do you used 2d keypoint model?

[wangm-word]

thank you for your work! how to use the demo code ? do you used 2d keypoints when you train and vis?

[SangbumChoi]

https://github.com/SangbumChoi/MobileHumanPose/issues/2

For the question of using 2d keypoints please refer to the above issue

For the demo code you might try

python3 demo.py --gpu 0-1 --modelpath {model_path} --input_image {input_image.jpg} --backbone LPSKI

However, I found that current demo code is quite old so it requires to give input of headnet but I need to fix this.

[wangm-word]

2

For the question of using 2d keypoints please refer to the above issue

For the demo code you might try

python3 demo.py --gpu 0-1 --modelpath {model_path} --input_image {input_image.jpg} --backbone LPSKI

However, I found that current demo code is quite old so it requires to give input of headnet but I need to fix this.

thank you very much

[SangbumChoi]

@wangm-word I also updated demo.py but not sure if it works :( Feel free to test