inferno.functional

Protocols

HalfBounding	Callable used to apply bounding to the lower or upper limit of a parameter.
FullBounding	Callable used to apply bounding to the lower and upper limit of a parameter.
Interpolation	Callable used to interpolate in time between two tensors.
Extrapolation	Callable used to extrapolate in time to two tensors.
DimensionReduction	Callable used to reduce the dimensions of a tensor.
SpikeTimeHalfKernel	Callable used for computing the presynaptic or postsynaptic update of pre-post training methods with spike time difference.

Bounding

bound_power	Computes the scaled update of power parameter dependence.
bound_upper_power	Computes the scaled update of upper-bound power parameter dependence.
bound_lower_power	Computes the scaled update of lower-bound power parameter dependence.
bound_scaled_power	Computes the scaled update of scaled power parameter dependence.
bound_upper_scaled_power	Computes the scaled update of upper-bound scaled power parameter dependence.
bound_lower_scaled_power	Computes the scaled update of lower-bound scaled power parameter dependence.
bound_multiplicative	Computes the scaled update of multiplicative parameter dependence.
bound_upper_multiplicative	Computes the scaled update of upper-bound multiplicative parameter dependence.
bound_lower_multiplicative	Computes the scaled update of lower-bound multiplicative parameter dependence.
bound_scaled_multiplicative	Computes the scaled update of multiplicative parameter dependence.
bound_upper_scaled_multiplicative	Computes the scaled update of upper-bound scaled multiplicative parameter dependence.
bound_lower_scaled_multiplicative	Computes the scaled update of lower-bound scaled multiplicative parameter dependence.
bound_sharp	Computes the scaled update of sharp parameter dependence.
bound_upper_sharp	Computes the scaled update of upper-bound sharp parameter dependence.
bound_lower_sharp	Computes the scaled update of lower-bound sharp parameter dependence.

Interpolation

interp_previous	Interpolates by selecting the previous state.
interp_next	Interpolates by selecting the next state.
interp_nearest	Interpolates by selecting the nearest state.
interp_linear	Interpolates between previous and next states linearlly.
interp_expdecay	Interpolates by exponentially decaying value from previous state, parameterized by a time constant.
interp_expratedecay	Interpolates by exponentially decaying value from previous state, parameterized by a rate constant.

Extrapolation

extrap_previous	Extrapolates out to the previous state.
extrap_next	Extrapolates out to the next state.
extrap_neighbors	Extrapolates out to the neighboring states.
extrap_nearest	Extrapolates out to the closest neighbor.
extrap_linear_forward	Extrapolates out linearly to the next state.
extrap_linear_backward	Extrapolates out linearly to the previous state.
extrap_expdecay	Extrapolates out assuming exponential decay dynamics, parameterized by a time constant.
extrap_expratedecay	Extrapolates out assuming exponential decay dynamics, parameterized by a rate constant.

Dimension Reductions

sum	Returns a tensor with dimensions reduced via summation.
nansum	Returns a tensor with dimensions reduced via summation, excluding NaN values.
divsum	Returns a tensor with dimensions reduced via summation then divided by a constant.
nandivsum	Returns a tensor with dimensions reduced via summation then divided by a constant, excluding NaN values.
min	Returns a tensor with dimensions reduced by taking the minimum.
max	Returns a tensor with dimensions reduced by taking the maximum.
absmin	Returns a tensor with dimensions reduced by taking the minimum absolute distance from zero.
absmax	Returns a tensor with dimensions reduced by taking the maximum absolute distance from zero.
mean	Returns a tensor with dimensions reduced by taking the arithmetic mean.
nanmean	Returns a tensor with dimensions reduced by taking the arithmetic mean, excluding NaN values.
median	Returns a tensor with dimensions reduced by taking median.
nanmedian	Returns a tensor with dimensions reduced by taking median, excluding NaN values.
geomean	Returns a tensor with dimensions reduced by taking the geometric mean.
nangeomean	Returns a tensor with dimensions reduced by taking the geometric mean, excluding NaN values.
quantile	Returns a tensor with dimensions reduced by taking the arithmetic mean.
nanquantile	Returns a tensor with dimensions reduced by taking the arithmetic mean, excluding NaN values.

Spike Time Kernels

exp_stdp_post_kernel	Returns the postsynaptic update for exponential spike-timing dependent plasticity.
exp_stdp_pre_kernel	Returns the presynaptic update for exponential spike-timing dependent plasticity.