def __init__(self, permutation, dim=1):
if permutation.dim() != 1:
raise ValueError("Permutation must be a 1D tensor.")
if not is_positive_int(dim):
raise ValueError("dim must be a positive integer.")
super().__init__()
self._dim = dim
self.register_buffer("_permutation", permutation)
def __init__(self, features):
"""
Transform that performs activation normalization. Works for 2D and 4D inputs. For 4D
inputs (images) normalization is performed per-channel, assuming BxCxHxW input shape.
Reference:
> D. Kingma et. al., Glow: Generative flow with invertible 1x1 convolutions, NeurIPS 2018.
"""
if not is_positive_int(features):
raise TypeError("Number of features must be a positive integer.")
super().__init__()
self.initialized = False
self.log_scale = nn.Parameter(torch.zeros(features))
self.shift = nn.Parameter(torch.zeros(features))
def __init__(self, in_channels, dim=1):
if not is_positive_int(in_channels):
raise ValueError("Number of features must be a positive integer.")
super().__init__(torch.arange(in_channels - 1, -1, -1), dim)
def __init__(self, in_channels: int, dim: int = 1) -> None:
if not is_positive_int(in_channels):
raise ValueError("Number of features must be a positive integer.")
super().__init__(torch.randperm(in_channels), dim)