Recently _Geoffrey Hinton_ and his team introduced a new interesting deep neural network model, called CapsNet.
I think the implementation of this Net on PyTorch and train it on ImageNet Data set, and after that add its pre-trained model to TorchVision Models would be great for the PyTorch community.
Thanks to the PyTorch team
As far as I know, Hinton's group doesn't show experiments on large-scale dataset -- they only tested MNIST.
Most helpful comment
As far as I know, Hinton's group doesn't show experiments on large-scale dataset -- they only tested MNIST.