CapsNet-PyTorch

I hvae crated CapsuleConv for PrimaryCapsule, CapsuleLinear for DigitCapsule

minist_toy_example.py: minist example from pytorch github code repo

Usage

Step-1: Set your environment

virtualenv -p python3 --no-site-packages .capsule_env

source .capsule_env/bin/activate
pip3 install http://download.pytorch.org/whl/cu80/torch-0.2.0.post3-cp35-cp35m-manylinux1_x86_64.whl 
pip3 install torchvision

Step-2: Run

CUDA_VISIBLE_DEVICES=2 python main.py

Todo

1. Test for cifar10
2. Re-Implement EM-Routing

Other Implementations

• Kaggle (this version as self-contained notebook):

  • MNIST Dataset running on the standard MNIST and predicting for test data
  • MNIST Fashion running on the more challenging Fashion images.

• TensorFlow:

  • naturomics/CapsNet-Tensorflow
    Very good implementation. I referred to this repository in my code.
  • InnerPeace-Wu/CapsNet-tensorflow
    I referred to the use of tf.scan when optimizing my CapsuleLayer.
  • LaoDar/tf_CapsNet_simple

• PyTorch:

  • nishnik/CapsNet-PyTorch
  • timomernick/pytorch-capsule
  • gram-ai/capsule-networks
  • andreaazzini/capsnet.pytorch
  • leftthomas/CapsNet

• MXNet:

  • AaronLeong/CapsNet_Mxnet

• Lasagne (Theano):

  • DeniskaMazur/CapsNet-Lasagne

• Chainer:

  • soskek/dynamic_routing_between_capsules