A deep learning library for spiking neural networks which is based on PyTorch, focuses on fast training and supports inference on neuromorphic hardware.
Numpy API probably changed in the recent times and adding an integer and a numpy.uint64 creates a numpy.float64 object which is incompatible with the events timestamps we are sending to the device.
We use this in the chip_factory methods xytp_to_events and raster_to_events when we are converting structured array events to samna device event type. The fix for that is to cast the integer directly to numpy.uint64.
Numpy API probably changed in the recent times and adding an integer and a
numpy.uint64creates anumpy.float64object which is incompatible with the events timestamps we are sending to the device. We use this in thechip_factorymethodsxytp_to_eventsandraster_to_eventswhen we are converting structured array events to samna device event type. The fix for that is to cast the integer directly tonumpy.uint64.