search

naturomics / CapsNet-Tensorflow

A Tensorflow implementation of CapsNet(Capsules Net) in paper Dynamic Routing Between Capsules
Apache License 2.0
3.8k stars 1.17k forks source link

Training on different input dimensions than MNIST #19

Closed servetcoskun closed 6 years ago

servetcoskun commented 6 years ago

Thanks for writing the code so shortly after the article was released. I'm trying to change the structure such that the capsule network can be trained for any image(x,y,z), but I am having trouble re-structuring the code. Can you help me identify which lines needs to be modified. I am guessing all lines with ... 28, 28, 1) -> ... 32, 32, 3) for CIFAR 10. But I am still not able to make it work.

Thank you again 👍

naturomics commented 6 years ago

@servetcoskun I'm going to release a new version which is dimensional compatible for different image size in the next few days. Maybe you can wait for a few days, I'm trying my best to make this code to be more reusable and easier to use.

Thanks for your follow

servetcoskun commented 6 years ago

Is it possible to get an update :)? Good work on the commits 👍

utkarshcontractor commented 6 years ago

Thanks for sharing the implementation @naturomics ! dimensional compatible for different image sizes would be very useful to adapt caps net to different use cases. Thanks in advance

djoffrey commented 6 years ago

Has any update to this issue commited?

naturomics commented 6 years ago

I've created a new project CapsLayer, which should be able to meet the dimensional compatible requirement. Please see the details in the project homepage.