Convolution Layers

Convolutional layers apply learnable filters to input data, extracting local features through sliding window operations. They are fundamental to CNNs for image, audio, and sequential data processing.

Conv1d/2d/3d: Standard convolution for 1D sequences, 2D images, or 3D volumes
ConvTranspose1d/2d/3d: Transposed convolution (deconvolution) for upsampling

Key parameters:

in_channels: Number of input channels
out_channels: Number of output channels (number of filters)
kernel_size: Size of the convolving kernel
stride: Stride of the convolution (default: 1)
padding: Zero-padding added to input (default: 0)
dilation: Spacing between kernel elements (default: 1)
groups: Number of blocked connections (default: 1, use in_channels for depthwise)

Conv1d

Applies 1D convolution over an input signal composed of several input planes.

{doxygenclass}

:members:
:undoc-members:

{doxygenclass}

:members:
:undoc-members:

Conv2d

Applies 2D convolution over an input image. The most commonly used layer for image processing tasks.

{doxygenclass}

:members:
:undoc-members:

{doxygenclass}

:members:
:undoc-members:

Example:

cpp

// Create Conv2d: 3 input channels, 64 output channels, 3x3 kernel
auto conv = torch::nn::Conv2d(
    torch::nn::Conv2dOptions(3, 64, 3)
        .stride(1)
        .padding(1)
        .bias(true));

auto output = conv->forward(input);  // input: [N, 3, H, W]

Conv3d

Applies 3D convolution over an input volume (e.g., video frames or 3D medical images).

{doxygenclass}

:members:
:undoc-members:

{doxygenclass}

:members:
:undoc-members:

ConvTranspose1d

Applies 1D transposed convolution (fractionally-strided convolution) for upsampling.

{doxygenclass}

:members:
:undoc-members:

{doxygenclass}

:members:
:undoc-members:

ConvTranspose2d

Applies 2D transposed convolution for upsampling. Commonly used in decoder networks and generative models.

{doxygenclass}

:members:
:undoc-members:

{doxygenclass}

:members:
:undoc-members:

Example:

cpp

// Create ConvTranspose2d for upsampling
auto conv_transpose = torch::nn::ConvTranspose2d(
    torch::nn::ConvTranspose2dOptions(64, 32, 4)
        .stride(2)
        .padding(1));

ConvTranspose3d

Applies 3D transposed convolution for upsampling volumetric data.

{doxygenclass}

:members:
:undoc-members:

{doxygenclass}

:members:
:undoc-members: