Group Abstract Group Abstract

Message Boards Message Boards

0
|
120 Views
|
3 Replies
|
1 Total Like
View groups...
Share
Share this post:

Error creating convolution layer: inferred in consistent value for input size

Iuval Clejan
Posted 8 days ago

I'd like to have a convolutional layer with 3 outputs o1,o2,o3, and 9 inputs i1,,,i9, so that
o1=k11 i1+k12 i2+k13 i3,
o2=k21 i4+k22 i5+k23 i6,
o3=k31 i7+k32 i8+k33 i9
where the ks are either 3 kernels or 1, I don't care. I am getting an error. Please let me know how to do it right.
POSTED BY: Iuval Clejan
3 Replies
Sort By:

From the documentation, "Input" must be an array.

POSTED BY: Gianluca Gorni
Iuval Clejan
Iuval Clejan
Posted 8 days ago

OK, I thought a vector would be fine, but I made it into a 1x9 array. But now the output is a 3x3 array instead of a 3-vector. I need the diagonal vector of the 3x3 output matrix (see enclosed file)
POSTED BY: Iuval Clejan
Iuval Clejan
Iuval Clejan
Posted 8 days ago

Also, see enclosed pic of the 13.3 Mathematica documentation I have. It's for NetChain, but you would think it would apply to ConvolutionLayer is well (I.e. vector input is allowed).

Attachment

Attachments:
Screen Shot 2025...png
POSTED BY: Iuval Clejan
Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Tag limit exceeded
Note: Only the first five people you tag will receive an email notification; the other tagged names will appear as links to their profiles.
Reply Preview
Attachments
Remove
or Discard