spandrel.architectures.__arch_helpers.block

module documentation
(source)

Undocumented

Class ConcatBlock Undocumented
Class ResidualDenseBlock_5C Residual Dense Block Style: 5 convs The core module of paper: (Residual Dense Network for Image Super-Resolution, CVPR 18) Modified options that can be used:
Class ResNetBlock ResNet Block, 3-3 style with extra residual scaling used in EDSR (Enhanced Deep Residual Networks for Single Image Super-Resolution, CVPRW 17)
Class RRDB Residual in Residual Dense Block (ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks)
Class ShortcutBlock Undocumented
Class ShortcutBlockSPSR Undocumented
Function act Undocumented
Function conv1x1 Undocumented
Function conv_block Conv layer with padding, normalization, activation mode: - CNA --> Conv -> Norm -> Act - NAC --> Norm -> Act --> Conv (Identity Mappings in Deep Residual Networks, ECCV16)
Function conv_block_2c2 Undocumented
Function get_valid_padding Undocumented
Function norm Undocumented
Function pad Undocumented
Function pixelshuffle_block Pixel shuffle layer (Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network, CVPR17)
Function sequential Undocumented
Function upconv_block Undocumented
Type Alias ConvMode Undocumented
def act(act_type: str, inplace=True, neg_slope=0.2, n_prelu=1): (source)

Undocumented

def conv1x1(in_planes, out_planes, stride=1): (source)

Undocumented

def conv_block(in_nc: int, out_nc: int, kernel_size, stride=1, dilation=1, groups=1, bias=True, pad_type='zero', norm_type: str|None = None, act_type: str|None = 'relu', mode: ConvMode = 'CNA'): (source)

Conv layer with padding, normalization, activation mode: - CNA --> Conv -> Norm -> Act - NAC --> Norm -> Act --> Conv (Identity Mappings in Deep Residual Networks, ECCV16)

def conv_block_2c2(in_nc, out_nc, act_type='relu'): (source)

Undocumented

def get_valid_padding(kernel_size, dilation): (source)

Undocumented

def norm(norm_type: str, nc: int): (source)

Undocumented

def pad(pad_type: str, padding): (source)

Undocumented

def pixelshuffle_block(in_nc: int, out_nc: int, upscale_factor=2, kernel_size=3, stride=1, bias=True, pad_type='zero', norm_type: str|None = None, act_type='relu'): (source)

Pixel shuffle layer (Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network, CVPR17)

def sequential(*args): (source)

Undocumented

def upconv_block(in_nc: int, out_nc: int, upscale_factor=2, kernel_size=3, stride=1, bias=True, pad_type='zero', norm_type: str|None = None, act_type='relu', mode='nearest'): (source)

Undocumented

ConvMode = (source)

Undocumented

Value
Literal['CNA', 'NAC', 'CNAC']