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
models
提交
a40349a4
M
models
项目概览
PaddlePaddle
/
models
大约 1 年 前同步成功
通知
222
Star
6828
Fork
2962
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
602
列表
看板
标记
里程碑
合并请求
255
Wiki
0
Wiki
分析
仓库
DevOps
项目成员
Pages
M
models
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
602
Issue
602
列表
看板
标记
里程碑
合并请求
255
合并请求
255
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
未验证
提交
a40349a4
编写于
9月 11, 2019
作者:
R
ruri
提交者:
GitHub
9月 11, 2019
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
Unify model name (#3316)
上级
f10df497
变更
8
隐藏空白更改
内联
并排
Showing
8 changed file
with
25 addition
and
13 deletion
+25
-13
PaddleCV/image_classification/README.md
PaddleCV/image_classification/README.md
+1
-1
PaddleCV/image_classification/README_en.md
PaddleCV/image_classification/README_en.md
+1
-1
PaddleCV/image_classification/models/__init__.py
PaddleCV/image_classification/models/__init__.py
+3
-3
PaddleCV/image_classification/models/mobilenet_v1.py
PaddleCV/image_classification/models/mobilenet_v1.py
+9
-1
PaddleCV/image_classification/models/mobilenet_v2.py
PaddleCV/image_classification/models/mobilenet_v2.py
+8
-4
PaddleCV/image_classification/scripts/train/MobileNetV1.sh
PaddleCV/image_classification/scripts/train/MobileNetV1.sh
+1
-1
PaddleCV/image_classification/scripts/train/MobileNetV2.sh
PaddleCV/image_classification/scripts/train/MobileNetV2.sh
+1
-1
README.md
README.md
+1
-1
未找到文件。
PaddleCV/image_classification/README.md
浏览文件 @
a40349a4
...
...
@@ -334,7 +334,7 @@ PaddlePaddle/Models ImageClassification 支持自定义数据
|
[
SE_ResNet50_vd
](
https://paddle-imagenet-models-name.bj.bcebos.com/SE_ResNet50_vd_pretrained.tar
)
| 79.52% | 94.75% | 10.345 | 7.662 |
|
[
SE_ResNeXt50_32x4d
](
https://paddle-imagenet-models-name.bj.bcebos.com/SE_ResNeXt50_32x4d_pretrained.tar
)
| 78.44% | 93.96% | 14.916 | 12.126 |
|
[
SE_ResNeXt101_32x4d
](
https://paddle-imagenet-models-name.bj.bcebos.com/SE_ResNeXt101_32x4d_pretrained.tar
)
| 79.12% | 94.20% | 30.085 | 24.110 |
|
[
SENet
_154_vd
](
https://paddle-imagenet-models-name.bj.bcebos.com/SENet_
154_vd_pretrained.tar
)
| 81.40% | 95.48% | 71.892 | 64.855 |
|
[
SENet
154_vd
](
https://paddle-imagenet-models-name.bj.bcebos.com/SENet
154_vd_pretrained.tar
)
| 81.40% | 95.48% | 71.892 | 64.855 |
### Inception Series
| Model | Top-1 | Top-5 | Paddle Fluid inference time(ms) | Paddle TensorRT inference time(ms) |
...
...
PaddleCV/image_classification/README_en.md
浏览文件 @
a40349a4
...
...
@@ -316,7 +316,7 @@ Pretrained models can be downloaded by clicking related model names.
|
[
SE_ResNet50_vd
](
https://paddle-imagenet-models-name.bj.bcebos.com/SE_ResNet50_vd_pretrained.tar
)
| 79.52% | 94.75% | 10.345 | 7.662 |
|
[
SE_ResNeXt50_32x4d
](
https://paddle-imagenet-models-name.bj.bcebos.com/SE_ResNeXt50_32x4d_pretrained.tar
)
| 78.44% | 93.96% | 14.916 | 12.126 |
|
[
SE_ResNeXt101_32x4d
](
https://paddle-imagenet-models-name.bj.bcebos.com/SE_ResNeXt101_32x4d_pretrained.tar
)
| 79.12% | 94.20% | 30.085 | 24.110 |
|
[
SENet
_154_vd
](
https://paddle-imagenet-models-name.bj.bcebos.com/SENet_
154_vd_pretrained.tar
)
| 81.40% | 95.48% | 71.892 | 64.855 |
|
[
SENet
154_vd
](
https://paddle-imagenet-models-name.bj.bcebos.com/SENet
154_vd_pretrained.tar
)
| 81.40% | 95.48% | 71.892 | 64.855 |
### Inception Series
| Model | Top-1 | Top-5 | Paddle Fluid inference time(ms) | Paddle TensorRT inference time(ms) |
...
...
PaddleCV/image_classification/models/__init__.py
浏览文件 @
a40349a4
...
...
@@ -13,8 +13,8 @@
#limitations under the License.
from
.alexnet
import
AlexNet
from
.mobilenet_v1
import
MobileNetV1_x0_25
,
MobileNetV1_x0_5
,
MobileNetV1_x0_75
,
MobileNetV1
from
.mobilenet_v2
import
MobileNetV2_x0_25
,
MobileNetV2_x0_5
,
MobileNetV2_x0_75
,
MobileNetV2
,
MobileNetV2_x1_5
,
MobileNetV2_x2_0
from
.mobilenet_v1
import
MobileNetV1_x0_25
,
MobileNetV1_x0_5
,
MobileNetV1_x
1_0
,
MobileNetV1_x
0_75
,
MobileNetV1
from
.mobilenet_v2
import
MobileNetV2_x0_25
,
MobileNetV2_x0_5
,
MobileNetV2_x0_75
,
MobileNetV2
_x1_0
,
MobileNetV2_x1_5
,
MobileNetV2_x2_0
,
MobileNetV2
from
.mobilenet_v3
import
MobileNetV3_small_x0_25
,
MobileNetV3_small_x0_5
,
MobileNetV3_small_x0_75
,
MobileNetV3_small_x1_0
,
MobileNetV3_small_x1_25
,
MobileNetV3_large_x0_25
,
MobileNetV3_large_x0_5
,
MobileNetV3_large_x0_75
,
MobileNetV3_large_x1_0
,
MobileNetV3_large_x1_25
from
.googlenet
import
GoogLeNet
from
.vgg
import
VGG11
,
VGG13
,
VGG16
,
VGG19
...
...
@@ -32,7 +32,7 @@ from .shufflenet_v2_swish import ShuffleNetV2_swish, ShuffleNetV2_x0_5_swish, Sh
from
.shufflenet_v2
import
ShuffleNetV2_x0_25
,
ShuffleNetV2_x0_33
,
ShuffleNetV2_x0_5
,
ShuffleNetV2_x1_0
,
ShuffleNetV2_x1_5
,
ShuffleNetV2_x2_0
,
ShuffleNetV2
from
.fast_imagenet
import
FastImageNet
from
.xception
import
Xception41
,
Xception65
,
Xception71
from
.xception_deeplab
import
Xception41_deeplab
,
Xception65_deeplab
,
Xception71_deeplab
from
.xception_deeplab
import
Xception41_deeplab
,
Xception65_deeplab
,
Xception71_deeplab
from
.densenet
import
DenseNet121
,
DenseNet161
,
DenseNet169
,
DenseNet201
,
DenseNet264
from
.squeezenet
import
SqueezeNet1_0
,
SqueezeNet1_1
from
.darknet
import
DarkNet53
...
...
PaddleCV/image_classification/models/mobilenet_v1.py
浏览文件 @
a40349a4
...
...
@@ -20,7 +20,10 @@ import paddle.fluid as fluid
from
paddle.fluid.initializer
import
MSRA
from
paddle.fluid.param_attr
import
ParamAttr
__all__
=
[
'MobileNetV1'
,
'MobileNetV1_x0_25'
,
'MobileNetV1_x0_5'
,
'MobileNetV1_x0_75'
]
__all__
=
[
'MobileNetV1'
,
'MobileNetV1_x0_25'
,
'MobileNetV1_x0_5'
,
'MobileNetV1_x1_0'
,
'MobileNetV1_x0_75'
]
class
MobileNetV1
():
...
...
@@ -205,6 +208,11 @@ def MobileNetV1_x0_5():
return
model
def
MobileNetV1_x1_0
():
model
=
MobileNetV1
(
scale
=
1.0
)
return
model
def
MobileNetV1_x0_75
():
model
=
MobileNetV1
(
scale
=
0.75
)
return
model
...
...
PaddleCV/image_classification/models/mobilenet_v2.py
浏览文件 @
a40349a4
...
...
@@ -21,8 +21,8 @@ from paddle.fluid.param_attr import ParamAttr
__all__
=
[
'MobileNetV2_x0_25'
,
'MobileNetV2_x0_5'
'MobileNetV2_x0_75'
,
'MobileNetV2'
,
'MobileNetV2_x1_5'
,
'MobileNetV2_x2_0'
,
'MobileNetV2_x0_75'
,
'MobileNetV2
_x1_0
'
,
'MobileNetV2_x1_5'
,
'MobileNetV2_x2_0'
,
'MobileNetV2'
]
...
...
@@ -215,12 +215,11 @@ def MobileNetV2_x0_75():
return
model
def
MobileNetV2
():
def
MobileNetV2
_x1_0
():
model
=
MobileNetV2
(
scale
=
1.0
)
return
model
def
MobileNetV2_x1_5
():
model
=
MobileNetV2
(
scale
=
1.5
)
return
model
...
...
@@ -229,3 +228,8 @@ def MobileNetV2_x1_5():
def
MobileNetV2_x2_0
():
model
=
MobileNetV2
(
scale
=
2.0
)
return
model
def
MobileNetV2
():
model
=
MobileNetV2
(
scale
=
1.0
)
return
model
PaddleCV/image_classification/scripts/train/MobileNetV1.sh
浏览文件 @
a40349a4
...
...
@@ -7,7 +7,7 @@ export FLAGS_fraction_of_gpu_memory_to_use=0.98
python train.py
\
--model
=
MobileNetV1
_x1_0
\
--model
=
MobileNetV1
\
--batch_size
=
256
\
--total_images
=
1281167
\
--class_dim
=
1000
\
...
...
PaddleCV/image_classification/scripts/train/MobileNetV2.sh
浏览文件 @
a40349a4
...
...
@@ -7,7 +7,7 @@ export FLAGS_fraction_of_gpu_memory_to_use=0.98
python train.py
\
--model
=
MobileNetV2
_x1_0
\
--model
=
MobileNetV2
\
--batch_size
=
500
\
--total_images
=
1281167
\
--class_dim
=
1000
\
...
...
README.md
浏览文件 @
a40349a4
...
...
@@ -40,7 +40,7 @@ PaddlePaddle 提供了丰富的计算单元,使得用户可以采用模块化
|
[
Inceptionv4
](
https://github.com/PaddlePaddle/models/tree/develop/PaddleCV/image_classification
)
| 将Inception模块与Residual Connection进行结合,通过ResNet的结构极大地加速训练并获得性能的提升 | ImageNet-2012验证集 | 80.77%/95.26% |
|
[
MobileNetV1
](
https://github.com/PaddlePaddle/models/tree/develop/PaddleCV/image_classification
)
| 将传统的卷积结构改造成两层卷积结构的网络,在基本不影响准确率的前提下大大减少计算时间,更适合移动端和嵌入式视觉应用 | ImageNet-2012验证集 | 70.99%/89.68% |
|
[
MobileNetV2
](
https://github.com/PaddlePaddle/models/tree/develop/PaddleCV/image_classification
)
| MobileNet结构的微调,直接在thinner的bottleneck层上进行skip learning连接以及对bottleneck layer不进行ReLu非线性处理可取得更好的结果 | ImageNet-2012验证集 | 72.15%/90.65% |
|
[
SENet
_
154_vd
](
https://github.com/PaddlePaddle/models/tree/develop/PaddleCV/image_classification
)
| 在ResNeXt 基础、上加入了SE(Sequeeze-and-Excitation)模块,提高了识别准确率,在ILSVRC 2017 的分类项目中取得了第一名 | ImageNet-2012验证集 | 81.40%/95.48% |
|
[
SENet154_vd
](
https://github.com/PaddlePaddle/models/tree/develop/PaddleCV/image_classification
)
| 在ResNeXt 基础、上加入了SE(Sequeeze-and-Excitation)模块,提高了识别准确率,在ILSVRC 2017 的分类项目中取得了第一名 | ImageNet-2012验证集 | 81.40%/95.48% |
|
[
ShuffleNetV2
](
https://github.com/PaddlePaddle/models/tree/develop/PaddleCV/image_classification
)
| ECCV2018,轻量级CNN网络,在速度和准确度之间做了很好地平衡。在同等复杂度下,比ShuffleNet和MobileNetv2更准确,更适合移动端以及无人车领域 | ImageNet-2012验证集 | 70.03%/89.17% |
更多图像分类模型请参考
[
Image Classification
](
https://github.com/PaddlePaddle/models/tree/develop/PaddleCV/image_classification
)
...
...
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