hxtorch.spiking.ConversionConfig

class hxtorch.spiking.ConversionConfig(dt: float = 1e-06, calib_path: str = None, weight_scale: float = 64.0, trace_scale: float = 0.02, trace_shift: float = 0.0, cadc_recording: bool = True, mock: bool = True, input_loopback: bool = True, device: torch.device = <factory>)

Bases: object

Configuration for the conversion of NIRGraph to hxtorch SNN.

Some parameters are a dict with node names as keys such that the parametrization for each layer can be set individually.

__init__(dt: float = 1e-06, calib_path: str = None, weight_scale: float = 64.0, trace_scale: float = 0.02, trace_shift: float = 0.0, cadc_recording: bool = True, mock: bool = True, input_loopback: bool = True, device: torch.device = <factory>) → None: Initialize self. See help(type(self)) for accurate signature.

Methods

__init__([dt, calib_path, weight_scale, …])

Initialize self.

Attributes

`cadc_recording`
`calib_path`
`dt`
`input_loopback`
`mock`
`trace_scale`
`trace_shift`
`weight_scale`

cadc_recording: bool = True

calib_path: str = None

device: torch.device

dt: float = 1e-06

input_loopback: bool = True

mock: bool = True

trace_scale: float = 0.02

trace_shift: float = 0.0

weight_scale: float = 64.0