Skip to content

P
PaddleDetection

s920243400 / PaddleDetection
与 Fork 源项目一致

Fork自 PaddlePaddle / PaddleDetection

通知 2
Star 0
Fork 0
P
PaddleDetection

体验新版 GitCode，发现更多精彩内容 >>

You need to sign in or sign up before continuing.
未验证 提交 ba9646fd 编写于 作者: Y Yang Zhang 提交者： GitHub
浏览文件
操作

Add cn doc for mobile (#664)

上级 f5d79800
显示空白变更内容
内联 并排
Showing with 113 addition and 29 deletion
configs/mobile/README.md
浏览文件 @ ba9646fd
# Mobile Model Zoo
[English](README_en.md) | 简体中文
# 移动端模型库
## Models
This directory contains models optimized for mobile applications, at present the following models included:
## 模型
| Backbone | Architecture | Input | Image/gpu <sup>1</sup> | Lr schd | Box AP | Download <sup>2</sup> |
PaddleDetection目前提供一系列针对移动应用进行优化的模型，主要支持以下结构:
| 骨干网络 | 结构 | 输入大小 | 图片/gpu <sup>1</sup> | 学习率策略 | Box AP | 下载 <sup>2</sup> |
|--------------------------|---------------------------|-------|------------------------|---------------|--------|-----------------------|
| MobileNetV3 Small | SSDLite | 320 | 64 | 400K (cosine) | 16.6 | [Link](https://paddlemodels.bj.bcebos.com/object_detection/ssdlite_mobilenet_v3_small.tar.gz) |
| MobileNetV3 Large | SSDLite | 320 | 64 | 400K (cosine) | 22.8 | [Link](https://paddlemodels.bj.bcebos.com/object_detection/ssdlite_mobilenet_v3_large.tar.gz) |
| MobileNetV3 Large w/ FPN | Cascade RCNN | 320 | 2 | 500k (cosine) | 25.0 | [Link](https://paddlemodels.bj.bcebos.com/object_detection/cascade_rcnn_mobilenetv3_fpn_320.tar.gz) |
| MobileNetV3 Large w/ FPN | Cascade RCNN | 640 | 2 | 500k (cosine) | 30.2 | [Link](https://paddlemodels.bj.bcebos.com/object_detection/cascade_rcnn_mobilenetv3_fpn_640.tar.gz) |
| MobileNetV3 Large | YOLOv3 | 320 | 8 | 500K | 27.1 | [Link](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_mobilenet_v3.tar.gz) |
| MobileNetV3 Large | YOLOv3 Prune <sup>3</sup> | 320 | 8 | - | 24.6 | [Link](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_mobilenet_v3_prune86_FPGM_320.tar.gz) |
| MobileNetV3 Small | SSDLite | 320 | 64 | 400K (cosine) | 16.6 | [链接](https://paddlemodels.bj.bcebos.com/object_detection/ssdlite_mobilenet_v3_small.tar.gz) |
| MobileNetV3 Large | SSDLite | 320 | 64 | 400K (cosine) | 22.8 | [链接](https://paddlemodels.bj.bcebos.com/object_detection/ssdlite_mobilenet_v3_large.tar.gz) |
| MobileNetV3 Large w/ FPN | Cascade RCNN | 320 | 2 | 500k (cosine) | 25.0 | [链接](https://paddlemodels.bj.bcebos.com/object_detection/cascade_rcnn_mobilenetv3_fpn_320.tar.gz) |
| MobileNetV3 Large w/ FPN | Cascade RCNN | 640 | 2 | 500k (cosine) | 30.2 | [链接](https://paddlemodels.bj.bcebos.com/object_detection/cascade_rcnn_mobilenetv3_fpn_640.tar.gz) |
| MobileNetV3 Large | YOLOv3 | 320 | 8 | 500K | 27.1 | [链接](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_mobilenet_v3.tar.gz) |
| MobileNetV3 Large | YOLOv3 Prune <sup>3</sup> | 320 | 8 | - | 24.6 | [链接](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_mobilenet_v3_prune86_FPGM_320.tar.gz) |
**Notes**:
**注意**:
- <a name="gpu">[1]</a> All models are trained on 8 GPUs.
- <a name="tarball">[2]</a> Each tarball contains the following files
- model weight file (`.pdparams` or `.tar`)
- inference model files (`__model__` and `__params__`)
- Paddle-Lite model file (`.nb`)
- <a name="prune">[3]</a> See the note section on how YOLO head is pruned
- <a name="gpu">[1]</a> 模型统一使用8卡训练.
- <a name="tarball">[2]</a> 压缩包包括下列文件
- 模型权重文件 (`.pdparams` or `.tar`)
- inference model 文件 (`__model__` and `__params__`)
- Paddle-Lite 模型文件 (`.nb`)
- <a name="prune">[3]</a> 参考下面关于YOLO剪裁的说明
## Benchmarks Results
## 评测结果
- Models are benched on following chipsets with [Paddle-Lite](https://github.com/PaddlePaddle/Paddle-Lite) 2.6 (to be released)
- 模型使用 [Paddle-Lite](https://github.com/PaddlePaddle/Paddle-Lite) 2.6 (即将发布) 在下列平台上进行了测试
- Qualcomm Snapdragon 625
- Qualcomm Snapdragon 835
- Qualcomm Snapdragon 845
- Qualcomm Snapdragon 855
- HiSilicon Kirin 970
- HiSilicon Kirin 980
- With 1 CPU thread (latency numbers are in ms)
- 单CPU线程 (单位： ms)
| | SD625 | SD835 | SD845 | SD855 | Kirin 970 | Kirin 980 |
|------------------|---------|---------|---------|---------|-----------|-----------|
......@@ -43,7 +45,7 @@ This directory contains models optimized for mobile applications, at present the
| YOLOv3 prune | 253.98 | 131.279 | 89.4124 | 48.2856 | 122.732 | 55.8626 |
| Cascade RCNN 320 | 286.526 | 125.635 | 87.404 | 46.184 | 149.179 | 52.9994 |
| Cascade RCNN 640 | 1115.66 | 495.926 | 351.361 | 189.722 | 573.558 | 207.917 |
- With 4 CPU threads (latency numbers are in ms)
- 4 CPU线程 (单位： ms)
| | SD625 | SD835 | SD845 | SD855 | Kirin 970 | Kirin 980 |
|------------------|---------|---------|---------|---------|-----------|-----------|
......@@ -55,26 +57,26 @@ This directory contains models optimized for mobile applications, at present the
| Cascade RCNN 640 | 473.083 | 224.543 | 156.205 | 100.686 | 231.108 | 138.391 |
## Notes on YOLOv3 pruning
## YOLOv3剪裁说明
We pruned the YOLO-head and distill the pruned model with YOLOv3-ResNet34 as the teacher, which has a higher mAP on COCO (31.4 with 320\*320 input).
首先对YOLO检测头进行剪裁，然后再使用 YOLOv3-ResNet34 作为teacher网络对剪裁后的模型进行蒸馏, teacher网络在COCO的mAP为31.4 (输入大小320\*320).
The following configurations can be used for pruning:
可以使用如下两种方式进行剪裁:
- Prune with fixed ratio, overall prune ratios is 86%
- 固定比例剪裁, 整体剪裁率是86%
```shell
--pruned_params="yolo_block.0.0.0.conv.weights,yolo_block.0.0.1.conv.weights,yolo_block.0.1.0.conv.weights,yolo_block.0.1.1.conv.weights,yolo_block.0.2.conv.weights,yolo_block.0.tip.conv.weights,yolo_block.1.0.0.conv.weights,yolo_block.1.0.1.conv.weights,yolo_block.1.1.0.conv.weights,yolo_block.1.1.1.conv.weights,yolo_block.1.2.conv.weights,yolo_block.1.tip.conv.weights,yolo_block.2.0.0.conv.weights,yolo_block.2.0.1.conv.weights,yolo_block.2.1.0.conv.weights,yolo_block.2.1.1.conv.weights,yolo_block.2.2.conv.weights,yolo_block.2.tip.conv.weights" \
--pruned_ratios="0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.875,0.875,0.875,0.875,0.875,0.875"
```
- Prune filters using [FPGM](https://arxiv.org/abs/1811.00250) algorithm:
- 使用 [FPGM](https://arxiv.org/abs/1811.00250) 算法剪裁:
```shell
--prune_criterion=geometry_median
```
## Upcoming
## 敬请关注后续发布
- [ ] More models configurations
- [ ] Quantized models
- [ ] 更多模型
- [ ] 量化模型
configs/mobile/README_en.md 0 → 100755
浏览文件 @ ba9646fd
English | [简体中文](README.md)
# Mobile Model Zoo
## Models
This directory contains models optimized for mobile applications, at present the following models included:
| Backbone | Architecture | Input | Image/gpu <sup>1</sup> | Lr schd | Box AP | Download <sup>2</sup> |
|--------------------------|---------------------------|-------|------------------------|---------------|--------|-----------------------|
| MobileNetV3 Small | SSDLite | 320 | 64 | 400K (cosine) | 16.6 | [Link](https://paddlemodels.bj.bcebos.com/object_detection/ssdlite_mobilenet_v3_small.tar.gz) |
| MobileNetV3 Large | SSDLite | 320 | 64 | 400K (cosine) | 22.8 | [Link](https://paddlemodels.bj.bcebos.com/object_detection/ssdlite_mobilenet_v3_large.tar.gz) |
| MobileNetV3 Large w/ FPN | Cascade RCNN | 320 | 2 | 500k (cosine) | 25.0 | [Link](https://paddlemodels.bj.bcebos.com/object_detection/cascade_rcnn_mobilenetv3_fpn_320.tar.gz) |
| MobileNetV3 Large w/ FPN | Cascade RCNN | 640 | 2 | 500k (cosine) | 30.2 | [Link](https://paddlemodels.bj.bcebos.com/object_detection/cascade_rcnn_mobilenetv3_fpn_640.tar.gz) |
| MobileNetV3 Large | YOLOv3 | 320 | 8 | 500K | 27.1 | [Link](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_mobilenet_v3.tar.gz) |
| MobileNetV3 Large | YOLOv3 Prune <sup>3</sup> | 320 | 8 | - | 24.6 | [Link](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_mobilenet_v3_prune86_FPGM_320.tar.gz) |
**Notes**:
- <a name="gpu">[1]</a> All models are trained on 8 GPUs.
- <a name="tarball">[2]</a> Each tarball contains the following files
- model weight file (`.pdparams` or `.tar`)
- inference model files (`__model__` and `__params__`)
- Paddle-Lite model file (`.nb`)
- <a name="prune">[3]</a> See the note section on how YOLO head is pruned
## Benchmarks Results
- Models are benched on following chipsets with [Paddle-Lite](https://github.com/PaddlePaddle/Paddle-Lite) 2.6 (to be released)
- Qualcomm Snapdragon 625
- Qualcomm Snapdragon 835
- Qualcomm Snapdragon 845
- Qualcomm Snapdragon 855
- HiSilicon Kirin 970
- HiSilicon Kirin 980
- With 1 CPU thread (latency numbers are in ms)
| | SD625 | SD835 | SD845 | SD855 | Kirin 970 | Kirin 980 |
|------------------|---------|---------|---------|---------|-----------|-----------|
| SSDLite Large | 289.071 | 134.408 | 91.933 | 48.2206 | 144.914 | 55.1186 |
| SSDLite Small | 122.932 | 57.1914 | 41.003 | 22.0694 | 61.5468 | 25.2106 |
| YOLOv3 baseline | 1082.5 | 435.77 | 317.189 | 155.948 | 536.987 | 178.999 |
| YOLOv3 prune | 253.98 | 131.279 | 89.4124 | 48.2856 | 122.732 | 55.8626 |
| Cascade RCNN 320 | 286.526 | 125.635 | 87.404 | 46.184 | 149.179 | 52.9994 |
| Cascade RCNN 640 | 1115.66 | 495.926 | 351.361 | 189.722 | 573.558 | 207.917 |
- With 4 CPU threads (latency numbers are in ms)
| | SD625 | SD835 | SD845 | SD855 | Kirin 970 | Kirin 980 |
|------------------|---------|---------|---------|---------|-----------|-----------|
| SSDLite Large | 107.535 | 51.1382 | 34.6392 | 20.4978 | 50.5598 | 24.5318 |
| SSDLite Small | 51.5704 | 24.5156 | 18.5486 | 11.4218 | 24.9946 | 16.7158 |
| YOLOv3 baseline | 413.486 | 184.248 | 133.624 | 75.7354 | 202.263 | 126.435 |
| YOLOv3 prune | 98.5472 | 53.6228 | 34.4306 | 21.3112 | 44.0722 | 31.201 |
| Cascade RCNN 320 | 131.515 | 59.6026 | 39.4338 | 23.5802 | 58.5046 | 36.9486 |
| Cascade RCNN 640 | 473.083 | 224.543 | 156.205 | 100.686 | 231.108 | 138.391 |
## Notes on YOLOv3 pruning
We pruned the YOLO-head and distill the pruned model with YOLOv3-ResNet34 as the teacher, which has a higher mAP on COCO (31.4 with 320\*320 input).
The following configurations can be used for pruning:
- Prune with fixed ratio, overall prune ratios is 86%
```shell
--pruned_params="yolo_block.0.0.0.conv.weights,yolo_block.0.0.1.conv.weights,yolo_block.0.1.0.conv.weights,yolo_block.0.1.1.conv.weights,yolo_block.0.2.conv.weights,yolo_block.0.tip.conv.weights,yolo_block.1.0.0.conv.weights,yolo_block.1.0.1.conv.weights,yolo_block.1.1.0.conv.weights,yolo_block.1.1.1.conv.weights,yolo_block.1.2.conv.weights,yolo_block.1.tip.conv.weights,yolo_block.2.0.0.conv.weights,yolo_block.2.0.1.conv.weights,yolo_block.2.1.0.conv.weights,yolo_block.2.1.1.conv.weights,yolo_block.2.2.conv.weights,yolo_block.2.tip.conv.weights" \
--pruned_ratios="0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.75,0.875,0.875,0.875,0.875,0.875,0.875"
```
- Prune filters using [FPGM](https://arxiv.org/abs/1811.00250) algorithm:
```shell
--prune_criterion=geometry_median
```
## Upcoming
- [ ] More models configurations
- [ ] Quantized models
Markdown is supported
0% 或 .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑！
想要评论请 注册 或