Stable release

To install TLViz, run this command in your terminal:

pip install tensorly-viz

Alternatively, to get the latest development version of TLViz, run:

pip install git+

Interoperability with TensorLy

TLViz doesn’t strictly depend on TensorLy. However, we recommend installing both if you want to analyse multi-way data with tensor decomposition.

Optional dependencies

By default, TLViz uses SciPy to generate QQ-plots. However, the GPL lisenced Pingouin package can be used instead if you want better looking QQ-plots with confidence intervals. To install the optional (GPL lisenced) dependencies run

pip install tensorly-viz[all]

Running the examples

The examples depend on a variety of extra packages, such as TensorLy and PlotLy. To install these as well, run

pip install tensorly-viz[docs]

Running the test suite

The test suite have some extra dependencies, such as pytest and pytorch. To install these, run

pip install tensorly-viz[test]

Installing multiple extra dependencies

If you want to install multiple extra dependencies at once (e.g. docs and test), consider running

pip install tensorly-viz[docs,test]