Blocks.

Base blocks for Resnet - Basic and Bottleneck Blocks.

BasicBlock

block = BasicBlock(64, 64)
block

output

BasicBlock(
      (convs): Sequential(
        (conv_0): ConvBnAct(
          (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (bn): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (act_fn): ReLU(inplace=True)
        )
        (conv_1): ConvBnAct(
          (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (bn): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
      )
      (act_fn): ReLU(inplace=True)
    )

BottleneckBlock

block = BottleneckBlock(64, 64, dw=True)
block

output

BottleneckBlock(
      (convs): Sequential(
        (conv_0): ConvBnAct(
          (conv): Conv2d(64, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (bn): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (act_fn): ReLU(inplace=True)
        )
        (conv_1): ConvBnAct(
          (conv): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=16, bias=False)
          (bn): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (act_fn): ReLU(inplace=True)
        )
        (conv_2): ConvBnAct(
          (conv): Conv2d(16, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (bn): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
      )
      (act_fn): ReLU(inplace=True)
    )

block = BottleneckBlock(64, 64, groups=4)
block

output

BottleneckBlock(
      (convs): Sequential(
        (conv_0): ConvBnAct(
          (conv): Conv2d(64, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (bn): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (act_fn): ReLU(inplace=True)
        )
        (conv_1): ConvBnAct(
          (conv): Conv2d(16, 16, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), groups=4, bias=False)
          (bn): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (act_fn): ReLU(inplace=True)
        )
        (conv_2): ConvBnAct(
          (conv): Conv2d(16, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (bn): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
      )
      (act_fn): ReLU(inplace=True)
    )

block = BottleneckBlock(64, 64, expansion=2, act_fn=nn.LeakyReLU, bn_1st=False)
block

output

BottleneckBlock(
      (convs): Sequential(
        (conv_0): ConvBnAct(
          (conv): Conv2d(64, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (act_fn): LeakyReLU(negative_slope=0.01, inplace=True)
          (bn): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
        (conv_1): ConvBnAct(
          (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (act_fn): LeakyReLU(negative_slope=0.01, inplace=True)
          (bn): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
        (conv_2): ConvBnAct(
          (conv): Conv2d(32, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (bn): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
      )
      (act_fn): LeakyReLU(negative_slope=0.01, inplace=True)
    )

lock = BottleneckBlock(32, 64, expansion=2, dw=True)
block

output

BottleneckBlock(
      (convs): Sequential(
        (conv_0): ConvBnAct(
          (conv): Conv2d(64, 32, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (act_fn): LeakyReLU(negative_slope=0.01, inplace=True)
          (bn): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
        (conv_1): ConvBnAct(
          (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
          (act_fn): LeakyReLU(negative_slope=0.01, inplace=True)
          (bn): BatchNorm2d(32, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
        (conv_2): ConvBnAct(
          (conv): Conv2d(32, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (bn): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
      )
      (act_fn): LeakyReLU(negative_slope=0.01, inplace=True)
    )

pool = partial(nn.AvgPool2d, kernel_size=2, ceil_mode=True)

block = BottleneckBlock(32, 64, stride=2, dw=True, pool=pool)
block

output

BottleneckBlock(
      (convs): Sequential(
        (conv_0): ConvBnAct(
          (conv): Conv2d(32, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (bn): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (act_fn): ReLU(inplace=True)
        )
        (conv_1): ConvBnAct(
          (conv): Conv2d(16, 16, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=16, bias=False)
          (bn): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (act_fn): ReLU(inplace=True)
        )
        (conv_2): ConvBnAct(
          (conv): Conv2d(16, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (bn): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
      )
      (id_conv): Sequential(
        (pool): AvgPool2d(kernel_size=2, stride=2, padding=0)
        (id_conv): ConvBnAct(
          (conv): Conv2d(32, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (bn): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
      )
      (act_fn): ReLU(inplace=True)
    )

from model_constructor.layers import SEModule, SimpleSelfAttention

block = BottleneckBlock(32, 64, stride=2, dw=True, pool=pool, se=SEModule)
block

output

BottleneckBlock(
      (convs): Sequential(
        (conv_0): ConvBnAct(
          (conv): Conv2d(32, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (bn): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (act_fn): ReLU(inplace=True)
        )
        (conv_1): ConvBnAct(
          (conv): Conv2d(16, 16, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=16, bias=False)
          (bn): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (act_fn): ReLU(inplace=True)
        )
        (conv_2): ConvBnAct(
          (conv): Conv2d(16, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (bn): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
        (se): SEModule(
          (squeeze): AdaptiveAvgPool2d(output_size=1)
          (excitation): Sequential(
            (reduce): Linear(in_features=64, out_features=4, bias=True)
            (se_act): ReLU(inplace=True)
            (expand): Linear(in_features=4, out_features=64, bias=True)
            (se_gate): Sigmoid()
          )
        )
      )
      (id_conv): Sequential(
        (pool): AvgPool2d(kernel_size=2, stride=2, padding=0)
        (id_conv): ConvBnAct(
          (conv): Conv2d(32, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (bn): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
      )
      (act_fn): ReLU(inplace=True)
    )

block = BottleneckBlock(32, 64, stride=2, dw=True, pool=pool, se=SEModule, sa=SimpleSelfAttention)
block

output

BottleneckBlock(
      (convs): Sequential(
        (conv_0): ConvBnAct(
          (conv): Conv2d(32, 16, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (bn): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (act_fn): ReLU(inplace=True)
        )
        (conv_1): ConvBnAct(
          (conv): Conv2d(16, 16, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), groups=16, bias=False)
          (bn): BatchNorm2d(16, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
          (act_fn): ReLU(inplace=True)
        )
        (conv_2): ConvBnAct(
          (conv): Conv2d(16, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (bn): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
        (se): SEModule(
          (squeeze): AdaptiveAvgPool2d(output_size=1)
          (excitation): Sequential(
            (reduce): Linear(in_features=64, out_features=4, bias=True)
            (se_act): ReLU(inplace=True)
            (expand): Linear(in_features=4, out_features=64, bias=True)
            (se_gate): Sigmoid()
          )
        )
        (sa): SimpleSelfAttention(
          (conv): Conv1d(64, 64, kernel_size=(1,), stride=(1,), bias=False)
        )
      )
      (id_conv): Sequential(
        (pool): AvgPool2d(kernel_size=2, stride=2, padding=0)
        (id_conv): ConvBnAct(
          (conv): Conv2d(32, 64, kernel_size=(1, 1), stride=(1, 1), bias=False)
          (bn): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
        )
      )
      (act_fn): ReLU(inplace=True)
    )

model_constructor by ayasyrev