rockpool.nn.layers.updown

Feedforward layer that converts each analogue input channel to one spiking up and one down channel

Runs in batch mode like UpDownTorch to save memory, but does not use pytorch. UpDownTorch seems to be slower…

Classes

FFUpDown(*args, **kwargs)	Define a spiking feedforward layer to convert analogue inputs to up and down channels
FFUpDownV1(weights[, repeat_output, dt, ...])	Define a spiking feedforward layer to convert analogue inputs to up and down channels