A Python Package for Advanced Tensor Learning Methods
Project description
tensorlearn
tensorlearn is a Python library distributed on Pypi for implementing tensor learning
This is a package under development. Yet, the available methods are final and functional. The backend is Numpy.
Installation
Use the package manager pip to install tensorlearn in Python.
pip install tensorlearn
methods
Decomposition Methods
Tensor-Train Decomposition with Auto Rank Determination
Tensor-Train Decomposition with Auto Rank Determination
Tensorlearn.decomposition.tensor_train.auto_rank_tt(tensor,epsilon)
This implementation of tensor-train decomposition determines rank automatically based on a given error bound written according to Oseledets (2011). Therefore the user does not need to specify a rank. Instead the user specifies an upper error bound (epsilon) which is the frobenius norm of the error divided by the frobenius norm of the given tensor to be decomposed.
Arguments
@tensor - dimension must be at least 3.
@epsilon - Error bound = frobenius norm of the error / frobenius norm of the given tensor.
Outputs
@factors - The list includes 2D numpy arrays of factors according to TT decomposition. Length of the list equals the dimension of the given tensor to be decomposed.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for tensorlearn-1.0.65-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | bdda727fd96592ea60a20f880562b80a145ab9f034b0f60dd0ab3dd80f99b84f |
|
MD5 | 01b47f72562f1bad16cf3c8fe58f81f3 |
|
BLAKE2b-256 | 88a317a38abf5a8d3db35b63a382843c6186e5626cdbdc5d4324bf6410541ef5 |