Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add support for dilated convolutions #3922

Merged
merged 5 commits into from Aug 24, 2017

Conversation

@AlexDBlack
Copy link
Contributor

commented Aug 23, 2017

Complete, should be good to go (maybe pending review).

  • Adds dilation (atrous convolutions) for both ConvolutionLayer and subsampling layer
  • Unlike Keras: no restriction on combinations of strides and dilation (Keras requires stride 1 with dilation > 1)
  • Gradient checks: added and passing
  • CuDNN Convolution layers support dilation (subsampling don't). Added + tested

deeplearning4j/nd4j#2064
deeplearning4j/libnd4j#439

@maxpumperla

This comment has been minimized.

Copy link

commented Aug 23, 2017

closes #2950

@maxpumperla
Copy link

left a comment

LGTM! +1 for documentation and minor clean-up also.

(As mentioned in gitter, you seem to return null whenever cudnn fails and I can only assume this has the intended effect / throws the right error, but I simply don't know).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
2 participants
You can’t perform that action at this time.