Closed Digital2Slave closed 5 years ago
i have same problem ++
Guys @monder0116, @JohnTian Hi ! I encountered the same problem, I found a solution, but I am not sure that it is 100% correct. So what I did: `from keras.layers import concatenate concat_feat = concatenate([concat_feat, x], mode='concat', axis=concataxis, name='concat'+str(stage)+'_'+str(branch))
I replaced this line with
mergeto
concatenate, because
merge` seems to be deprecated invocation.
i solved this problem via reinstalled the Keras to 2.0.5 version.
@django-kz Thanks for sharing.
@YingXiuHe Thanks for sharing.
、、、 from keras.layers import concatenate concat_feat = concatenate([concat_feat, x], axis=concataxis, name='concat'+str(stage)+'_'+str(branch)) 、、、 in my win10,it works. But 、、、 from keras.layers import concatenate concat_feat = concatenate([concat_feat, x], mode='concat', axis=concataxis, name='concat'+str(stage)+'_'+str(branch)) 、、、 it don't work. I guess that is the version problem of keras. Mine:2.2.4
The latest version of Keras does not support concatenate
Use this Lambda
policy = Lambda(lambda x: x[0]-K.mean(x[0])+x[1], (NUM_ACTIONS,))([advantage, value])
The latest version of Keras does not support
concatenate
Use this
Lambda
policy = Lambda(lambda x: x[0]-K.mean(x[0])+x[1], (NUM_ACTIONS,))([advantage, value])
hi, how to write the code? thanks.
The latest version of Keras does not support
concatenate
Use thisLambda
policy = Lambda(lambda x: x[0]-K.mean(x[0])+x[1], (NUM_ACTIONS,))([advantage, value])
hi, how to write the code? thanks.
Replace concatenate([concat_feat, x], mode='concat', axis=concat_axis, name='concat'+str(stage)+'_'+str(branch))
with policy = Lambda(lambda x: x[0]-K.mean(x[0])+x[1], (NUM_ACTIONS,))([advantage, value])
(This is my code. )
https://www.tensorflow.org/api_docs/python/tf/keras/layers/Lambda
Version
Error