search

cbovar / ConvNetSharp

Deep Learning in C#
MIT License
468 stars 109 forks source link

Any way to concatinate new information after convolution? #68

Open axxroytovu opened 7 years ago

axxroytovu commented 7 years ago

I'm looking to make a CNN which concatenates new information in on the first fully connected layer. Am I missing something already included or can you implement this feature? Graph flow would go:

Image vvvvvv Convolution Layer vvvvvv Fully Connected Layer <<< Additional Information vvvvvv Output Layer

cbovar commented 7 years ago

Hi. There used to be a 'MergeLayer' to do that. It seems it somehow disappeared after some refacto. I am currently on on holidays with no access to a computer. I can take a look when I'll be back.

cbovar commented 7 years ago

I've found those two classes from version 0.2.0: MergeLayer VolumeWrapper

It can give an idea how to implement a MergeLayer in the current version.

gianlucabertani commented 4 years ago

Hello,

after 3 years, are there any news on this? I need this feature too. Alternatively, is there a correct way to implement the same network using a computational graph? If yes, can you provide an example?

Thanks!