Spikeometric is a Pytorch Geometric based framework for simulating spiking neural networks (SNNs) using Generalized Linear Models and Linear-Nonlinear-Poisson models.
Project description
Spikeometric - GLM-based Spiking Neural Networks with PyTorch Geometric
This package provides a simple and scaleable way to simulate networks of neurons using either Linear-Nonlinear-Poisson models (LNP) or its cousin the Generalized Linear Model (GLM).
The framework is built on top of torch modules and let's you tune parameters in your model to match a certain firing rate, provided the model is differentiable.
One key application is the problem of infering connectivity from spike data, where these models are often used both as generative and inference models.
Install
Before installing spikeometric
you will need to download versions of PyTorch and PyTorch Geometric that work with your hardware. When you have done that (for example in a conda environment), you are ready to download spikeometric with:
pip install spikeometric
Documentation
For more information about the package and a full API reference check out our documentation.
Project details
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 spikeometric-0.1.0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | da416b6f5bf09d45cdba3293bea8678422022ab1602dd1db556c523c5778c3ad |
|
MD5 | 850e311c698bdda12e586bfefa7c38ee |
|
BLAKE2b-256 | 1acd0f996f03a3b322671dbbcb41cbe60420e4b86dbb81a1792e41f7aab889ba |