====================================== Tensor methods in Python with TensorLy

This repository contains a series of tutorials and examples on tensor learning, with implementations in Python using TensorLy <https://github.com/tensorly/tensorly>, and how to combine tensor methods and deep learning using the MXNet <https://mxnet.incubator.apache.org/>, PyTorch <http://pytorch.org/>_ and TensorFlow <https://www.tensorflow.org/>__ frameworks as backends.

Installation

You will need to have the latest version of TensorLy installed to run these examples as explained in the instructions <https://tensorly.github.io/dev/installation.html>_.

The easiest way is to clone the repository::

git clone https://github.com/tensorly/tensorly cd tensorly pip install -e .

Then simply clone this repository::

git clone https://github.com/JeanKossaifi/tensorly_notebooks

You are ready to go!

Table of contents

1 - Tensor basics

• Manipulating tensors (unfolding, n-mode product, etc) <https://github.com/tensorly/tensorly-notebooks/blob/master/01_tensor_basics/tensor_manipulation.ipynb>_

2 - Tensor decomposition

• CP decomposition <https://github.com/tensorly/tensorly-notebooks/blob/master/02_tensor_decomposition/cp_decomposition.ipynb>_
• Tucker decomposition <https://github.com/tensorly/tensorly-notebooks/blob/master/02_tensor_decomposition/tucker_decomposition.ipynb>_

3 - Tensor regression

• Low-rank tensor regression <https://github.com/tensorly/tensorly-notebooks/blob/master/03_tensor_regression/Low_rank_tensor_regression.ipynb>_

4 - Tensor methods and deep learning with the MXNet backend

• Tucker decomposition via gradient descent <https://github.com/tensorly/tensorly-notebooks/blob/master/04_mxnet_backend/tucker_decomposition_with_mxnet_and_tensorly.ipynb>_
• Tensor regression networks <https://github.com/tensorly/tensorly-notebooks/blob/master/04_mxnet_backend/tensor_regression_layer_MXNet.ipynb>_

5 - Tensor methods and deep learning with the PyTorch backend

• Tucker decomposition via gradient descent <https://github.com/tensorly/tensorly-notebooks/blob/master/05_pytorch_backend/tucker_decomposition_tensorly_and_pytorch.ipynb>_
• Tensor regression networks <https://github.com/tensorly/tensorly-notebooks/blob/master/05_pytorch_backend/tensor_regression_layer_pytorch.ipynb>_

6 - Tensor methods and deep learning with the TensorFlow backend

• Tucker decomposition via gradient descent <https://github.com/JeanKossaifi/tensorly-notebooks/blob/master/06_tensorflow_backend/tensorflow_tucker.ipynb>__

Useful resources

The following are very useful sources of information and I highly recomment you check them out:

• TensorLy documentation <https://tensorly.github.io/dev/index.html>_ : extensive documentation, API, etc.
• Deep Learning - The Straight Dope <https://github.com/zackchase/mxnet-the-straight-dope>_ : a great tutorial for Deep Learning using MXNet, by Zack Lipton.
• Deep Learning with PyTorch <http://pytorch.org/tutorials/beginner/deep_learning_60min_blitz.html>_ : another great tutorial, this time with PyTorch, by Soumith Chintala.
• The fast.ai cource <https://www.fast.ai/>__ : a great course that teaches Deep Learning from the start, and build up all the way to state-of-the-art models.