Another SPDnet implementation

This repository contains an alternative implementation of SPDnet from:

Main differences are:

Extensive relying on torch.einsum for increased computational speed
Grad of BiMap coded with einsum rather than using autograd
Works for any number of batch dimensions (rather than 1 for Davoudi and 2 for torchspdnet)
bias term on the eigenvalues in the reEig layers ?

Navigation Menu