Revision ac62d292d7de9b49855b7ca53f2645a10c63246a authored by rballester on 16 September 2022, 15:17:15 UTC, committed by rballester on 16 September 2022, 15:17:15 UTC
1 parent 363d461
goals.rst
Project Goals
=============
This package was born to bring together some of the most popular tensor decomposition models (including CP, Tucker, and the tensor train) under a common interface. Thus, we use *one class* for all those models. They are all particular cases of `tensor networks <https://arxiv.org/abs/1609.00893>`_, and the idea is that decomposing, manipulating, and reconstructing tensors can be (to some extent) abstracted away from the particular decomposition format.
Building on top of `PyTorch <http://pytorch.org/>`_'s flexibility and built-in automatic differentiation, the overall goal is to exploit those features and allow users to quickly develop, model, and fit various tensor decompositions in a range of data science applications.
Computing file changes ...
