Quick question about learning rates...

[rafaelblevin821]

So nu accepts a pair of learning rates for both the pre-synaptic and post-synaptic weight updates in the PostPre learning rule. If only one learning rate is defined, does the network use this value for both pre-synaptic and post-synaptic weight calculations, or does it only use this only for the pre-synaptic weight calculation and no learning rate multiplier for the post-synaptic weight calculation?

[Hananel-Hazan]

From the code documentation

:param nu: Single or pair of learning rates for pre- and post-synaptic events. It also accepts a pair of tensors to individualize learning rates of each neuron. In this case, their shape should be the same size as the connection weights.

Looking on the code here if a nu is set with single int or float, then both pre and post has the same value.

[rafaelblevin821]

@Hananel-Hazan Thank you