Geoopt: Riemannian Optimization in PyTorch

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.

Proceedings of the 37th International Conference on Machine Learning, Vienna, Austria, PMLR 108, 2020, GRLB Workshop

References in corpus (4)

• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• Fast Graph Representation Learning with PyTorch Geometric
• Poincaré Embeddings for Learning Hierarchical Representations
• Hyperbolic Graph Neural Networks

Cited by in corpus (2)

• From Trees to Continuous Embeddings and Back: Hyperbolic Hierarchical Clustering
• Manifolds for Unsupervised Visual Anomaly Detection