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
夏有凉风,冬有雪
Paddle
提交
e249ad12
P
Paddle
项目概览
夏有凉风,冬有雪
/
Paddle
与 Fork 源项目一致
Fork自
PaddlePaddle / Paddle
通知
1
Star
1
Fork
0
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
1
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
P
Paddle
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
1
Issue
1
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
体验新版 GitCode,发现更多精彩内容 >>
未验证
提交
e249ad12
编写于
1月 09, 2018
作者:
Y
Yancey
提交者:
GitHub
1月 09, 2018
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
Test dist word2vec (#7334)
* test dist word2vec * multiple trainers work
上级
b5fda272
变更
5
隐藏空白更改
内联
并排
Showing
5 changed file
with
115 addition
and
6 deletion
+115
-6
paddle/framework/executor.cc
paddle/framework/executor.cc
+1
-1
paddle/framework/executor.h
paddle/framework/executor.h
+0
-2
paddle/operators/recv_op.cc
paddle/operators/recv_op.cc
+17
-3
paddle/operators/sum_op.h
paddle/operators/sum_op.h
+1
-0
python/paddle/v2/fluid/tests/book_distribute/test_dist_word2vec.py
...ddle/v2/fluid/tests/book_distribute/test_dist_word2vec.py
+96
-0
未找到文件。
paddle/framework/executor.cc
浏览文件 @
e249ad12
...
...
@@ -35,7 +35,7 @@ const std::string kFetchOpType = "fetch";
Executor
::
Executor
(
const
platform
::
Place
&
place
)
:
place_
(
place
)
{}
void
CreateTensor
(
Variable
*
var
,
proto
::
VarDesc
::
VarType
var_type
)
{
static
void
CreateTensor
(
Variable
*
var
,
proto
::
VarDesc
::
VarType
var_type
)
{
if
(
var_type
==
proto
::
VarDesc
::
LOD_TENSOR
)
{
var
->
GetMutable
<
LoDTensor
>
();
}
else
if
(
var_type
==
proto
::
VarDesc
::
SELECTED_ROWS
)
{
...
...
paddle/framework/executor.h
浏览文件 @
e249ad12
...
...
@@ -45,7 +45,5 @@ class Executor {
const
platform
::
Place
place_
;
};
void
CreateTensor
(
Variable
*
var
,
proto
::
VarDesc
::
VarType
var_type
);
}
// namespace framework
}
// namespace paddle
paddle/operators/recv_op.cc
浏览文件 @
e249ad12
...
...
@@ -32,6 +32,20 @@ limitations under the License. */
namespace
paddle
{
namespace
operators
{
static
void
CreateTensorFromMessageType
(
framework
::
Variable
*
var
,
sendrecv
::
VarType
var_type
)
{
if
(
var_type
==
sendrecv
::
VarType
::
LOD_TENSOR
)
{
var
->
GetMutable
<
framework
::
LoDTensor
>
();
}
else
if
(
var_type
==
sendrecv
::
VarType
::
SELECTED_ROWS
)
{
var
->
GetMutable
<
framework
::
SelectedRows
>
();
}
else
{
PADDLE_THROW
(
"VraibleMessage type %d is not in "
"[LoDTensor, SelectedRows]"
,
var_type
);
}
}
void
RunServer
(
Server
**
rpc_server
,
std
::
shared_ptr
<
detail
::
SendRecvServerImpl
>
service
,
const
std
::
string
&
server_address
)
{
...
...
@@ -111,10 +125,10 @@ class RecvOp : public framework::OperatorBase {
auto
*
merged_grad
=
recv_scope
.
FindVar
(
grad_var_name
);
if
(
merged_grad
==
nullptr
)
{
auto
*
ptr
=
recv_scope
.
Var
(
grad_var_name
);
framework
::
CreateTensor
(
ptr
,
framework
::
ToVarType
(
merged_grad
->
Type
()));
CreateTensorFromMessageType
(
ptr
,
v
.
second
.
type
());
VLOG
(
3
)
<<
"Create Variable "
<<
grad_var_name
<<
" on recv scope, which pointer is "
<<
ptr
;
<<
" on recv scope, which pointer is "
<<
ptr
<<
" type is "
<<
v
.
second
.
type
();
}
if
(
trainer_count
>
1
)
{
...
...
paddle/operators/sum_op.h
浏览文件 @
e249ad12
...
...
@@ -70,6 +70,7 @@ class SumKernel : public framework::OpKernel<T> {
}
else
if
(
out_var
->
IsType
<
framework
::
SelectedRows
>
())
{
PADDLE_ENFORCE
(
!
in_place
,
"SelectedRows not support inplace sum now"
);
auto
*
out
=
context
.
Output
<
SelectedRows
>
(
"Out"
);
out
->
mutable_rows
()
->
clear
();
auto
*
out_value
=
out
->
mutable_value
();
// Runtime InferShape
...
...
python/paddle/v2/fluid/tests/book_distribute/test_dist_word2vec.py
0 → 100644
浏览文件 @
e249ad12
from
__future__
import
print_function
import
numpy
as
np
import
paddle.v2
as
paddle
import
paddle.v2.fluid
as
fluid
import
os
PASS_NUM
=
100
EMBED_SIZE
=
32
HIDDEN_SIZE
=
256
N
=
5
BATCH_SIZE
=
32
IS_SPARSE
=
True
TRAINERS
=
2
word_dict
=
paddle
.
dataset
.
imikolov
.
build_dict
()
dict_size
=
len
(
word_dict
)
first_word
=
fluid
.
layers
.
data
(
name
=
'firstw'
,
shape
=
[
1
],
dtype
=
'int64'
)
second_word
=
fluid
.
layers
.
data
(
name
=
'secondw'
,
shape
=
[
1
],
dtype
=
'int64'
)
third_word
=
fluid
.
layers
.
data
(
name
=
'thirdw'
,
shape
=
[
1
],
dtype
=
'int64'
)
forth_word
=
fluid
.
layers
.
data
(
name
=
'forthw'
,
shape
=
[
1
],
dtype
=
'int64'
)
next_word
=
fluid
.
layers
.
data
(
name
=
'nextw'
,
shape
=
[
1
],
dtype
=
'int64'
)
embed_first
=
fluid
.
layers
.
embedding
(
input
=
first_word
,
size
=
[
dict_size
,
EMBED_SIZE
],
dtype
=
'float32'
,
is_sparse
=
IS_SPARSE
,
param_attr
=
'shared_w'
)
embed_second
=
fluid
.
layers
.
embedding
(
input
=
second_word
,
size
=
[
dict_size
,
EMBED_SIZE
],
dtype
=
'float32'
,
is_sparse
=
IS_SPARSE
,
param_attr
=
'shared_w'
)
embed_third
=
fluid
.
layers
.
embedding
(
input
=
third_word
,
size
=
[
dict_size
,
EMBED_SIZE
],
dtype
=
'float32'
,
is_sparse
=
IS_SPARSE
,
param_attr
=
'shared_w'
)
embed_forth
=
fluid
.
layers
.
embedding
(
input
=
forth_word
,
size
=
[
dict_size
,
EMBED_SIZE
],
dtype
=
'float32'
,
is_sparse
=
IS_SPARSE
,
param_attr
=
'shared_w'
)
concat_embed
=
fluid
.
layers
.
concat
(
input
=
[
embed_first
,
embed_second
,
embed_third
,
embed_forth
],
axis
=
1
)
hidden1
=
fluid
.
layers
.
fc
(
input
=
concat_embed
,
size
=
HIDDEN_SIZE
,
act
=
'sigmoid'
)
predict_word
=
fluid
.
layers
.
fc
(
input
=
hidden1
,
size
=
dict_size
,
act
=
'softmax'
)
cost
=
fluid
.
layers
.
cross_entropy
(
input
=
predict_word
,
label
=
next_word
)
avg_cost
=
fluid
.
layers
.
mean
(
x
=
cost
)
sgd_optimizer
=
fluid
.
optimizer
.
SGD
(
learning_rate
=
0.001
)
optimize_ops
,
params_grads
=
sgd_optimizer
.
minimize
(
avg_cost
)
train_reader
=
paddle
.
batch
(
paddle
.
dataset
.
imikolov
.
train
(
word_dict
,
N
),
BATCH_SIZE
)
place
=
fluid
.
CPUPlace
()
exe
=
fluid
.
Executor
(
place
)
t
=
fluid
.
DistributeTranspiler
()
# all parameter server endpoints list for spliting parameters
pserver_endpoints
=
os
.
getenv
(
"PSERVERS"
)
# server endpoint for current node
current_endpoint
=
os
.
getenv
(
"SERVER_ENDPOINT"
)
# run as trainer or parameter server
training_role
=
os
.
getenv
(
"TRAINING_ROLE"
,
"TRAINER"
)
# get the training role: trainer/pserver
t
.
transpile
(
optimize_ops
,
params_grads
,
pservers
=
pserver_endpoints
,
trainers
=
TRAINERS
)
if
training_role
==
"PSERVER"
:
if
not
current_endpoint
:
print
(
"need env SERVER_ENDPOINT"
)
exit
(
1
)
pserver_prog
=
t
.
get_pserver_program
(
current_endpoint
,
optimize_ops
)
exe
.
run
(
fluid
.
default_startup_program
())
exe
.
run
(
pserver_prog
)
elif
training_role
==
"TRAINER"
:
feeder
=
fluid
.
DataFeeder
(
feed_list
=
[
first_word
,
second_word
,
third_word
,
forth_word
,
next_word
],
place
=
place
)
exe
.
run
(
fluid
.
default_startup_program
())
for
pass_id
in
range
(
PASS_NUM
):
for
data
in
train_reader
():
avg_cost_np
=
exe
.
run
(
fluid
.
default_main_program
(),
feed
=
feeder
.
feed
(
data
),
fetch_list
=
[
avg_cost
])
print
(
"avg_cost_np"
,
avg_cost_np
)
if
avg_cost_np
[
0
]
<
5.0
:
exit
(
0
)
# if avg cost less than 10.0, we think our code is good.
else
:
print
(
"environment var TRAINER_ROLE should be TRAINER os PSERVER"
)
exit
(
1
)
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