mxnet
|
support for broadcast and repmat More...
#include "../extension.h"
Go to the source code of this file.
Classes | |
struct | mshadow::expr::Broadcast1DExp< SrcExp, DType, dimdst, dimdst_m_cast > |
broadcast Tensor1D into a higher dimension Tensor input: Tensor<Device,1>: ishape[0] output: Tensor<Device,dimdst> : oshape[dimcast] = ishape[0] More... | |
struct | mshadow::expr::BroadcastScalarExp< SrcExp, DType, dimdst > |
broadcast scalar into a higher dimension Tensor input: Tensor<Device,1>: ishape = {1} output: Tensor<Device, dimdst> : oshape[dimcast] = ishape[0] More... | |
struct | mshadow::expr::Plan< Broadcast1DExp< SrcExp, DType, dimdst, dimdst_m_cast >, DType > |
struct | mshadow::expr::Plan< Broadcast1DExp< SrcExp, DType, dimdst, 1 >, DType > |
execution plan of Broadcast1DExp More... | |
struct | mshadow::expr::Plan< BroadcastScalarExp< SrcExp, DType, dimdst >, DType > |
execution plan of Broadcast1DExp More... | |
Namespaces | |
mshadow | |
namespace for mshadow | |
mshadow::expr | |
namespace for abstract expressions and expressions template, have no dependency on tensor.h, These data structure takes no charge in computations, they are only used to define operations and represent expression in a symbolic way | |
Functions | |
template<int dimcast, typename SrcExp , typename DType , int etype, int dimdst> | |
Broadcast1DExp< SrcExp, DType, dimdst, dimdst-dimcast > | mshadow::expr::broadcast (const expr::Exp< SrcExp, DType, etype > &src, Shape< dimdst > shape) |
a expression that replicate a 1 dimension tensor in dimension dimcast More... | |
template<typename SrcExp , typename DType , int etype, int dimdst> | |
BroadcastScalarExp< SrcExp, DType, dimdst > | mshadow::expr::broadcast_scalar (const expr::Exp< SrcExp, DType, etype > &src, Shape< dimdst > shape) |
a expression that replicate a scalar tensor to target dimension. More... | |
template<typename SrcExp , typename DType , int etype> | |
Broadcast1DExp< SrcExp, DType, 2, 1 > | mshadow::expr::repmat (const expr::Exp< SrcExp, DType, etype > &src, index_t nrow) |
a expression that replicate a 1 dimension tensor for nrow times More... | |
support for broadcast and repmat
Copyright (c) 2014 by Contributors