Trjtdtk

Watching something be piped to a file live with tail

Forgetting the musical notes while performing in concert

How could indestructible materials be used in power generation?

Dealing with conflict between co-workers for non-work-related issue affecting their work

What's the point of deactivating Num Lock on login screens?

Neighboring nodes in the network

How can saying a song's name be a copyright violation?

Why is Collection not simply treated as Collection<?>

Why are electrically insulating heatsinks so rare? Is it just cost?

Why doesn't using multiple commands with a || or && conditional work?

Is it canonical bit space?

How do I deal with an unproductive colleague in a small company?

Is there a hemisphere-neutral way of specifying a season?

What does it mean to describe someone as a butt steak?

Has there ever been an airliner design involving reducing generator load by installing solar panels?

Alternative to sending password over mail?

How do I find out when a node was added to an availability group?

Personal Teleportation: From Rags to Riches

Does casting Light, or a similar spell, have any effect when the caster is swallowed by a monster?

Why does Optional.map make this assignment work?

How can I tell some body that I want to be his or her friend?

Why would the Red Woman birth a shadow if she worshipped the Lord of the Light?

Is it inappropriate for a student to attend their mentor's dissertation defense?

Contemplating leaving job of two years, but a long travel is in the works. Is it professional for me to resign?

Controlling number of channels in weight/kernel in tensorflow

1

$begingroup$

To implement a specific function, I need "input_channels" number of kernels in my layer, each having only a single channel depth, and not depth = "input_channels". I need to convolve one kernel with one channel of the input, thus the output of the layer would have "input_channels" number of kernels.

Image attached for reference -

Thanks in advance for any help.

(if anyone wishes to know what all i have tried yet -
In the conv2d function of tensorflow, if I specify number of kernels = 1 to do this, then it will sum over all input_channels and number of output_channels will be 1, since it always initialises kernel depth = "input_channels".

Another option is to specify number of number of kernels = input_channels in conv2d function but this would create "input_channels" number of kernels of depth "input_channels", thus adding lot of complexity and incorrect implementation of my layer.

Yet another thing I tried was to initialise a kernel of volume (kernel_height, kernel_width, input_channels) and loop over the third dimension to convolve only a single input channel with a single kernel. But the tensorflow conv2d function requires a rank 4 kernel to work and gives the following error -

ValueError: Shape must be rank 4 but is rank 3 for 'generic_act_func_4/Conv2D' (op: 'Conv2D') with input shapes: [?,28,28], [28,28].

)

python tensorflow

share|improve this question

edited Mar 26 at 8:29

psj

asked Mar 26 at 7:58

psjpsj

112

$endgroup$

• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  Dear, probably the error occurs in (kernel_height, kernel_width, input_channels), its 3 parameters (means rank 3), but the error said it should be rank for, means it needs the fourth parameters, which I think the number of the filters in this conv.
  $endgroup$
  – honar.cs
  Mar 26 at 13:09
  

  
  
  
  

add a comment | 

1

$begingroup$

To implement a specific function, I need "input_channels" number of kernels in my layer, each having only a single channel depth, and not depth = "input_channels". I need to convolve one kernel with one channel of the input, thus the output of the layer would have "input_channels" number of kernels.

Image attached for reference -

Thanks in advance for any help.

(if anyone wishes to know what all i have tried yet -
In the conv2d function of tensorflow, if I specify number of kernels = 1 to do this, then it will sum over all input_channels and number of output_channels will be 1, since it always initialises kernel depth = "input_channels".

Another option is to specify number of number of kernels = input_channels in conv2d function but this would create "input_channels" number of kernels of depth "input_channels", thus adding lot of complexity and incorrect implementation of my layer.

Yet another thing I tried was to initialise a kernel of volume (kernel_height, kernel_width, input_channels) and loop over the third dimension to convolve only a single input channel with a single kernel. But the tensorflow conv2d function requires a rank 4 kernel to work and gives the following error -

ValueError: Shape must be rank 4 but is rank 3 for 'generic_act_func_4/Conv2D' (op: 'Conv2D') with input shapes: [?,28,28], [28,28].

)

python tensorflow

share|improve this question

edited Mar 26 at 8:29

psj

asked Mar 26 at 7:58

psjpsj

112

$endgroup$

• 
  
  
  

  
  

  
  
  
  
  
  

  
  $begingroup$
  Dear, probably the error occurs in (kernel_height, kernel_width, input_channels), its 3 parameters (means rank 3), but the error said it should be rank for, means it needs the fourth parameters, which I think the number of the filters in this conv.
  $endgroup$
  – honar.cs
  Mar 26 at 13:09
  

  
  
  
  

add a comment | 

$begingroup$

To implement a specific function, I need "input_channels" number of kernels in my layer, each having only a single channel depth, and not depth = "input_channels". I need to convolve one kernel with one channel of the input, thus the output of the layer would have "input_channels" number of kernels.

Image attached for reference -

Thanks in advance for any help.

(if anyone wishes to know what all i have tried yet -
In the conv2d function of tensorflow, if I specify number of kernels = 1 to do this, then it will sum over all input_channels and number of output_channels will be 1, since it always initialises kernel depth = "input_channels".

Another option is to specify number of number of kernels = input_channels in conv2d function but this would create "input_channels" number of kernels of depth "input_channels", thus adding lot of complexity and incorrect implementation of my layer.

Yet another thing I tried was to initialise a kernel of volume (kernel_height, kernel_width, input_channels) and loop over the third dimension to convolve only a single input channel with a single kernel. But the tensorflow conv2d function requires a rank 4 kernel to work and gives the following error -

ValueError: Shape must be rank 4 but is rank 3 for 'generic_act_func_4/Conv2D' (op: 'Conv2D') with input shapes: [?,28,28], [28,28].

)

python tensorflow

share|improve this question

edited Mar 26 at 8:29

psj

asked Mar 26 at 7:58

psjpsj

112

$endgroup$

ValueError: Shape must be rank 4 but is rank 3 for 'generic_act_func_4/Conv2D' (op: 'Conv2D') with input shapes: [?,28,28], [28,28].

share|improve this question

edited Mar 26 at 8:29

psj

asked Mar 26 at 7:58

psjpsj

112

share|improve this question

edited Mar 26 at 8:29

psj

asked Mar 26 at 7:58

psjpsj

112

asked Mar 26 at 7:58

psjpsj

112

asked Mar 26 at 7:58

psjpsj

112