Hfrxdjt

Airplane gently rocking its wings during whole flight

Spaces in which all closed sets are regular closed

Why is the US ranked as #45 in Press Freedom ratings, despite its extremely permissive free speech laws?

Won the lottery - how do I keep the money?

Could a dragon use its wings to swim?

Why do we say 'Un seul M' and not 'Une seule M' even though M is a "consonne"

Is there such a thing as a proper verb, like a proper noun?

How do I fit a non linear curve?

Is there an equivalent of cd - for cp or mv

IC has pull-down resistors on SMBus lines?

Is it ok to trim down a tube patch?

How to use ReplaceAll on an expression that contains a rule

If Nick Fury and Coulson already knew about aliens (Kree and Skrull) why did they wait until Thor's appearance to start making weapons?

What steps are necessary to read a Modern SSD in Medieval Europe?

Yu-Gi-Oh cards in Python 3

Is it okay to majorly distort historical facts while writing a fiction story?

Is it professional to write unrelated content in an almost-empty email?

How do you define an element with an ID attribute using LWC?

Getting Stale Gas Out of a Gas Tank w/out Dropping the Tank

how one can write a nice vector parser, something that does pgfvecparseA=B-C; D=E x F;

What CSS properties can the br tag have?

Do I need to write [sic] when including a quotation with a number less than 10 that isn't written out?

Would a grinding machine be a simple and workable propulsion system for an interplanetary spacecraft?

Are the names of these months realistic?

implementing conv2d in fourier domain using np.einsum --> ValueError: einstein sum subscripts string contains too many subscripts for operand 0

0

$begingroup$

According to the convolution theorem, convolution operation changes to pointwise multiplication in fourier domain - here I have 'fft_x' of shape (batchsize, height, width, in_channels) which is the fft of input data and similarly 'fft_kernel' of shape (height, width, in_channels, out_channels) which is fft of the kernel after being padded to image size. To get pointwise multiplication of these in efficient way, I was using einsum in the following way -

...
 print(fft_x)
 print(fft_kernel)
 output = 0
 n=int(self.no_of_kernels/2)+1 # n = out_channels here
 for i in range(n):
 output += np.einsum('ijkl,jkl->ijk', fft_x, fft_kernel[i])
 return output 
...

It gives the following output and error-

Tensor("input_11:0", shape=(?, 28, 28, 1), dtype=complex64)
Tensor("fourier__conv2d_11/transpose:0", shape=(28, 28, 1, 17), dtype=complex64)
...
...
ValueError: einstein sum subscripts string contains too many subscripts for operand 0

Could anyone please explain why this error is arising? Thanks in advance for any help.

python tensorflow numpy

share|improve this question

asked 26 mins ago

psjpsj

112

$endgroup$

add a comment | 

0

$begingroup$

According to the convolution theorem, convolution operation changes to pointwise multiplication in fourier domain - here I have 'fft_x' of shape (batchsize, height, width, in_channels) which is the fft of input data and similarly 'fft_kernel' of shape (height, width, in_channels, out_channels) which is fft of the kernel after being padded to image size. To get pointwise multiplication of these in efficient way, I was using einsum in the following way -

...
 print(fft_x)
 print(fft_kernel)
 output = 0
 n=int(self.no_of_kernels/2)+1 # n = out_channels here
 for i in range(n):
 output += np.einsum('ijkl,jkl->ijk', fft_x, fft_kernel[i])
 return output 
...

It gives the following output and error-

Tensor("input_11:0", shape=(?, 28, 28, 1), dtype=complex64)
Tensor("fourier__conv2d_11/transpose:0", shape=(28, 28, 1, 17), dtype=complex64)
...
...
ValueError: einstein sum subscripts string contains too many subscripts for operand 0

Could anyone please explain why this error is arising? Thanks in advance for any help.

python tensorflow numpy

share|improve this question

asked 26 mins ago

psjpsj

112

$endgroup$

add a comment | 

$begingroup$

According to the convolution theorem, convolution operation changes to pointwise multiplication in fourier domain - here I have 'fft_x' of shape (batchsize, height, width, in_channels) which is the fft of input data and similarly 'fft_kernel' of shape (height, width, in_channels, out_channels) which is fft of the kernel after being padded to image size. To get pointwise multiplication of these in efficient way, I was using einsum in the following way -

...
 print(fft_x)
 print(fft_kernel)
 output = 0
 n=int(self.no_of_kernels/2)+1 # n = out_channels here
 for i in range(n):
 output += np.einsum('ijkl,jkl->ijk', fft_x, fft_kernel[i])
 return output 
...

It gives the following output and error-

Tensor("input_11:0", shape=(?, 28, 28, 1), dtype=complex64)
Tensor("fourier__conv2d_11/transpose:0", shape=(28, 28, 1, 17), dtype=complex64)
...
...
ValueError: einstein sum subscripts string contains too many subscripts for operand 0

Could anyone please explain why this error is arising? Thanks in advance for any help.

python tensorflow numpy

share|improve this question

asked 26 mins ago

psjpsj

112

$endgroup$

...
 print(fft_x)
 print(fft_kernel)
 output = 0
 n=int(self.no_of_kernels/2)+1 # n = out_channels here
 for i in range(n):
 output += np.einsum('ijkl,jkl->ijk', fft_x, fft_kernel[i])
 return output 
...

Tensor("input_11:0", shape=(?, 28, 28, 1), dtype=complex64)
Tensor("fourier__conv2d_11/transpose:0", shape=(28, 28, 1, 17), dtype=complex64)
...
...
ValueError: einstein sum subscripts string contains too many subscripts for operand 0

share|improve this question

asked 26 mins ago

psjpsj

112

share|improve this question

asked 26 mins ago

psjpsj

112

asked 26 mins ago

psjpsj

112

asked 26 mins ago

psjpsj

112