Python Neural Network

This library sports a fully connected neural network written in Python with NumPy. The network can be trained by a variety of learning algorithms: backpropagation, resilient backpropagation, scaled conjugate gradient and SciPy's optimize function. The library was developed with PYPY in mind and should play nicely with their super-fast JIT compiler.


With pip

pip install nimblenet

Easily install the latest version of nimblenet with pip. The only requirement is NumPy. Additionally, SciPy is needed if you would like to train the network using SciPy's optimize() function.

From Github

git clone

Fork or clone the Github repo to get started. Feel free to report Github Issues if you run into bugs.



The library has been implemented with performance in focus and uses NumPy to provide fast calculations.

Easily extendible

The activation and cost functions are effortlessly interchangeable and it is easy to define new functions.

In active development

The code is actively maintained and the issues reported on Github are always attended to.