Question: Weights Decreasing using PostPre Learning Rule

[tasha061]

Hi,
I am using the PostPre learning rule in BindsNet and I am getting good accuracy from my SNN network. However, my weights seem to be decreasing instead of increasing is this normal or would you know a reason why this would be happening?

Many thanks.

[SimonInParis]

Is the average of the weights decreasing? or only some of them?
Did you put a constraint on the norm of the layer's weights?
Keep in mind that increase/decrease of the w is also related to the amount of incoming/outgoing spikes. You can check that too.