Tensor Train decomposition on TensorFlow (T3F)

TL;DR

T3F is a TensorFlow-based library that simplifies implementing Tensor Train decompositions in machine learning, supporting GPU, batch processing, and Riemannian optimization.

Contribution

It introduces a comprehensive, GPU-compatible Tensor Train library with Riemannian optimization support, easing research and development in related machine learning applications.

Findings

Supports GPU acceleration and batch processing

Includes Riemannian optimization framework

Provides extensive documentation and testing

Abstract

Tensor Train decomposition is used across many branches of machine learning. We present T3F -- a library for Tensor Train decomposition based on TensorFlow. T3F supports GPU execution, batch processing, automatic differentiation, and versatile functionality for the Riemannian optimization framework, which takes into account the underlying manifold structure to construct efficient optimization methods. The library makes it easier to implement machine learning papers that rely on the Tensor Train decomposition. T3F includes documentation, examples and 94% test coverage.