未验证 提交 7018a080 编写于 作者: J JYChen 提交者: GitHub

fix link and eval guide (#6452)

上级 e6ad10e5
......@@ -97,7 +97,7 @@ MPII数据集
### 2、数据准备
​ 目前KeyPoint模型支持[COCO](https://cocodataset.org/#keypoints-2017)数据集和[MPII](http://human-pose.mpi-inf.mpg.de/#overview)数据集,数据集的准备方式请参考[关键点数据准备](../../docs/tutorials/data/PrepareDetDataSet.md)
​ 目前KeyPoint模型支持[COCO](https://cocodataset.org/#keypoints-2017)数据集和[MPII](http://human-pose.mpi-inf.mpg.de/#overview)数据集,数据集的准备方式请参考[关键点数据准备](../../docs/tutorials/data/PrepareKeypointDataSet.md)
​ 关于config配置文件内容说明请参考[关键点配置文件说明](../../docs/tutorials/KeyPointConfigGuide_cn.md)
......
......@@ -106,7 +106,7 @@ We also release [PP-TinyPose](./tiny_pose/README_en.md), a real-time keypoint de
### 2.Dataset Preparation
​ Currently, KeyPoint Detection Models support [COCO](https://cocodataset.org/#keypoints-2017) and [MPII](http://human-pose.mpi-inf.mpg.de/#overview). Please refer to [Keypoint Dataset Preparation](../../docs/tutorials/data/PrepareDetDataSet_en.md) to prepare dataset.
​ Currently, KeyPoint Detection Models support [COCO](https://cocodataset.org/#keypoints-2017) and [MPII](http://human-pose.mpi-inf.mpg.de/#overview). Please refer to [Keypoint Dataset Preparation](../../docs/tutorials/data/PrepareKeypointDataSet_en.md) to prepare dataset.
​ About the description for config files, please refer to [Keypoint Config Guild](../../docs/tutorials/KeyPointConfigGuide_en.md).
......
......@@ -232,6 +232,11 @@ CUDA_VISIBLE_DEVICES=0,1,2,3 python3 -m paddle.distributed.launch tools/train.py
python3 tools/eval.py -c configs/keypoint/hrnet/hrnet_w32_256x192.yml
```
注意:由于测试依赖pycocotools工具,其默认为`COCO`数据集的17点,如果修改后的模型并非预测17点,直接使用评估命令会报错。
需要修改以下内容以获得正确的评估结果:
- [sigma列表](https://github.com/PaddlePaddle/PaddleDetection/blob/develop/ppdet/modeling/keypoint_utils.py#L219),表示每个关键点的范围方差,越大则容忍度越高。其长度与预测点数一致。根据实际关键点可信区域设置,区域精确的一般0.25-0.5,例如眼睛。区域范围大的一般0.5-1.0,例如肩膀。若不确定建议0.75。
- [pycocotools sigma列表](https://github.com/cocodataset/cocoapi/blob/master/PythonAPI/pycocotools/cocoeval.py#L523),含义及内容同上,取值与sigma列表一致。
### 模型导出及预测
#### Top-Down模型联合部署
```shell
......
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