Note

This documentation is for a development version. Click here for the latest stable release (v0.1.0).

Latest PyPI version Travis-CI build status Test coverage

PyTorchSpiking

PyTorchSpiking provides tools for training and running spiking neural networks directly within the PyTorch framework. The main feature is pytorch_spiking.SpikingActivation, which can be used to transform any activation function into a spiking equivalent. For example, we can translate a non-spiking model, such as

torch.nn.Sequential(
    torch.nn.Linear(5, 10),
    torch.nn.ReLU(),
)

into the spiking equivalent:

torch.nn.Sequential(
    torch.nn.Linear(5, 10),
    pytorch_spiking.SpikingActivation(torch.nn.ReLU()),
)

Models with SpikingActivation layers can be optimized and evaluated in the same way as any other PyTorch model. They will automatically take advantage of PyTorchSpiking’s “spiking aware training”: using the spiking activations on the forward pass and the non-spiking (differentiable) activation function on the backwards pass.

PyTorchSpiking also includes various tools to assist in the training of spiking models, such as filtering layers.

If you are interested in building and optimizing spiking neuron models, you may also be interested in NengoDL. See this page for a comparison of the different use cases supported by these two packages.