snntorch.functional.loss.ce_rate_loss

[Maya7991]

Hi,

I am working on a network which uses both CNN and SCNN as intermediate layers. I want to know if can use snntorch.functional.loss.ce_rate_loss to train the network which uses CNN layer also. Because the output here will not be spikes but it will have num_steps as a dimension which could be handled well by snntorch.functional.loss.ce_rate_loss.
So my question is does this loss function only work with spiking output?

I tried going through the code of ce_rate_loss but still have this doubt :(

Thanks,
Maya