

[WenjiaWang0312]

请问作者大大，是不是少装了什么库？
还有这种错误

ValueError: Variable G/generator_tcn_t_encoder_conv1_1_conv/ker
nel already exists, disallowed. Did you mean to
 set reuse=True or reuse=tf.AUTO_REUSE in VarScope? Originally defined at:

  File "model.py", line 51, in _conv_bn_relu
    x = tf.layers.conv2d(x, cnum, kernel_size = 3, strides = 1, activation = None, padding = padding, name = n
ame + '_conv')
  File "model.py", line 66, in build_encoder_net
    x = self._conv_bn_relu(x, self.cnum, name = name + '_conv1_1')
  File "model.py", line 119, in build_text_conversion_net
    x_t = self.build_encoder_net(x_t, name = name + '_t_encoder')
  File "model.py", line 174, in build_generator
    o_sk, o_t = self.build_text_conversion_net(i_t, i_s, name = name + '_tcn')
  File "model.py", line 186, in build_whole_net_with_loss
    o_sk, o_t, o_b, o_f = self.build_generator(inputs)

[ursaminor0401]

我什么也没跑，只是改了下model.py，想看下网络结构
python3 model.py
在最后加入下面内容

if __name__ =='__main__':
    from IPython import embed
    embed()
    s=SRNet()

然后就这样

WARNING:tensorflow:Entity <bound method BatchNormalization.call of <tensorflow.python.layers.normalization.Ba$
chNormalization object at 0x7f3905978cf8>> could not be transformed and will be executed as-is. Please report
this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOS$
TY=10`) and attach the full output. Cause: converting <bound method BatchNormalization.call of <tensorflow.pyt
hon.layers.normalization.BatchNormalization object at 0x7f3905978cf8>>: AssertionError: Bad argument number fo
r Name: 3, expecting 4
WARNING:tensorflow:Entity <bound method Conv.call of <tensorflow.python.layers.convolutional.Conv2D object at
0x7f3905978cf8>> could not be transformed and will be executed as-is. Please report this to the AutgoGraph tea
m. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the ful
l output. Cause: converting <bound method Conv.call of <tensorflow.python.layers.convolutional.Conv2D object a
t 0x7f3905978cf8>>: AssertionError: Bad argument number for Name: 3, expecting 4
WARNING:tensorflow:Entity <bound method BatchNormalization.call of <tensorflow.python.layers.normalization.Bat
chNormalization object at 0x7f3905978cf8>> could not be transformed and will be executed as-is. Please report
this to the AutgoGraph team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSI
TY=10`) and attach the full output. Cause: converting <bound method BatchNormalization.call of <tensorflow.pyt
hon.layers.normalization.BatchNormalization object at 0x7f3905978cf8>>: AssertionError: Bad argument number fo
r Name: 3, expecting 4

这种错误报几百行

你好，这种warning是因为gast包版本问题，pip install gast==0.2.2后会消除这些warning

[ursaminor0401]

请问作者大大，是不是少装了什么库？
还有这种错误

ValueError: Variable G/generator_tcn_t_encoder_conv1_1_conv/ker
nel already exists, disallowed. Did you mean to
 set reuse=True or reuse=tf.AUTO_REUSE in VarScope? Originally defined at:

  File "model.py", line 51, in _conv_bn_relu
    x = tf.layers.conv2d(x, cnum, kernel_size = 3, strides = 1, activation = None, padding = padding, name = n
ame + '_conv')
  File "model.py", line 66, in build_encoder_net
    x = self._conv_bn_relu(x, self.cnum, name = name + '_conv1_1')
  File "model.py", line 119, in build_text_conversion_net
    x_t = self.build_encoder_net(x_t, name = name + '_t_encoder')
  File "model.py", line 174, in build_generator
    o_sk, o_t = self.build_text_conversion_net(i_t, i_s, name = name + '_tcn')
  File "model.py", line 186, in build_whole_net_with_loss
    o_sk, o_t, o_b, o_f = self.build_generator(inputs)

我这里没有出现这种问题，看样子是节点名重复了，有修改过模型结构么？

[wode-czw]

太谢谢了，舒服了，查了各种关闭 warninig的博客都没解决

[gioipv]

I did use pytorch version instead of this version, becasause of, in this version, i can not use gpu despite of i set os.environ['CUDA...']='0' . I also try fix this issue by search on google search to fix. set tf.session(config) ..... but it still cannot run on gpu.