ResidualConv¶
tiatoolbox
.models
.architecture
.nuclick
.ResidualConv
- class ResidualConv(num_input_channels, num_output_channels=32, kernel_size=(3, 3), strides=(1, 1), dilation_rate=(1, 1), *, use_bias=False)[source]¶
Residual Convolution block.
- Parameters:
num_input_channels (int) – Number of channels in input.
num_output_channels (int) – Number of channels in output.
kernel_size (int) – Size of the kernel in all convolution layers.
strides (int) – Size of the stride in all convolution layers.
use_bias (bool) – Whether to use bias in the convolution layers.
dilation_rate (int) – Dilation rate in all convolution layers.
- Returns:
A pytorch model.
- Return type:
model (torch.nn.Module)
Initialize
ResidualConv
.Methods
Logic for using layers defined in ResidualConv init.
Attributes
training
- forward(input_tensor)[source]¶
Logic for using layers defined in ResidualConv init.
This method defines how layers are used in forward operation.
- Parameters:
input_tensor (torch.Tensor) – Input, the tensor is of the shape NCHW.
self (ResidualConv)
- Returns:
The inference output.
- Return type:
output (torch.Tensor)