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
cccc61c0
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
cccc61c0
编写于
4月 02, 2020
作者:
T
tangwei12
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
fix ctr-dnn local training
上级
afec7a49
变更
5
展开全部
隐藏空白更改
内联
并排
Showing
5 changed file
with
109 addition
and
108 deletion
+109
-108
examples/ctr-dnn_train.yaml
examples/ctr-dnn_train.yaml
+2
-5
models/ctr_dnn/data/train/sample_train.txt
models/ctr_dnn/data/train/sample_train.txt
+100
-100
models/ctr_dnn/dataset.py
models/ctr_dnn/dataset.py
+4
-0
models/ctr_dnn/model.py
models/ctr_dnn/model.py
+2
-2
trainer/single_train.py
trainer/single_train.py
+1
-1
未找到文件。
examples/ctr-dnn_train.yaml
浏览文件 @
cccc61c0
...
...
@@ -34,7 +34,7 @@ train:
reader
:
mode
:
"
dataset"
batch_size
:
32
batch_size
:
2
pipe_command
:
"
python
/paddle/eleps/models/ctr_dnn/dataset.py"
train_data_path
:
"
/paddle/eleps/models/ctr_dnn/data/train"
...
...
@@ -45,10 +45,7 @@ train:
sparse_feature_number
:
1000001
sparse_feature_dim
:
8
dense_input_dim
:
13
fc_sizes
:
[
101
,
512
,
32
]
# - 1024
# - 512
# - 32
fc_sizes
:
[
512
,
256
,
128
,
32
]
learning_rate
:
0.001
save
:
...
...
models/ctr_dnn/data/train/sample_train.txt
浏览文件 @
cccc61c0
此差异已折叠。
点击以展开。
models/ctr_dnn/dataset.py
浏览文件 @
cccc61c0
...
...
@@ -11,6 +11,8 @@
# 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.
from
__future__
import
print_function
import
sys
import
paddle.fluid.incubate.data_generator
as
dg
...
...
@@ -37,6 +39,7 @@ class CriteoDataset(dg.MultiSlotDataGenerator):
This function needs to be implemented by the user, based on data format
"""
features
=
line
.
rstrip
(
'
\n
'
).
split
(
'
\t
'
)
dense_feature
=
[]
sparse_feature
=
[]
for
idx
in
continuous_range_
:
...
...
@@ -46,6 +49,7 @@ class CriteoDataset(dg.MultiSlotDataGenerator):
dense_feature
.
append
(
(
float
(
features
[
idx
])
-
cont_min_
[
idx
-
1
])
/
cont_diff_
[
idx
-
1
])
for
idx
in
categorical_range_
:
sparse_feature
.
append
(
[
hash
(
str
(
idx
)
+
features
[
idx
])
%
hash_dim_
])
...
...
models/ctr_dnn/model.py
浏览文件 @
cccc61c0
...
...
@@ -39,7 +39,7 @@ class Train(object):
fluid
.
layers
.
data
(
name
=
"C"
+
str
(
i
),
shape
=
[
1
],
lod_level
=
1
,
dtype
=
"int64"
)
for
i
in
range
(
ids
)
dtype
=
"int64"
)
for
i
in
range
(
1
,
ids
)
]
return
sparse_input_ids
,
[
var
.
name
for
var
in
sparse_input_ids
]
...
...
@@ -60,7 +60,7 @@ class Train(object):
self
.
label_input
,
self
.
label_input_varname
=
label_input
()
def
input_vars
(
self
):
return
self
.
sparse_inputs
+
[
self
.
dense_input
]
+
[
self
.
label_input
]
return
[
self
.
dense_input
]
+
self
.
sparse_inputs
+
[
self
.
label_input
]
def
input_varnames
(
self
):
return
[
input
.
name
for
input
in
self
.
input_vars
()]
...
...
trainer/single_train.py
浏览文件 @
cccc61c0
...
...
@@ -169,7 +169,7 @@ class SingleTrainerWithDataset(SingleTrainer):
dataset
=
dataset
,
fetch_list
=
self
.
metrics
,
fetch_info
=
[
"auc "
,
"batch auc"
],
print_period
=
1
00
)
print_period
=
1
)
context
[
'status'
]
=
'infer_pass'
...
...
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