论文标题
Geoopt：Pytorch中的Riemannian优化
Geoopt: Riemannian Optimization in PyTorch
论文作者
论文摘要
Geoopt是一个面向研究的模块化开源软件包，用于Pytorch中的Riemannian优化。 GEOOPT的核心是标准歧管接口，允许对优化算法的一般实现。 Geoopt支持基本的Riemannian SGD以及自适应优化算法。 Geoopt还为支持的流形提供了几种算法和算术方法，这些方法允许组成几何感知的神经网络层，这些神经网络层可以与现有模型集成。
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.
