未验证 提交 2a3cafdc 编写于 作者: G Guanghua Yu 提交者: GitHub

fix modelzoo and custom dataset docs error (#1019)

上级 79270ca2
......@@ -196,7 +196,7 @@ results of image size 608/416/320 above. Deformable conv is added on stage 5 of
| MobileNet_v1 | 300 | 64 | Cosine decay(40w) | - | 23.6 | [model](https://paddlemodels.bj.bcebos.com/object_detection/ssdlite_mobilenet_v1.pdparams) | [config](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/ssd/ssdlite_mobilenet_v1.yml) |
| MobileNet_v3 small | 320 | 64 | Cosine decay(40w) | - | 16.2 | [model](https://paddlemodels.bj.bcebos.com/object_detection/mobile_models/ssdlite_mobilenet_v3_small.pdparams) | [config](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/ssd/ssdlite_mobilenet_v3_small.yml) |
| MobileNet_v3 large | 320 | 64 | Cosine decay(40w) | - | 23.3 | [model](https://paddlemodels.bj.bcebos.com/object_detection/mobile_models/ssdlite_mobilenet_v3_large.pdparams) | [config](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/ssd/ssdlite_mobilenet_v3_large.yml) |
| MobileNet_v3 large w/ FPN | 320 | 64 | Cosine decay(40w) | - | 18.9 | [model](https://paddlemodels.bj.bcebos.com/object_detection/mobile_models/ssdlite_mobilenet_v3_small_fpn.pdparams) | [config](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/ssd/ssdlite_mobilenet_v3_small_fpn.yml) |
| MobileNet_v3 small w/ FPN | 320 | 64 | Cosine decay(40w) | - | 18.9 | [model](https://paddlemodels.bj.bcebos.com/object_detection/mobile_models/ssdlite_mobilenet_v3_small_fpn.pdparams) | [config](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/ssd/ssdlite_mobilenet_v3_small_fpn.yml) |
| MobileNet_v3 large w/ FPN | 320 | 64 | Cosine decay(40w) | - | 24.3 | [model](https://paddlemodels.bj.bcebos.com/object_detection/mobile_models/ssdlite_mobilenet_v3_large_fpn.pdparams) | [config](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/ssd/ssdlite_mobilenet_v3_large_fpn.yml) |
**Notes:** `SSDLite` is trained in 8 GPU with total batch size as 512 and uses cosine decay strategy to train.
......
......@@ -188,7 +188,7 @@ Paddle提供基于ImageNet的骨架网络预训练模型。所有预训练模型
| MobileNet_v1 | 300 | 64 | Cosine decay(40w) | - | 23.6 | [下载链接](https://paddlemodels.bj.bcebos.com/object_detection/ssdlite_mobilenet_v1.pdparams) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/ssd/ssdlite_mobilenet_v1.yml) |
| MobileNet_v3 small | 320 | 64 | Cosine decay(40w) | - | 16.2 | [下载链接](https://paddlemodels.bj.bcebos.com/object_detection/mobile_models/ssdlite_mobilenet_v3_small.pdparams) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/ssd/ssdlite_mobilenet_v3_small.yml) |
| MobileNet_v3 large | 320 | 64 | Cosine decay(40w) | - | 23.3 | [下载链接](https://paddlemodels.bj.bcebos.com/object_detection/mobile_models/ssdlite_mobilenet_v3_large.pdparams) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/ssd/ssdlite_mobilenet_v3_large.yml) |
| MobileNet_v3 large w/ FPN | 320 | 64 | Cosine decay(40w) | - | 18.9 | [下载链接](https://paddlemodels.bj.bcebos.com/object_detection/mobile_models/ssdlite_mobilenet_v3_small_fpn.pdparams) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/ssd/ssdlite_mobilenet_v3_small_fpn.yml) |
| MobileNet_v3 small w/ FPN | 320 | 64 | Cosine decay(40w) | - | 18.9 | [下载链接](https://paddlemodels.bj.bcebos.com/object_detection/mobile_models/ssdlite_mobilenet_v3_small_fpn.pdparams) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/ssd/ssdlite_mobilenet_v3_small_fpn.yml) |
| MobileNet_v3 large w/ FPN | 320 | 64 | Cosine decay(40w) | - | 24.3 | [下载链接](https://paddlemodels.bj.bcebos.com/object_detection/mobile_models/ssdlite_mobilenet_v3_large_fpn.pdparams) | [配置文件](https://github.com/PaddlePaddle/PaddleDetection/tree/master/configs/ssd/ssdlite_mobilenet_v3_large_fpn.yml) |
**注意事项:** SSDLite模型使用学习率余弦衰减策略在8卡GPU下总batch size为512。
......
# 如何训练自定义数据集
## 目录
- [1.数据准备](#1.准备数据)
- [将数据集转换为COCO格式](#方式一将数据集转换为COCO格式)
- [将数据集转换为VOC格式](#方式二将数据集转换为VOC格式)
- [添加新数据源](#方式三添加新数据源)
- [2.选择模型](#2.选择模型)
- [3.修改参数配置](#3.修改参数配置)
- [4.开始训练与部署](#4.开始训练与部署)
- [附:一个自定义数据集demo](#附一个自定义数据集demo)
- [1.数据准备](#1准备数据)
- [将数据集转换为COCO格式](#方式一将数据集转换为COCO格式)
- [将数据集转换为VOC格式](#方式二将数据集转换为VOC格式)
- [添加新数据源](#方式三添加新数据源)
- [2.选择模型](#2选择模型)
- [3.修改参数配置](#3修改参数配置)
- [4.开始训练与部署](#4开始训练与部署)
- [附:一个自定义数据集demo](#附一个自定义数据集demo)
## 1.准备数据
如果数据符合COCO或VOC数据集格式,可以直接进入[2.选择模型](#2.选择模型),否则需要将数据集转换至COCO格式或VOC格式。
如果数据符合COCO或VOC数据集格式,可以直接进入[2.选择模型](#2选择模型),否则需要将数据集转换至COCO格式或VOC格式。
### 方式一:将数据集转换为COCO格式
......@@ -86,7 +86,7 @@ PaddleDetection中提供了丰富的模型库,具体可在[模型库](../MODEL
选择好模型后,需要在`configs`目录中找到对应的配置文件,为了适配在自定义数据集上训练,需要对参数配置做一些修改:
- 数据路径配置: 在yaml配置文件中,依据[1.数据准备](#1.准备数据)中准备好的路径,配置`TrainReader``EvalReader``TestReader`的路径。
- 数据路径配置: 在yaml配置文件中,依据[1.数据准备](#1准备数据)中准备好的路径,配置`TrainReader``EvalReader``TestReader`的路径。
- COCO数据集:
```yaml
dataset:
......@@ -124,7 +124,7 @@ PaddleDetection中提供了丰富的模型库,具体可在[模型库](../MODEL
## 附:一个自定义数据集demo
我们以`AI识虫数据集`为例,对自定义数据集上训练过程进行演示,该数据集提供了2183张图片,其中训练集1693张,验证集与测试集分别有245张,共包含7种昆虫。在AIStudio上有很多用户公开了此数据集,您可以进行搜索并下载,如:[链接1](https://aistudio.baidu.com/aistudio/datasetdetail/34213)[链接2](https://aistudio.baidu.com/aistudio/datasetdetail/19748)等。
我们以`AI识虫数据集`为例,对自定义数据集上训练过程进行演示,该数据集提供了2183张图片,其中训练集1693张,验证集与测试集分别有245张,共包含7种昆虫。下载链接为:[数据集链接](https://paddlemodels.bj.bcebos.com/insect.tar), 在AIStudio上也有很多用户公开了此数据集,您可以进行搜索并下载,如:[链接1](https://aistudio.baidu.com/aistudio/datasetdetail/34213)[链接2](https://aistudio.baidu.com/aistudio/datasetdetail/19748)等。
#### 第一步:准备数据
......@@ -230,8 +230,6 @@ PaddleDetection中提供了丰富的模型库,具体可在[模型库](../MODEL
--use_gpu=True
```
预测结果如下图所示:
<div align="center">
<img src="docs/images/2572.jpeg" />
</div>
![](../images/2572.jpeg)
如仍有疑惑,欢迎给我们提issue。
......@@ -43,9 +43,7 @@ visualdl --logdir vdl_fruit_dir/scalar/ --host <host_IP> --port <port_num>
Result on VisualDL is shown below:
<div align="center">
<img src='../images/visualdl_fruit.jpg' width='800'>
</div>
![](../images/visualdl_fruit.jpg)
Model can be downloaded [here](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_mobilenet_v1_fruit.tar)
......@@ -65,13 +63,8 @@ python -u tools/infer.py -c configs/yolov3_mobilenet_v1_fruit.yml \
Inference images are shown below:
<div align="center">
<img src='../../demo/orange_71.jpg' width='600'>
</div>
![](../../demo/orange_71.jpg)
<div align="center">
<img src='../images/orange_71_detection.jpg' width='600'>
</div>
![](../images/orange_71_detection.jpg)
For detailed infomation of training and evalution, please refer to [GETTING_STARTED.md](GETTING_STARTED.md).
......@@ -47,10 +47,7 @@ visualdl --logdir vdl_fruit_dir/scalar/ --host <host_IP> --port <port_num>
VisualDL结果显示如下:
<div align="center">
<img src='../images/visualdl_fruit.jpg' width='800'>
</div>
![](../images/visualdl_fruit.jpg)
训练模型[下载链接](https://paddlemodels.bj.bcebos.com/object_detection/yolov3_mobilenet_v1_fruit.tar)
......@@ -72,15 +69,9 @@ python -u tools/infer.py -c configs/yolov3_mobilenet_v1_fruit.yml \
预测图片如下:
![](../../demo/orange_71.jpg)
<div align="center">
<img src='../../demo/orange_71.jpg' width='600'>
</div>
<div align="center">
<img src='../images/orange_71_detection.jpg' width='600'>
</div>
![](../images/orange_71_detection.jpg)
更多训练及评估流程,请参考[入门使用文档](GETTING_STARTED_cn.md)
Markdown is supported
0% .
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