SNN_starter

Overview

SNN moudle rebuild of A Low Power, Fully Event-Based Gesture Recognition System using yjwu17/BP-for-SpikingNN.

Download the dataset from here and extract as dataset.

Raw dataset must be formatted before training and it may spend some time. You can download the formatted data from here for the first training.

Ensure that all files (eg. *.adept, *.csv, trails_to_train.txt and trails_to_test.txt) are NOT in any directory other than ./dataset.

python train.py

Change preload = True to use formatted data, data is saved at root/*.data

Change the trials_to_train.txt and trials_to_test.txt to control the size of training and test dataset.

Name		Name	Last commit message	Last commit date
Latest commit History 17 Commits
checkpoint		checkpoint
README.md		README.md
dataset_test.py		dataset_test.py
self_spiking_model.py		self_spiking_model.py
snn_dataset.py		snn_dataset.py
test.py		test.py
train.ipynb		train.ipynb
train.py		train.py