spandrel.architectures.SPAN.SPAN

class documentation

class SPAN(nn.Module): (source)

Constructor: SPAN(num_in_ch, num_out_ch, feature_channels, upscale, ...)

View In Hierarchy

Swift Parameter-free Attention Network for Efficient Super-Resolution

Method __init__ Undocumented
Method forward Undocumented
Class Variable hyperparameters Undocumented
Instance Variable block_1 Undocumented
Instance Variable block_2 Undocumented
Instance Variable block_3 Undocumented
Instance Variable block_4 Undocumented
Instance Variable block_5 Undocumented
Instance Variable block_6 Undocumented
Instance Variable conv_1 Undocumented
Instance Variable conv_2 Undocumented
Instance Variable conv_cat Undocumented
Instance Variable img_range Undocumented
Instance Variable in_channels Undocumented
Instance Variable mean Undocumented
Instance Variable no_norm Undocumented
Instance Variable out_channels Undocumented
Instance Variable upsampler Undocumented
Property is_norm Undocumented
def __init__(self, *, num_in_ch: int, num_out_ch: int, feature_channels=48, upscale=4, bias=True, norm=True, img_range=255.0, rgb_mean=(0.4488, 0.4371, 0.404)): (source)

Undocumented

def forward(self, x): (source)

Undocumented

hyperparameters: dict = (source)

Undocumented

block_1 = (source)

Undocumented

block_2 = (source)

Undocumented

block_3 = (source)

Undocumented

block_4 = (source)

Undocumented

block_5 = (source)

Undocumented

block_6 = (source)

Undocumented

conv_1 = (source)

Undocumented

conv_2 = (source)

Undocumented

conv_cat = (source)

Undocumented

img_range = (source)

Undocumented

in_channels = (source)

Undocumented

mean = (source)

Undocumented

no_norm: torch.Tensor|None = (source)

Undocumented

out_channels = (source)

Undocumented

upsampler = (source)

Undocumented

@property
is_norm = (source)

Undocumented