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
PaddlePaddle
PaddleRec
提交
e48639c6
P
PaddleRec
项目概览
PaddlePaddle
/
PaddleRec
通知
68
Star
12
Fork
5
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
27
列表
看板
标记
里程碑
合并请求
10
Wiki
1
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
27
Issue
27
列表
看板
标记
里程碑
合并请求
10
合并请求
10
Pages
分析
分析
仓库分析
DevOps
Wiki
1
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
e48639c6
编写于
5月 15, 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 license and bug fix
上级
0cab80c0
变更
5
隐藏空白更改
内联
并排
Showing
5 changed file
with
60 addition
and
4 deletion
+60
-4
doc/ps_background.md
doc/ps_background.md
+1
-0
models/contentunderstanding/classification/model.py
models/contentunderstanding/classification/model.py
+17
-4
models/contentunderstanding/classification/reader.py
models/contentunderstanding/classification/reader.py
+14
-0
models/contentunderstanding/tagspace/model.py
models/contentunderstanding/tagspace/model.py
+14
-0
models/contentunderstanding/tagspace/reader.py
models/contentunderstanding/tagspace/reader.py
+14
-0
未找到文件。
doc/ps_background.md
浏览文件 @
e48639c6
# 参数服务器训练简介
# 参数服务器训练简介
飞桨参数服务器原理介绍请参见:
[
飞桨参数服务器
](
https://github.com/PaddlePaddle/Fleet/blob/develop/markdown_doc/transpiler/transpiler_cpu.md
)
如图1所示,参数服务器是分布式训练领域普遍采用的编程架构,主要包含Server和Worker两个部分,其中Server负责参数的存储和更新,而Worker负责训练。飞桨的参数服务器功能也是基于这种经典的架构进行设计和开发的,同时在这基础上进行了SGD(Stochastic Gradient Descent)算法的创新(Geometric Stochastic Gradient Descent)。当前经过大量的实验验证,最佳的方案是每台机器上启动Server和Worker两个进程,而一个Worker进程中可以包含多个用于训练的线程。
如图1所示,参数服务器是分布式训练领域普遍采用的编程架构,主要包含Server和Worker两个部分,其中Server负责参数的存储和更新,而Worker负责训练。飞桨的参数服务器功能也是基于这种经典的架构进行设计和开发的,同时在这基础上进行了SGD(Stochastic Gradient Descent)算法的创新(Geometric Stochastic Gradient Descent)。当前经过大量的实验验证,最佳的方案是每台机器上启动Server和Worker两个进程,而一个Worker进程中可以包含多个用于训练的线程。
<p
align=
"center"
>
<p
align=
"center"
>
...
...
models/contentunderstanding/classification/model.py
浏览文件 @
e48639c6
# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import
paddle.fluid
as
fluid
import
paddle.fluid
as
fluid
import
math
import
math
...
@@ -34,8 +48,7 @@ class Model(ModelBase):
...
@@ -34,8 +48,7 @@ class Model(ModelBase):
fc_1
=
fluid
.
layers
.
fc
(
input
=
[
conv
],
size
=
hid_dim
)
fc_1
=
fluid
.
layers
.
fc
(
input
=
[
conv
],
size
=
hid_dim
)
# softmax layer
# softmax layer
prediction
=
fluid
.
layers
.
fc
(
input
=
[
fc_1
],
size
=
class_dim
,
act
=
"softmax"
)
prediction
=
fluid
.
layers
.
fc
(
input
=
[
fc_1
],
size
=
class_dim
,
act
=
"softmax"
)
#if is_prediction:
# return prediction
cost
=
fluid
.
layers
.
cross_entropy
(
input
=
prediction
,
label
=
label
)
cost
=
fluid
.
layers
.
cross_entropy
(
input
=
prediction
,
label
=
label
)
avg_cost
=
fluid
.
layers
.
mean
(
x
=
cost
)
avg_cost
=
fluid
.
layers
.
mean
(
x
=
cost
)
acc
=
fluid
.
layers
.
accuracy
(
input
=
prediction
,
label
=
label
)
acc
=
fluid
.
layers
.
accuracy
(
input
=
prediction
,
label
=
label
)
...
@@ -50,9 +63,9 @@ class Model(ModelBase):
...
@@ -50,9 +63,9 @@ class Model(ModelBase):
return
self
.
metrics
return
self
.
metrics
def
optimizer
(
self
):
def
optimizer
(
self
):
learning_rate
=
0.01
#envs.get_global_env("hyper_parameters.base_lr", None, self._namespace)
learning_rate
=
0.01
sgd_optimizer
=
fluid
.
optimizer
.
Adagrad
(
learning_rate
=
learning_rate
)
sgd_optimizer
=
fluid
.
optimizer
.
Adagrad
(
learning_rate
=
learning_rate
)
#sgd_optimizer.minimize(avg_cost)
return
sgd_optimizer
return
sgd_optimizer
...
...
models/contentunderstanding/classification/reader.py
浏览文件 @
e48639c6
# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import
re
import
re
import
sys
import
sys
import
collections
import
collections
...
...
models/contentunderstanding/tagspace/model.py
浏览文件 @
e48639c6
# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import
paddle.fluid
as
fluid
import
paddle.fluid
as
fluid
import
math
import
math
...
...
models/contentunderstanding/tagspace/reader.py
浏览文件 @
e48639c6
# Copyright (c) 2020 PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import
re
import
re
import
sys
import
sys
import
collections
import
collections
...
...
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