neuromancer.modules.functions module

Set of useful function transformations

neuromancer.modules.functions.bounds_clamp(x, xmin=None, xmax=None)

hard bounds on variable x via ReLU clamping between xmin and xmax values :param x: :param xmin: :param xmax: :return:

neuromancer.modules.functions.bounds_scaling(x, xmin, xmax, scaling=1.0)

hard bounds on variable x via sigmoid scaling between xmin and xmax values :param x: :param xmin: :param xmax: :param scaling: :return:

neuromancer.modules.functions.window_functions(x, num_domains, delta=1.9)

Window functions for finite-basis domain decomposition. :param x: input tensor representing the domain. :param num_domains: number of domains. Must be a perfect square. :param delta: overlapping ratio. Higher = more overlapping. :return w/(s+eps): weighted window functions.