ggoupy / spikingconvnet

Convolutional spiking neural network implementing STDP

SpikingConvNet

Implementation of the paper STDP-based spiking deep neural networks for object recognition, available here, for the MNIST classification task.

Results

The model achieves ~95% accuracy, better performance can be reached with more tuning.

MNIST Dataset

See https://pypi.org/project/python-mnist/ to download the dataset.

References:

[1] Kheradpisheh, S. R., Ganjtabesh, M., Thorpe, S. J., & Masquelier, T. (2018). STDP-based spiking deep convolutional neural networks for object recognition. Neural Networks, 99, 56–67. https://doi.org/10.1016/J.NEUNET.2017.12.005

[2] Mozafari, M., Ganjtabesh, M., Nowzari-Dalini, A., & Masquelier, T. (2019). SpykeTorch: Efficient simulation of convolutional spiking neural networks with at most one spike per neuron. Frontiers in Neuroscience, 13, 625. https://doi.org/10.3389/FNINS.2019.00625

[3] https://github.com/npvoid/SDNN_python

spikingconvnet's Issues

Some problems and how to solve them.

First, you need to install python-mnist before you start.
You can use the following commands
pip install python-mnist

Then in utils.py at line 101.
mndata = MNIST()
You need to add the path to the folder containing the MNIST file like this
mndata = MNIST("your_files_folder_path")

The MNIST files contain t10k-images-idx3-ubyte, t10k-labels-idx1-ubyte, train-images-idx3-ubyte, train-labels-idx1-ubyte.