Ufdjrw

Terse Method to Swap Lowest for Highest?

Angel of Condemnation - Exile creature with second ability

How do apertures which seem too large to physically fit work?

What should you do if you miss a job interview (deliberately)?

Quoting Keynes in a lecture

Why did the EU agree to delay the Brexit deadline?

Does the Linux kernel need a file system to run?

Why does the Sun have different day lengths, but not the gas giants?

Has any country ever had 2 former presidents in jail simultaneously?

How can mimic phobia be cured?

creating a ":KeepCursor" command

How could a planet have erratic days?

Quasinilpotent , non-compact operators

Does Doodling or Improvising on the Piano Have Any Benefits?

Why is the "ls" command showing permissions of files in a FAT32 partition?

What is the highest possible scrabble score for placing a single tile

Is aluminum electrical wire used on aircraft?

Redundant comparison & "if" before assignment

How do you respond to a colleague from another team when they're wrongly expecting that you'll help them?

PTIJ: Haman's bad computer

Why does AES have exactly 10 rounds for a 128-bit key, 12 for 192 bits and 14 for a 256-bit key size?

Recommended PCB layout understanding - ADM2572 datasheet

Fear of getting stuck on one programming language / technology that is not used in my country

Using substitution ciphers to generate new alphabets in a novel

finding the inner product of two 3D tensor in keras custom layer

What is the best way to reshape tensor in kerasUsing the K.function() within a custom Keras layerKeras custom layer with no different output_shapeCustom Lambda layer for Kronecker product in Keras - troubles with the dimension reserved for batch_sizeIndexing a Keras TensorHow to create a custom convolution operation with a trainable Kernel in Keras?Multiply multiple tensors pairwise kerasHow to apply convolution on the last three dimensions of a 5D tensor using the Conv2D in Keras?The initial state or constants of an RNN layer cannot be specified with a mix of Keras tensors and non-Keras tensorsUnable to stack keras 1D tensors to form 2D keras tensor

0

in my keras custom layer, i have to find the inner product of two 3D tensors. For example , x=(?,80,150,12) is the input to the layer and the inner product have to be with all the 4 slices of kernel (80,150,12,4), 4 being the slicing dimension. I am trying to do it with K.dot but i am getting dimension mismatch error.The code snippet is-

kernel1_unpacked = tf.unstack(self.kernel1,axis=3) 
 all_probs = []

 for t in kernel1_unpacked:
 t=K.permute_dimensions(t,[2,0,1])
 B3=K.sum(K.dot(x,t),[1,2,3])
 B3=K.expand_dims(B3,1)
 print(B3)
 all_probs.append(B3)

 B4=tf.concat(all_probs[:],axis=1)

 print(B4)

the output should have dimension (?,4). please help

python tensorflow keras inner-product

share|improve this question

asked Mar 7 at 6:15

Sandeep PandeySandeep Pandey

185

• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  @giser_yugang Sir, any idea how to get around this ?
  
  – Sandeep Pandey
  Mar 8 at 5:43
  

  
  
  
  

add a comment | 

0

in my keras custom layer, i have to find the inner product of two 3D tensors. For example , x=(?,80,150,12) is the input to the layer and the inner product have to be with all the 4 slices of kernel (80,150,12,4), 4 being the slicing dimension. I am trying to do it with K.dot but i am getting dimension mismatch error.The code snippet is-

kernel1_unpacked = tf.unstack(self.kernel1,axis=3) 
 all_probs = []

 for t in kernel1_unpacked:
 t=K.permute_dimensions(t,[2,0,1])
 B3=K.sum(K.dot(x,t),[1,2,3])
 B3=K.expand_dims(B3,1)
 print(B3)
 all_probs.append(B3)

 B4=tf.concat(all_probs[:],axis=1)

 print(B4)

the output should have dimension (?,4). please help

python tensorflow keras inner-product

share|improve this question

asked Mar 7 at 6:15

Sandeep PandeySandeep Pandey

185

• 
  
  
  

  
  

  
  
  
  
  
  

  
  
  @giser_yugang Sir, any idea how to get around this ?
  
  – Sandeep Pandey
  Mar 8 at 5:43
  

  
  
  
  

add a comment | 

in my keras custom layer, i have to find the inner product of two 3D tensors. For example , x=(?,80,150,12) is the input to the layer and the inner product have to be with all the 4 slices of kernel (80,150,12,4), 4 being the slicing dimension. I am trying to do it with K.dot but i am getting dimension mismatch error.The code snippet is-

kernel1_unpacked = tf.unstack(self.kernel1,axis=3) 
 all_probs = []

 for t in kernel1_unpacked:
 t=K.permute_dimensions(t,[2,0,1])
 B3=K.sum(K.dot(x,t),[1,2,3])
 B3=K.expand_dims(B3,1)
 print(B3)
 all_probs.append(B3)

 B4=tf.concat(all_probs[:],axis=1)

 print(B4)

the output should have dimension (?,4). please help

python tensorflow keras inner-product

share|improve this question

asked Mar 7 at 6:15

Sandeep PandeySandeep Pandey

185

kernel1_unpacked = tf.unstack(self.kernel1,axis=3) 
 all_probs = []

 for t in kernel1_unpacked:
 t=K.permute_dimensions(t,[2,0,1])
 B3=K.sum(K.dot(x,t),[1,2,3])
 B3=K.expand_dims(B3,1)
 print(B3)
 all_probs.append(B3)

 B4=tf.concat(all_probs[:],axis=1)

 print(B4)

share|improve this question

asked Mar 7 at 6:15

Sandeep PandeySandeep Pandey

185

share|improve this question

asked Mar 7 at 6:15

Sandeep PandeySandeep Pandey

185

asked Mar 7 at 6:15

Sandeep PandeySandeep Pandey

185

asked Mar 7 at 6:15

Sandeep PandeySandeep Pandey

185