fn
max_pool2d_backward_f64
void max_pool2d_backward_f64(const double * g, const int * argmax, double * dx, int B, int C, int H, int W, int OH, int OW)Double-precision counterpart to max_pool2d_backward_f32.
Parameters
gconst double*Upstream gradient of shape
(B, C, OH, OW).argmaxconst int32_t*Per-output winning
h*W + w indices.dxdouble*Output gradient of shape
(B, C, H, W); accumulated.B, C, H, W, OH, OWintLayout extents.