Is there a 3d separable convolution implementation in TF?

[Sanketh Vedula]

Hi!

I want to implement a 3D separable convolution in my network using tensorflow. I was not able to find such function in the documentation.

Is there an implementation for 3D separable convolutions, just like separable_conv2d?

Please let me know if you know about an implementation for such operations.

Cheers,
Sanketh

[Martin Wicke]

I don't believe there is one in core tensorflow. You can check whether there's already an open issue for this, if not, feel free to add a feature request on github.

--
You received this message because you are subscribed to the Google Groups "Discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to discuss+unsubscribe@tensorflow.org.
To post to this group, send email to dis...@tensorflow.org.
To view this discussion on the web visit https://groups.google.com/a/tensorflow.org/d/msgid/discuss/27b3af6f-e9ef-4814-a6ca-0e0b62f39d2f%40tensorflow.org.