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
兽拳
Paddle
提交
e72832ad
P
Paddle
项目概览
兽拳
/
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 >>
未验证
提交
e72832ad
编写于
4月 28, 2020
作者:
F
FDInSky
提交者:
GitHub
4月 28, 2020
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
fix generate_proposals (#23797)
* test=develop fix generate_proposals
上级
3047b69f
变更
3
隐藏空白更改
内联
并排
Showing
3 changed file
with
41 addition
and
6 deletion
+41
-6
python/paddle/fluid/layers/detection.py
python/paddle/fluid/layers/detection.py
+10
-3
python/paddle/fluid/tests/test_detection.py
python/paddle/fluid/tests/test_detection.py
+1
-1
python/paddle/fluid/tests/unittests/test_generate_proposals_op.py
...addle/fluid/tests/unittests/test_generate_proposals_op.py
+30
-2
未找到文件。
python/paddle/fluid/layers/detection.py
浏览文件 @
e72832ad
...
...
@@ -2746,7 +2746,8 @@ def generate_proposals(scores,
nms_thresh
=
0.5
,
min_size
=
0.1
,
eta
=
1.0
,
name
=
None
):
name
=
None
,
return_rois_num
=
False
):
"""
**Generate proposal Faster-RCNN**
...
...
@@ -2793,7 +2794,10 @@ def generate_proposals(scores,
width < min_size. The data type must be float32. `0.1` by default.
eta(float): Apply in adaptive NMS, if adaptive `threshold > 0.5`,
`adaptive_threshold = adaptive_threshold * eta` in each iteration.
return_rois_num(bool): When setting True, it will return a 1D Tensor with shape [N, ] that includes Rois's
num of each image in one batch. The N is the image's num. For example, the tensor has values [4,5] that represents
the first image has 4 Rois, the second image has 5 Rois. It only used in rcnn model.
'False' by default.
Returns:
tuple:
A tuple with format ``(rpn_rois, rpn_roi_probs)``.
...
...
@@ -2847,7 +2851,10 @@ def generate_proposals(scores,
rpn_roi_probs
.
stop_gradient
=
True
rpn_rois_lod
.
stop_gradient
=
True
return
rpn_rois
,
rpn_roi_probs
,
rpn_rois_lod
if
return_rois_num
:
return
rpn_rois
,
rpn_roi_probs
,
rpn_rois_lod
else
:
return
rpn_rois
,
rpn_roi_probs
def
box_clip
(
input
,
im_info
,
name
=
None
):
...
...
python/paddle/fluid/tests/test_detection.py
浏览文件 @
e72832ad
...
...
@@ -505,7 +505,7 @@ class TestGenerateProposals(unittest.TestCase):
name
=
'bbox_deltas'
,
shape
=
[
num_anchors
*
4
,
8
,
8
],
dtype
=
'float32'
)
rpn_rois
,
rpn_roi_probs
,
_
=
fluid
.
layers
.
generate_proposals
(
rpn_rois
,
rpn_roi_probs
=
fluid
.
layers
.
generate_proposals
(
name
=
'generate_proposals'
,
scores
=
scores
,
bbox_deltas
=
bbox_deltas
,
...
...
python/paddle/fluid/tests/unittests/test_generate_proposals_op.py
浏览文件 @
e72832ad
...
...
@@ -282,8 +282,6 @@ class TestGenerateProposalsOp(OpTest):
self
.
outputs
=
{
'RpnRois'
:
(
self
.
rpn_rois
[
0
],
[
self
.
lod
]),
'RpnRoiProbs'
:
(
self
.
rpn_roi_probs
[
0
],
[
self
.
lod
]),
'RpnRoisLod'
:
(
np
.
asarray
(
self
.
lod
,
dtype
=
np
.
int32
))
}
def
test_check_output
(
self
):
...
...
@@ -328,5 +326,35 @@ class TestGenerateProposalsOp(OpTest):
self
.
nms_thresh
,
self
.
min_size
,
self
.
eta
)
class
TestGenerateProposalsOutLodOp
(
TestGenerateProposalsOp
):
def
set_data
(
self
):
self
.
init_test_params
()
self
.
init_test_input
()
self
.
init_test_output
()
self
.
inputs
=
{
'Scores'
:
self
.
scores
,
'BboxDeltas'
:
self
.
bbox_deltas
,
'ImInfo'
:
self
.
im_info
.
astype
(
np
.
float32
),
'Anchors'
:
self
.
anchors
,
'Variances'
:
self
.
variances
}
self
.
attrs
=
{
'pre_nms_topN'
:
self
.
pre_nms_topN
,
'post_nms_topN'
:
self
.
post_nms_topN
,
'nms_thresh'
:
self
.
nms_thresh
,
'min_size'
:
self
.
min_size
,
'eta'
:
self
.
eta
,
'return_rois_num'
:
True
}
self
.
outputs
=
{
'RpnRois'
:
(
self
.
rpn_rois
[
0
],
[
self
.
lod
]),
'RpnRoiProbs'
:
(
self
.
rpn_roi_probs
[
0
],
[
self
.
lod
]),
'RpnRoisLod'
:
(
np
.
asarray
(
self
.
lod
,
dtype
=
np
.
int32
))
}
if
__name__
==
'__main__'
:
unittest
.
main
()
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