GitHub - dtyulman/ModNN: Modular (feedforward) neural network. Cluster the inputs to the network, then train each module in the network on each cluster

Name		Name	Last commit message	Last commit date
Latest commit History 6 Commits
.gitignore		.gitignore
activateSubnet.m		activateSubnet.m
classifyClustNN.m		classifyClustNN.m
drawMNISTdigit.m		drawMNISTdigit.m
errorNN.m		errorNN.m
feedforward.m		feedforward.m
gmeans.m		gmeans.m
initKmeans.m		initKmeans.m
initNN.m		initNN.m
initseed.mat		initseed.mat
initseedpreproc.mat		initseedpreproc.mat
loadMNIST.m		loadMNIST.m
loss.m		loss.m
mnist2nn.m		mnist2nn.m
mnist2nn.mat		mnist2nn.mat
onehot2class.m		onehot2class.m
packNN.m		packNN.m
plot_step_hist.m		plot_step_hist.m
predictClustNN.m		predictClustNN.m
risk.m		risk.m
script_analyzeClustering.m		script_analyzeClustering.m
script_repeat_trainClustNN.m		script_repeat_trainClustNN.m
script_trainClustNN.m		script_trainClustNN.m
softmax.m		softmax.m
trainClustNN.m		trainClustNN.m
unpackNN.m		unpackNN.m
weightsCorr.m		weightsCorr.m
zeropadSubnet.m		zeropadSubnet.m

About

Modular (feedforward) neural network. Cluster the inputs to the network, then train each module in the network on each cluster

0 stars

1 watching

No releases published