The Nengo team at ABR is excited to announce the release of Nengo Loihi 0.3.
What is Nengo Loihi?
Nengo Loihi is a backend for the Nengo neural simulator allowing standard spiking Nengo models to run on Intel’s new Loihi neuromorphic chip. Nengo Loihi includes an emulator so that you can develop spiking neuron models for Loihi without having access to hardware.
How do I use it?
Once installed, Nengo Loihi simulates Nengo networks. After defining your network, instead of simulating it with
nengo.Simulator(model)
do
nengo_loihi.Simulator(model)
and that’s it!
What’s new?
Nengo Loihi is still in heavy development, so features are changing rapidly and many bugs have been discovered and fixed.
In version 0.3.0, we have added support for the nengo.SpikingRectifiedLinear neuron type, which can be run in the emulator and on real hardware. Models can also be run with different dt values, which can be helpful for debugging models that run well on the reference Nengo backend but poorly on Nengo Loihi.
We have updated Nengo Loihi to work with the recently released NxSDK 0.7. We are currently not maintaining compatibility with older versions of NxSDK, so please upgrade it before upgrading Nengo Loihi.
Several bugs involving filtering connections and connecting between objects simulated on the host and on the chip have been found and fixed. Additionally, the tuning curves that many models use to solve for decoding weights are more accurate, meaning that decoded values will be
How do I get it?
If you are using the emulator, then you can simply use pip.
pip install nengo-loihi
If you are setting up Nengo Loihi to use Loihi hardware for the first time, see the Nengo Loihi installation instructions.
Where can I learn more?
- Nengo Loihi documentation
- Instructions for installing Nengo and Nengo Loihi
- An introduction to Nengo
- General Nengo documentation
Where can I get help?
You’re already there! If you have any questions about Nengo Loihi, please ask in the Loihi category. And if you run in to any bugs or have suggestions for new features, file an issue through Github.