Skip to content
体验新版
项目
组织
正在加载...
登录
切换导航
打开侧边栏
magicwindyyd
mindspore
提交
37250625
M
mindspore
项目概览
magicwindyyd
/
mindspore
与 Fork 源项目一致
Fork自
MindSpore / mindspore
通知
1
Star
1
Fork
0
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
0
列表
看板
标记
里程碑
合并请求
0
Wiki
0
Wiki
分析
仓库
DevOps
项目成员
Pages
M
mindspore
项目概览
项目概览
详情
发布
仓库
仓库
文件
提交
分支
标签
贡献者
分支图
比较
Issue
0
Issue
0
列表
看板
标记
里程碑
合并请求
0
合并请求
0
Pages
分析
分析
仓库分析
DevOps
Wiki
0
Wiki
成员
成员
收起侧边栏
关闭侧边栏
动态
分支图
创建新Issue
提交
Issue看板
提交
37250625
编写于
8月 27, 2020
作者:
M
mindspore-ci-bot
提交者:
Gitee
8月 27, 2020
浏览文件
操作
浏览文件
下载
差异文件
!5229 [AutoParallel]Fix CodeDex
Merge pull request !5229 from lichen/fix_code_index
上级
a15b9beb
49aa4b76
变更
4
隐藏空白更改
内联
并排
Showing
4 changed file
with
3 addition
and
5 deletion
+3
-5
mindspore/ccsrc/frontend/parallel/auto_parallel/rec_core/rec_cost.cc
...csrc/frontend/parallel/auto_parallel/rec_core/rec_cost.cc
+0
-1
mindspore/ccsrc/frontend/parallel/auto_parallel/rec_core/rec_generate_strategy.cc
.../parallel/auto_parallel/rec_core/rec_generate_strategy.cc
+0
-2
mindspore/ccsrc/frontend/parallel/auto_parallel/rec_core/rec_partition.cc
...frontend/parallel/auto_parallel/rec_core/rec_partition.cc
+3
-1
mindspore/ccsrc/frontend/parallel/step_parallel.cc
mindspore/ccsrc/frontend/parallel/step_parallel.cc
+0
-1
未找到文件。
mindspore/ccsrc/frontend/parallel/auto_parallel/rec_core/rec_cost.cc
浏览文件 @
37250625
...
...
@@ -26,7 +26,6 @@
namespace
mindspore
{
namespace
parallel
{
// Compute redistributed cost
double
CostRedis
(
const
Graph
::
NodeType
&
node
,
const
std
::
vector
<
std
::
pair
<
std
::
string
,
StrategyRec
>>
&
node_name_to_strategy
,
...
...
mindspore/ccsrc/frontend/parallel/auto_parallel/rec_core/rec_generate_strategy.cc
浏览文件 @
37250625
...
...
@@ -746,7 +746,6 @@ Strategys CheckBroadcast(const std::vector<std::shared_ptr<OperatorInfo>> &ops,
size_t
first_tensor_dim
=
ops
[
iter_ops
]
->
inputs_tensor_info
()[
0
].
shape
().
size
();
size_t
second_tensor_dim
=
ops
[
iter_ops
]
->
inputs_tensor_info
()[
1
].
shape
().
size
();
// Do Broadcasting in the second tensor.
if
(
second_tensor_dim
<
first_tensor_dim
)
{
bool
braoadcast_first_tensor
=
false
;
...
...
@@ -964,7 +963,6 @@ void GenerateEliminatedOperatorStrategyBackward(const std::vector<std::shared_pt
auto
iter_ops
=
no_stra_op_list
->
at
(
iter_list
-
1
);
Strategys
stra
;
Dimensions
s
=
CopyOutgoingOperatorInputStrategy
(
ops
,
input_tensor_names
,
iter_ops
);
if
(
s
.
size
()
!=
0
&&
ops
[
iter_ops
]
->
type
()
==
SQUEEZE
)
{
s
=
ModifyStrategyIfSqueezeOutgoing
(
ops
,
iter_ops
,
s
);
}
...
...
mindspore/ccsrc/frontend/parallel/auto_parallel/rec_core/rec_partition.cc
浏览文件 @
37250625
...
...
@@ -204,7 +204,9 @@ Status PartitionForAllDevices(const size_t num_device, const double device_memor
// Comopute iter times
int
iter_times
=
static_cast
<
int
>
(
log2
(
num_device
));
if
(
iter_times
>
10
)
{
MS_LOG
(
EXCEPTION
)
<<
"ERROR: Number of iter_times can't be larger than 10."
;
}
// N-cuts loop
for
(
int
loop
=
0
;
loop
<
iter_times
;
loop
++
)
{
// Sort by weights
...
...
mindspore/ccsrc/frontend/parallel/step_parallel.cc
浏览文件 @
37250625
...
...
@@ -341,7 +341,6 @@ void Redistribution(const std::pair<AnfNodePtr, int> &node_pair, const OperatorI
TensorInfo
tensorinfo_out
=
next_distribute_operator
->
inputs_tensor_info
()[
IntToSize
(
index
-
1
)];
TensorLayout
tensorlayout_out
=
tensorinfo_out
.
tensor_layout
();
TensorLayout
tensorlayout_in
=
GetTensorInLayout
(
middle_node
,
middle_prim
,
distribute_operator
);
if
(
tensor_redistribution
.
Init
(
tensorlayout_in
,
tensorlayout_out
,
dev_list
)
==
FAILED
)
{
MS_LOG
(
ERROR
)
<<
"Redistribution: middle_prim "
<<
middle_prim
->
name
()
<<
" next_prim : "
<<
next_prim_name
;
MS_LOG
(
ERROR
)
<<
"Redistribution: middle_node "
<<
middle_node
->
ToString
()
<<
" next_node "
...
...
编辑
预览
Markdown
is supported
0%
请重试
或
添加新附件
.
添加附件
取消
You are about to add
0
people
to the discussion. Proceed with caution.
先完成此消息的编辑!
取消
想要评论请
注册
或
登录