[Feature] Add Hourglass pose detection quantization evaluation result. (#1066)

Detail results: https://github.com/Qingrenn/mmdeploy-summer-camp/blob/main/week4/result.md
This commit is contained in:
Qingren 2022-09-19 14:09:46 +08:00 committed by GitHub
parent 0caeaf238c
commit 18ccb15b7b
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 16 additions and 0 deletions

View File

@ -25,3 +25,11 @@ Note:
| [PANet](https://github.com/open-mmlab/mmocr/blob/main/configs/textdet/panet/panet_r18_fpem_ffm_600e_icdar2015.py) | ICDAR2015 | 0.795 | 0.792 @thr=0.9 |
Note: [mmocr](https://github.com/open-mmlab/mmocr) Uses 'shapely' to compute IoU, which results in a slight difference in accuracy
### Pose detection
| model | dataset | fp32 AP | int8 AP |
| :----------------------------------------------------------------------------------------------------------------------------------------------: | :------: | :-----: | :-----: |
| [Hourglass](https://github.com/open-mmlab/mmpose/blob/master/configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/hourglass52_coco_256x256.py) | COCO2017 | 0.726 | 0.713 |
Note: MMPose models are tested with `flip_test` explicitly set to `False` in model configs.

View File

@ -25,3 +25,11 @@
| [PANet](https://github.com/open-mmlab/mmocr/blob/main/configs/textdet/panet/panet_r18_fpem_ffm_600e_icdar2015.py) | ICDAR2015 | 0.795 | 0.792 @thr=0.9 |
备注:[mmocr](https://github.com/open-mmlab/mmocr) 使用 `shapely` 计算 IoU实现方法会导致轻微的精度差异
### 姿态检测任务
| model | dataset | fp32 AP | int8 AP |
| :----------------------------------------------------------------------------------------------------------------------------------------------: | :------: | :-----: | :-----: |
| [Hourglass](https://github.com/open-mmlab/mmpose/blob/master/configs/body/2d_kpt_sview_rgb_img/topdown_heatmap/coco/hourglass52_coco_256x256.py) | COCO2017 | 0.726 | 0.713 |
备注:测试转换后的模型精度时,对于 mmpose 模型,在模型配置文件中 `flip_test` 需设置为 `False`