Closed accosmin closed 9 years ago
Implement and check if a custom 3x3, 5x5 ... kernel implementation is faster than conv2d_dyn or corr2d_dyn.
These convolutions are not faster in the end than conv2d_dyn, which is doing more or less something similar.
Implement and check if a custom 3x3, 5x5 ... kernel implementation is faster than conv2d_dyn or corr2d_dyn.