Metadata-Version: 2.1
Name: tensorstate
Version: 0.4.0
Summary: 
Author: Nicholas-Schaub
Author-email: nicholas.j.schaub@gmail.com
Requires-Python: >=3.8.1,<4.0.0
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Requires-Dist: numcodecs (>=0.11.0,<0.12.0)
Requires-Dist: numpy (>=1.24.2,<2.0.0)
Requires-Dist: zarr (>=2.14.2,<3.0.0)
Description-Content-Type: text/markdown

# TensorState (v0.4.0) - Neural Network Efficiency Tools

TensorState is a toolbox to capture neural network information to better
understand how information flows through the network. The core of the toolbox is
the ability to capture and analyze neural layer state space, which logs unique
firing states of neural network layers. This repository implements and extends
the work by Schaub and Hotaling in their paper,
[Assessing Intelligence in Artificial Neural Networks](https://arxiv.org/abs/2006.02909).

## Installation

Precompiled wheels exist for Windows, Linux, and MacOS for Python 3.6-3.8. No
special installation instructions are required in most cases:

`pip install pip --upgrade`

`pip install TensorState`

If the wheels don't download or you run into an error, try installing the
pre-requisites for compiling before installing with `pip`.

`pip install pip --upgrade`

`pip install numpy==1.19.2 Cython==3.0a1`

`pip install TensorState`

## Documentation

https://tensorstate.readthedocs.io/en/latest/

