Conv1D.hpp File Reference

1D temporal convolution for FlexNN::Layers. More...

#include <utility>
#include <Eigen/Dense>
#include "LayerTypes.hpp"
#include "activations/Activation.hpp"
#include "activations/ActivationParameters.hpp"

Include dependency graph for Conv1D.hpp:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Classes
class	FlexNN::Layers::Conv1D

Namespaces
namespace	FlexNN
	Namespace for the FlexNN neural network library.
namespace	FlexNN::Layers

Detailed Description

1D temporal convolution for FlexNN::Layers.

Implements Conv1D as described in LLD_FLEXNN §6.2. The layer operates on flattened [C*L × batch] matrices where C is channels and L is temporal length. Forward is zero-padded, backward uses im2col/col2im for dW and dX. Dilation !=1 is rejected at export (Status error) but training still works (forward handles dilation).

Weight layout: W [outChannels × inChannels*K] row-major logical, stored in Eigen column-major MatrixXd but serialized row-major. Bias b [outChannels] broadcast over L_out.