Jas
Congratulations on your great work! MixSTE is a nice work that captures the temporal motion of different body joints and significantly outperforms the prior arts. Looking forward to your awesome...
Thanks! We are sorry that we do not have enough resources to support all these awesome algorithms. Please vote [here](https://github.com/open-mmlab/mmpose/issues/9) to raise its priority. We sincerely appreciate contributions from the...
可以打印出来预测出来的 bbox 框的坐标，看一下是否合理
@Indigo6 For now, we do not have enough manpower to support all these awesome algorithms. Your contribution is really helpful! We appreciate it very much. `Can we support direct 3d...
Could you please provide more details about rokoko dataset?
The problem is about `data/coco/person_detection_results/COCO_val2017_detections_AP_H_56_person.json`, this file is used for evaluation only. It provides **detected** bounding boxes, which can be used for keypoint evaluation. Since you do not have this,...
1. _I also had to comment out these two lines from the code above as it gave an error about PCK metric:_ Yes, for coco dataset, AP is used for...
_are you referring me to uninstall and reinstall mmcv-full?_ Yes, please try again. _would my COCO dataset be compatible with these bottom-up approaches?_ Yes, please have a try.
`_base_ = ['../../../../_base_/datasets/coco.py']` This should be replaced with your own 4-kpt dataset
val_dataloader=dict(samples_per_gpu=1), For bottom-up models, during evaluation, only batchsize=1 is supported.