Skip to content
This repository has been archived by the owner on Nov 17, 2023. It is now read-only.

Deeplab and atrous convolution #6326

Closed
hondaathma opened this issue May 18, 2017 · 1 comment
Closed

Deeplab and atrous convolution #6326

hondaathma opened this issue May 18, 2017 · 1 comment

Comments

@hondaathma
Copy link

Hi,
I am unable to find atrous convolution layers in mxnet(for deeplab).Can I incorporate any caffe operations and python layers in mxnet?

@Piyush3dB
Copy link
Contributor

For atrous convolution you can use

mx.sym.Convolution(data=net, num_filter=channels, kernel=filter_dimension, stride=stride, dilate=dilate)

where the dilate parameter specifies how wide the convolution kernel reaches.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants