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
BaiXuePrincess
Paddle
提交
e8d24aa1
P
Paddle
项目概览
BaiXuePrincess
/
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
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
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
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
e8d24aa1
编写于
5月 17, 2018
作者:
Q
qiaolongfei
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
Inferencer support parallel_executor
上级
2a636529
变更
2
隐藏空白更改
内联
并排
Showing
2 changed file
with
16 addition
and
8 deletion
+16
-8
python/paddle/fluid/inferencer.py
python/paddle/fluid/inferencer.py
+10
-2
python/paddle/fluid/trainer.py
python/paddle/fluid/trainer.py
+6
-6
未找到文件。
python/paddle/fluid/inferencer.py
浏览文件 @
e8d24aa1
...
@@ -17,6 +17,7 @@ import core
...
@@ -17,6 +17,7 @@ import core
import
executor
import
executor
import
framework
import
framework
import
io
import
io
import
parallel_executor
import
unique_name
import
unique_name
from
trainer
import
check_and_get_place
from
trainer
import
check_and_get_place
...
@@ -24,7 +25,7 @@ __all__ = ['Inferencer', ]
...
@@ -24,7 +25,7 @@ __all__ = ['Inferencer', ]
class
Inferencer
(
object
):
class
Inferencer
(
object
):
def
__init__
(
self
,
infer_func
,
param_path
,
place
=
None
):
def
__init__
(
self
,
infer_func
,
param_path
,
place
=
None
,
parallel
=
False
):
"""
"""
:param infer_func: a function that will return predict Variable
:param infer_func: a function that will return predict Variable
:param param_path: the path where the inference model is saved by fluid.io.save_params
:param param_path: the path where the inference model is saved by fluid.io.save_params
...
@@ -32,13 +33,20 @@ class Inferencer(object):
...
@@ -32,13 +33,20 @@ class Inferencer(object):
"""
"""
self
.
param_path
=
param_path
self
.
param_path
=
param_path
self
.
scope
=
core
.
Scope
()
self
.
scope
=
core
.
Scope
()
self
.
parallel
=
parallel
self
.
place
=
check_and_get_place
(
place
)
self
.
inference_program
=
framework
.
Program
()
self
.
inference_program
=
framework
.
Program
()
with
framework
.
program_guard
(
self
.
inference_program
):
with
framework
.
program_guard
(
self
.
inference_program
):
with
unique_name
.
guard
():
with
unique_name
.
guard
():
self
.
predict_var
=
infer_func
()
self
.
predict_var
=
infer_func
()
self
.
exe
=
executor
.
Executor
(
check_and_get_place
(
place
))
if
parallel
:
self
.
exe
=
parallel_executor
.
ParallelExecutor
(
use_cuda
=
isinstance
(
self
.
place
,
core
.
CUDAPlace
),
loss_name
=
self
.
predict_var
.
name
)
else
:
self
.
exe
=
executor
.
Executor
(
self
.
place
)
with
executor
.
scope_guard
(
self
.
scope
):
with
executor
.
scope_guard
(
self
.
scope
):
# load params from param_path into scope
# load params from param_path into scope
io
.
load_params
(
self
.
exe
,
param_path
,
self
.
inference_program
)
io
.
load_params
(
self
.
exe
,
param_path
,
self
.
inference_program
)
...
...
python/paddle/fluid/trainer.py
浏览文件 @
e8d24aa1
...
@@ -12,18 +12,18 @@
...
@@ -12,18 +12,18 @@
# See the License for the specific language governing permissions and
# See the License for the specific language governing permissions and
# limitations under the License.
# limitations under the License.
import
contextlib
import
os
import
os
import
core
import
core
import
framework
import
executor
import
data_feeder
import
data_feeder
import
contextlib
import
executor
import
framework
import
io
import
io
import
unique_name
import
parallel_executor
# optimizer is same as the parameter of Trainer.__init__. Rename it to opt_module
# optimizer is same as the parameter of Trainer.__init__. Rename it to opt_module
import
optimizer
as
opt_module
import
optimizer
as
opt_module
import
parallel_executor
from
transpiler
import
distribute_transpiler
from
transpiler
import
distribute_transpiler
__all__
=
[
__all__
=
[
...
...
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