The Nengo team is thrilled to announce the release of Nengo 2.2!
Nengo is a Python library for building and simulating large-scale neural models for AI and robotics. It can be thought of as a neural compiler, transforming a functional description of a neural model to a network of spiking or non-spiking neurons that can run on multiple backends including GPUs and neuromorphic hardware.
Unlike the jump to Nengo 2.1, not a whole lot has changed for existing models in Nengo 2.2. Everything that worked in Nengo 2.1 should still work after upgrading!
Nengo 2.2 features a number of nice improvements,
however, that can make some models faster.
If you don’t know the exact function
that you want to compute in a connection,
but you have a set of example input / output pairs,
it is now possible to pass a set of outputs
function in a connection,
as long as you also pass in the inputs as
If you’ve used the
then you should use this new feature instead.
We have also introduced the idea of
“config presets” to group together
sets of parameters that are commonly used together.
For example, when making ensembles that
should be silent up to a certain input threshold,
several parameters need to be set
to get good performance.
Those parameters can be set all at once with
If you’re still using the ancient NumPy 1.6, then now is a good time to update, as this new release drops support for NumPy 1.6.
To see the full list of changes in Nengo 2.2, head to the Github release page.
To get the new version of Nengo, use
pip install --upgrade nengo
Please come to the new Nengo forum! We welcome any questions and suggestions you might have, and invite you to share your Nengo creations there.