im2col_3d_f32

void im2col_3d_f32(const float * x, float * cols, int C, int D, int H, int W, int KD, int KH, int KW, int OD, int OH, int OW, int stride_d, int stride_h, int stride_w, int pad_d, int pad_h, int pad_w, int dilation_d, int dilation_h, int dilation_w)

source edit

Single-precision 3-D im2col unfold.

Unfolds a volumetric feature map $x$ of shape $(C, D, H, W)$ into a column matrix of shape $(C \cdot K_D \cdot K_H \cdot K_W, O_D \cdot O_H \cdot O_W)$ . Used by conv3d (e.g., for video / medical-imaging models). Out-of-bounds positions are zero-padded.

Parameters

xconst float*

Input tensor, shape

(C, D, H, W)

row-major.

colsfloat*

Output column matrix.

Cint

Channel count.

D, H, Wint

Input depth / height / width.

KD, KH, KWint

Kernel depth / height / width.

OD, OH, OWint

Output depth / height / width.

stride_d, stride_h, stride_wint

Per-axis strides.

pad_d, pad_h, pad_wint

Per-axis zero-padding.

dilation_d, dilation_h, dilation_wint

Per-axis dilation.

Notes

Memory layout is channel-first (NCDHW); batch dimension is iterated by the caller.

void im2col_3d_f32(const float * x, float * cols, int C, int D, int H, int W, int KD, int KH, int KW, int OD, int OH, int OW, int stride_d, int stride_h, int stride_w, int pad_d, int pad_h, int pad_w, int dilation_d, int dilation_h, int dilation_w)