Apache SINGA
A distributed deep learning platform .
|
unpack local (overlap) patches of image to column of mat, can be used to implement convolution, this expression allow unpack of a batch this is a version support unpacking multiple images after getting unpacked mat, we can use: output = dot( weight, mat ) to get covolved results, the relations: More...
#include <tensor_expr_ext.h>
Public Member Functions | |
UnpackPatchToColXExp (const SrcExp &img, index_t psize, index_t pstride) | |
constructor | |
Public Member Functions inherited from mshadow::expr::MakeTensorExp< UnpackPatchToColXExp< SrcExp, srcdim >, SrcExp, 2 > | |
const UnpackPatchToColXExp < SrcExp, srcdim > & | real_self (void) const |
true self of subtype | |
Public Member Functions inherited from mshadow::expr::Exp< MakeTensorExp< UnpackPatchToColXExp< SrcExp, srcdim >, SrcExp, dim >, type::kMapper > | |
const MakeTensorExp < UnpackPatchToColXExp< SrcExp, srcdim >, SrcExp, dim > & | self (void) const |
MakeTensorExp < UnpackPatchToColXExp< SrcExp, srcdim >, SrcExp, dim > & | refself (void) |
Public Attributes | |
const SrcExp & | img_ |
source operand | |
index_t | psize_ |
patch size | |
index_t | pstride_ |
patch stride | |
index_t | i_channel_ |
number of input channel | |
index_t | i_height_ |
height of img | |
index_t | i_width_ |
width of img | |
Public Attributes inherited from mshadow::expr::MakeTensorExp< UnpackPatchToColXExp< SrcExp, srcdim >, SrcExp, 2 > | |
Shape< dim > | shape_ |
the shape of this expression | |
unpack local (overlap) patches of image to column of mat, can be used to implement convolution, this expression allow unpack of a batch this is a version support unpacking multiple images after getting unpacked mat, we can use: output = dot( weight, mat ) to get covolved results, the relations:
SrcExp | source expression |
dstdim | destination dimension |