SOTAVerified

Geoopt: Riemannian Optimization in PyTorch

2020-05-06Code Available2· sign in to hype

Max Kochurov, Rasul Karimov, Serge Kozlukov

Code Available — Be the first to reproduce this paper.

Reproduce

Code

Abstract

Geoopt is a research-oriented modular open-source package for Riemannian Optimization in PyTorch. The core of Geoopt is a standard Manifold interface that allows for the generic implementation of optimization algorithms. Geoopt supports basic Riemannian SGD as well as adaptive optimization algorithms. Geoopt also provides several algorithms and arithmetic methods for supported manifolds, which allow composing geometry-aware neural network layers that can be integrated with existing models.

Tasks

Reproductions