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
a18792752712
Paddle
提交
cc5adfb8
P
Paddle
项目概览
a18792752712
/
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 >>
提交
cc5adfb8
编写于
11月 14, 2016
作者:
A
alvations
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
added resnet lstm architecture from GNMT
上级
805856aa
变更
2
隐藏空白更改
内联
并排
Showing
2 changed file
with
95 addition
and
0 deletion
+95
-0
demo/quick_start/train.sh
demo/quick_start/train.sh
+1
-0
demo/quick_start/trainer_config.resnet-lstm.py
demo/quick_start/trainer_config.resnet-lstm.py
+94
-0
未找到文件。
demo/quick_start/train.sh
浏览文件 @
cc5adfb8
...
...
@@ -20,6 +20,7 @@ cfg=trainer_config.lr.py
#cfg=trainer_config.lstm.py
#cfg=trainer_config.bidi-lstm.py
#cfg=trainer_config.db-lstm.py
#cfg=trainer_config.resnet-lstm.py
paddle train
\
--config
=
$cfg
\
--save_dir
=
./output
\
...
...
demo/quick_start/trainer_config.resnet-lstm.py
0 → 100644
浏览文件 @
cc5adfb8
# edit-mode: -*- python -*-
# Copyright (c) 2016 Baidu, Inc. 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.
"""
This configuration is a demonstration of how to implement the stacked LSTM
with residual connections, i.e. an LSTM layer takes the sum of the hidden states
and inputs of the previous LSTM layer instead of only the hidden states.
This architecture is from:
Yonghui Wu, Mike Schuster, Zhifeng Chen, Quoc V. Le, Mohammad Norouzi,
Wolfgang Macherey, Maxim Krikun, Yuan Cao, Qin Gao, Klaus Macherey,
Jeff Klingner, Apurva Shah, Melvin Johnson, Xiaobing Liu, Lukasz Kaiser,
Stephan Gouws, Yoshikiyo Kato, Taku Kudo, Hideto Kazawa, Keith Stevens,
George Kurian, Nishant Patil, Wei Wang, Cliff Young, Jason Smith, Jason Riesa,
Alex Rudnick, Oriol Vinyals, Greg Corrado, Macduff Hughes, Jeffrey Dean. 2016.
Google's Neural Machine Translation System: Bridging the Gap between Human and
Machine Translation. In arXiv https://arxiv.org/pdf/1609.08144v2.pdf
Different from the architecture described in the paper, we use a stack single
direction LSTM layers as the first layer instead of bi-directional LSTM. Also,
since this is a demo code, to reduce computation time, we stacked 4 layers
instead of 8 layers.
"""
from
paddle.trainer_config_helpers
import
*
dict_file
=
"./data/dict.txt"
word_dict
=
dict
()
with
open
(
dict_file
,
'r'
)
as
f
:
for
i
,
line
in
enumerate
(
f
):
w
=
line
.
strip
().
split
()[
0
]
word_dict
[
w
]
=
i
is_predict
=
get_config_arg
(
'is_predict'
,
bool
,
False
)
trn
=
'data/train.list'
if
not
is_predict
else
None
tst
=
'data/test.list'
if
not
is_predict
else
'data/pred.list'
process
=
'process'
if
not
is_predict
else
'process_predict'
define_py_data_sources2
(
train_list
=
trn
,
test_list
=
tst
,
module
=
"dataprovider_emb"
,
obj
=
process
,
args
=
{
"dictionary"
:
word_dict
})
batch_size
=
128
if
not
is_predict
else
1
settings
(
batch_size
=
batch_size
,
learning_rate
=
2e-3
,
learning_method
=
AdamOptimizer
(),
regularization
=
L2Regularization
(
8e-4
),
gradient_clipping_threshold
=
25
)
bias_attr
=
ParamAttr
(
initial_std
=
0.
,
l2_rate
=
0.
)
data
=
data_layer
(
name
=
"word"
,
size
=
len
(
word_dict
))
emb
=
embedding_layer
(
input
=
data
,
size
=
128
)
lstm
=
simple_lstm
(
input
=
emb
,
size
=
128
,
lstm_cell_attr
=
ExtraAttr
(
drop_rate
=
0.1
))
previous_input
,
previous_hidden_state
=
emb
,
lstm
for
i
in
range
(
3
):
# The input to the current layer is the sum of the hidden state
# and input of the previous layer.
current_input
=
addto_layer
(
input
=
[
previous_input
,
previous_hidden_state
])
hidden_state
=
simple_lstm
(
input
=
current_input
,
size
=
128
,
lstm_cell_attr
=
ExtraAttr
(
drop_rate
=
0.1
))
previous_input
,
previous_hidden_state
=
current_input
,
hidden_state
lstm
=
previous_hidden_state
lstm_last
=
pooling_layer
(
input
=
lstm
,
pooling_type
=
MaxPooling
())
output
=
fc_layer
(
input
=
lstm_last
,
size
=
2
,
bias_attr
=
bias_attr
,
act
=
SoftmaxActivation
())
if
is_predict
:
maxid
=
maxid_layer
(
output
)
outputs
([
maxid
,
output
])
else
:
label
=
data_layer
(
name
=
"label"
,
size
=
2
)
cls
=
classification_cost
(
input
=
output
,
label
=
label
)
outputs
(
cls
)
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