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
wmsofts
Paddle
提交
0c70f34c
P
Paddle
项目概览
wmsofts
/
Paddle
与 Fork 源项目一致
Fork自
PaddlePaddle / Paddle
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
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
体验新版 GitCode,发现更多精彩内容 >>
提交
0c70f34c
编写于
6月 30, 2017
作者:
D
dangqingqing
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
Fix bug for flowers dataset and row_conv.
上级
c5dc0b73
变更
3
显示空白变更内容
内联
并排
Showing
3 changed file
with
14 addition
and
9 deletion
+14
-9
python/paddle/trainer/config_parser.py
python/paddle/trainer/config_parser.py
+2
-2
python/paddle/trainer_config_helpers/tests/configs/protostr/test_row_conv.protostr
...fig_helpers/tests/configs/protostr/test_row_conv.protostr
+1
-1
python/paddle/v2/dataset/flowers.py
python/paddle/v2/dataset/flowers.py
+11
-6
未找到文件。
python/paddle/trainer/config_parser.py
浏览文件 @
0c70f34c
...
...
@@ -2082,10 +2082,10 @@ class MaxOutLayer(LayerBase):
class
RowConvLayer
(
LayerBase
):
def
__init__
(
self
,
name
,
inputs
,
context_length
,
**
xargs
):
super
(
RowConvLayer
,
self
).
__init__
(
name
,
'
maxout
'
,
0
,
inputs
=
inputs
,
**
xargs
)
name
,
'
row_conv
'
,
0
,
inputs
=
inputs
,
**
xargs
)
config_assert
(
len
(
self
.
inputs
)
==
1
,
'
TransLayer must have one and only one input
'
)
'
row convolution layer must have one and only one input.
'
)
input_layer
=
self
.
get_input_layer
(
0
)
row_conv_conf
=
self
.
config
.
inputs
[
0
].
row_conv_conf
row_conv_conf
.
context_length
=
context_length
...
...
python/paddle/trainer_config_helpers/tests/configs/protostr/test_row_conv.protostr
浏览文件 @
0c70f34c
...
...
@@ -7,7 +7,7 @@ layers {
}
layers {
name: "__row_conv_layer_0__"
type: "
maxout
"
type: "
row_conv
"
size: 2560
active_type: "relu"
inputs {
...
...
python/paddle/v2/dataset/flowers.py
浏览文件 @
0c70f34c
...
...
@@ -30,6 +30,7 @@ http://www.robots.ox.ac.uk/~vgg/publications/papers/nilsback08.{pdf,ps.gz}.
"""
import
cPickle
import
itertools
import
functools
from
common
import
download
import
tarfile
import
scipy.io
as
scio
...
...
@@ -54,21 +55,25 @@ TEST_FLAG = 'trnid'
VALID_FLAG
=
'valid'
def
default_mapper
(
sample
):
def
default_mapper
(
is_train
,
sample
):
'''
map image bytes data to type needed by model input layer
'''
img
,
label
=
sample
img
=
load_image_bytes
(
img
)
img
=
simple_transform
(
img
,
256
,
224
,
True
)
img
=
simple_transform
(
img
,
256
,
224
,
is_train
)
return
img
.
flatten
().
astype
(
'float32'
),
label
train_mapper
=
functools
.
partial
(
default_mapper
,
True
)
test_mapper
=
functools
.
partial
(
default_mapper
,
False
)
def
reader_creator
(
data_file
,
label_file
,
setid_file
,
dataset_name
,
mapper
=
default_mapper
,
mapper
,
buffered_size
=
1024
,
use_xmap
=
True
):
'''
...
...
@@ -118,7 +123,7 @@ def reader_creator(data_file,
return
map_readers
(
mapper
,
reader
)
def
train
(
mapper
=
default
_mapper
,
buffered_size
=
1024
,
use_xmap
=
True
):
def
train
(
mapper
=
train
_mapper
,
buffered_size
=
1024
,
use_xmap
=
True
):
'''
Create flowers training set reader.
It returns a reader, each sample in the reader is
...
...
@@ -141,7 +146,7 @@ def train(mapper=default_mapper, buffered_size=1024, use_xmap=True):
buffered_size
,
use_xmap
)
def
test
(
mapper
=
defaul
t_mapper
,
buffered_size
=
1024
,
use_xmap
=
True
):
def
test
(
mapper
=
tes
t_mapper
,
buffered_size
=
1024
,
use_xmap
=
True
):
'''
Create flowers test set reader.
It returns a reader, each sample in the reader is
...
...
@@ -164,7 +169,7 @@ def test(mapper=default_mapper, buffered_size=1024, use_xmap=True):
buffered_size
,
use_xmap
)
def
valid
(
mapper
=
defaul
t_mapper
,
buffered_size
=
1024
,
use_xmap
=
True
):
def
valid
(
mapper
=
tes
t_mapper
,
buffered_size
=
1024
,
use_xmap
=
True
):
'''
Create flowers validation set reader.
It returns a reader, each sample in the reader is
...
...
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