Caffe2 - C++ API
A deep learning, cross platform ML framework
Related Pages
Modules
Data Structures
Files
C++ API
Python API
GitHub
File List
Globals
torch
csrc
jit
passes
inline_fork_wait.h
1
#pragma once
2
3
#include <torch/csrc/jit/ir.h>
4
5
namespace
torch
{
6
namespace
jit {
7
8
// Inline Fork and Wait calls. This is used, for example, in ONNX export, where
9
// we do not support the explicit parallelism structures and would rather
10
// just have a flat graph. This inlines the forked section in the fork()
11
// callsite and replaces uses of the result of wait() calls with the values
12
// produced from the (now-inlined) forked section.
13
TORCH_API
void
InlineForkWait(
const
std::shared_ptr<Graph>& graph);
14
15
16
}
// namespace jit
17
}
// namespace torch
torch
Definition:
jit_type.h:17
Generated on Thu Mar 21 2019 13:06:23 for Caffe2 - C++ API by
1.8.11