afp_convolve2_gradient_nn

Calculates backward pass gradient of 2D convolution. This function calculates the gradient with respect to the output of the afp_convolve2NN function that uses the machine learning formulation for the dimensions of the signals and filters.

Arguments

1. Gradient array, gradients to be distributed in backwards pass
2. Input signal array
3. Input filter array, filter to forward pass of convolution
4. Output array from forward pass of convolution
5. Integer, number of stride dimensions
6. Integer, number of padding dimensions
7. Integer, number of dilation dimensions
8. Integer, gradient to return: 1 - filter, 2 - data, 3 - bias

Output

Resulting array

Examples