mxnet
|
The Expr and related elements in DataFlow construction. More...
#include <tvm/expr.h>
#include <tvm/tensor.h>
#include <tvm/schedule.h>
#include <tvm/packed_func_ext.h>
#include <tvm/runtime/registry.h>
#include <nnvm/op_attr_types.h>
#include <nnvm/graph_attr_types.h>
#include <nnvm/graph.h>
#include <vector>
#include <string>
#include "packed_func_ext.h"
Go to the source code of this file.
Namespaces | |
nnvm | |
nnvm::compiler | |
Typedefs | |
using | nnvm::compiler::TOpPattern = int |
the operator pattern More... | |
using | nnvm::compiler::FTVMCompute = std::function< Array< Tensor >(const NodeAttrs &attrs, const Array< Tensor > &inputs, const Array< Tensor > &out_info)> |
Computation description interface. More... | |
using | nnvm::compiler::FTVMSchedule = std::function< Schedule(const NodeAttrs &attrs, const Array< Tensor > &outs, const std::string &target)> |
Build the computation schedule for op whose root is at current op. More... | |
using | nnvm::compiler::FTVMAlterOpLayout = std::function< bool(const NodeAttrs &attrs, const Symbol &inputs, const Array< Tensor > &tinfos, Symbol *ret)> |
Modify the op node to alter its input layout. it is invoked in AlterOpLayout pass. More... | |
using | nnvm::compiler::FTVMVectorizedOp = std::function< nnvm::NodePtr(const nnvm::Node *node)> |
Transform from normal operator to vectorized operator. More... | |
Enumerations | |
enum | nnvm::compiler::OpPatternKind { nnvm::compiler::kElemWise = 0, nnvm::compiler::kBroadcast = 1, nnvm::compiler::kInjective = 2, nnvm::compiler::kCommReduce = 3, nnvm::compiler::kOutEWiseFusable = 4, nnvm::compiler::kOpaque = 8 } |
operator pattern used in graph fusion More... | |
The Expr and related elements in DataFlow construction.
Copyright (c) 2017 by Contributors