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
PaddleRec
提交
f18eea21
P
PaddleRec
项目概览
BaiXuePrincess
/
PaddleRec
与 Fork 源项目一致
Fork自
PaddlePaddle / PaddleRec
通知
1
Star
0
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
PaddleRec
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
f18eea21
编写于
5月 12, 2020
作者:
T
tangwei
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
add mpi cluster
上级
e9011c05
变更
5
隐藏空白更改
内联
并排
Showing
5 changed file
with
10 addition
and
5 deletion
+10
-5
fleet_rec/core/engine/cluster/__init__.py
fleet_rec/core/engine/cluster/__init__.py
+0
-0
fleet_rec/core/engine/cluster/cluster.py
fleet_rec/core/engine/cluster/cluster.py
+1
-1
fleet_rec/core/engine/cluster/job.sh
fleet_rec/core/engine/cluster/job.sh
+0
-0
fleet_rec/core/engine/cluster/submit.sh
fleet_rec/core/engine/cluster/submit.sh
+0
-0
fleet_rec/run.py
fleet_rec/run.py
+9
-4
未找到文件。
fleet_rec/core/engine/
mpi_
cluster/__init__.py
→
fleet_rec/core/engine/cluster/__init__.py
浏览文件 @
f18eea21
文件已移动
fleet_rec/core/engine/
mpi_
cluster/cluster.py
→
fleet_rec/core/engine/cluster/cluster.py
浏览文件 @
f18eea21
...
...
@@ -23,7 +23,7 @@ import copy
from
fleetrec.core.engine.engine
import
Engine
class
QSub
ClusterEngine
(
Engine
):
class
ClusterEngine
(
Engine
):
def
__init_impl__
(
self
):
abs_dir
=
os
.
path
.
dirname
(
os
.
path
.
abspath
(
__file__
))
self
.
submit_script
=
os
.
path
.
join
(
abs_dir
,
"submit.sh"
)
...
...
fleet_rec/core/engine/
mpi_
cluster/job.sh
→
fleet_rec/core/engine/cluster/job.sh
浏览文件 @
f18eea21
文件已移动
fleet_rec/core/engine/
mpi_
cluster/submit.sh
→
fleet_rec/core/engine/cluster/submit.sh
浏览文件 @
f18eea21
文件已移动
fleet_rec/run.py
浏览文件 @
f18eea21
...
...
@@ -108,6 +108,7 @@ def single_engine(args):
def
cluster_engine
(
args
):
from
fleetrec.core.engine.cluster.cluster
import
ClusterEngine
trainer
=
get_trainer_prefix
(
args
)
+
"ClusterTrainer"
cluster_envs
=
{}
cluster_envs
[
"train.trainer.trainer"
]
=
trainer
...
...
@@ -117,8 +118,8 @@ def cluster_engine(args):
print
(
"launch {} engine with cluster to run model: {}"
.
format
(
trainer
,
args
.
model
))
set_runtime_envs
(
cluster_envs
,
args
.
model
)
trainer
=
TrainerFactory
.
create
(
args
.
model
)
return
trainer
launch
=
LocalClusterEngine
(
cluster_envs
,
args
.
model
)
return
launch
def
cluster_mpi_engine
(
args
):
...
...
@@ -201,8 +202,10 @@ if __name__ == "__main__":
parser
.
add_argument
(
"-e"
,
"--engine"
,
type
=
str
,
choices
=
[
"single"
,
"local_cluster"
,
"cluster"
,
"tdm_single"
,
"tdm_local_cluster"
,
"tdm_cluster"
])
parser
.
add_argument
(
"-d"
,
"--device"
,
type
=
str
,
choices
=
[
"cpu"
,
"gpu"
],
default
=
"cpu"
)
parser
.
add_argument
(
"-d"
,
"--device"
,
type
=
str
,
choices
=
[
"cpu"
,
"gpu"
],
default
=
"cpu"
)
parser
.
add_argument
(
"-b"
,
"--backend"
,
type
=
str
,
default
=
None
)
parser
.
add_argument
(
"-r"
,
"--role"
,
type
=
str
,
choices
=
[
"master"
,
"worker"
],
default
=
"master"
)
abs_dir
=
os
.
path
.
dirname
(
os
.
path
.
abspath
(
__file__
))
envs
.
set_runtime_environs
({
"PACKAGE_BASE"
:
abs_dir
})
...
...
@@ -210,6 +213,8 @@ if __name__ == "__main__":
args
=
parser
.
parse_args
()
args
.
engine
=
args
.
engine
.
upper
()
args
.
device
=
args
.
device
.
upper
()
args
.
role
=
args
.
role
.
upper
()
model_name
=
args
.
model
.
split
(
'.'
)[
-
1
]
args
.
model
=
get_abs_model
(
args
.
model
)
engine_registry
()
...
...
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