cd-xlyang / tednet

TedNet: A Pytorch Toolkit for Tensor Decomposition Networks

TedNet: A Pytorch Toolkit for Tensor Decomposition Networks

tednet is a toolkit for tensor decomposition networks. Tensor decomposition networks are neural networks whose layers are decomposed by tensor decomposition, including CANDECOMP/PARAFAC, Tucker2, Tensor Train, Tensor Ring and so on. For a convenience to do research on it, tednet provides excellent tools to deal with tensorial networks.

Now, tednet is easy to be installed by pip:

pip install tednet

More information could be found in Document.

---

Quick Start

Operation

There are some operations supported in tednet, and it is convinient to use them. First, import it:

import tednet as tdt

Create matrix whose diagonal elements are ones:

diag_matrix = tdt.eye(5, 5)

A way to transfer the Pytorch tensor into numpy array:

diag_matrix = tdt.to_numpy(diag_matrix)

Similarly, the numpy array can be taken into Pytorch tensor by:

diag_matrix = tdt.to_tensor(diag_matrix)

Tensor Decomposition Networks (Tensor Ring for Sample)

To use tensor ring decomposition models, simply calling the tensor ring module is enough.

import tednet.tnn.tensor_ring as tr

# Define a TR-LeNet5
model = tr.TRLeNet5(10, [6, 6, 6, 6])

---

Citing

If you use tednet in an academic work, we will appreciate you for citing our paper with:

@article{DBLP:journals/corr/abs-2104-05018,
  author    = {Yu Pan and
               Maolin Wang and
               Zenglin Xu},
  title     = {TedNet: {A} Pytorch Toolkit for Tensor Decomposition Networks},
  journal   = {CoRR},
  volume    = {abs/2104.05018},
  year      = {2021},
  url       = {https://arxiv.org/abs/2104.05018},
  archivePrefix = {arXiv},
  eprint    = {2104.05018},
  timestamp = {Mon, 19 Apr 2021 16:45:47 +0200},
  biburl    = {https://dblp.org/rec/journals/corr/abs-2104-05018.bib},
  bibsource = {dblp computer science bibliography, https://dblp.org}
}