From: Robert I. <ri...@gm...> - 2009-03-04 16:19:36
|
Hi, Thank you Dejan for clarifying what net.reset() does exactly. Actually what I am trying to do is present a fixed spike pattern stimulus to the network repeatedly while having plasticity enabled in the network. The only way I can think of to represent the fixed spike pattern is with neurons of SpikingInputNeuron class. The way I go about it now is: - reset the network - repeat N times - run simulation for 1 sec ( stimulus is 0.5 sec) - save the weights of all the plastic synapses - reset network - restore the weights of the plastic synapses Is there a more natural way from pypcsim's perspective to "train" a network with a fixed spiking pattern then saving weights and resetting the network? More precisely is there any functionality for declaring repetitive or cyclic stimuli? Thanks again for the prompt answers. Regards, Robert. |