dwpose is incompatible with torch 2.5

[Honey-666]

When I upgrade torch to 2.5, it will have this error: 'Adafactor is already registered in optimizer at torch.optim'
@sayakpaul

[asomoza]

We don't actively maintain this repo that much (only patches to fix things), but more importantly, the dwpose code is the same as the original dwpose and comes from OpenMMLab. If any other repository has an update to this I can look into it or even better, you can open a PR yourself if you want (or someone from the community).

Also recently a community member did a PR with the onnx version of it which doesn't have the commercial restriction the original one has:

from easy_dwpose import DWposeDetector

device = "cuda:0" if torch.cuda.is_available() else "cpu"
dwpose = DWposeDetector(device=device)

skeleton = dwpose(input_image, output_type="pil", include_hands=True, include_face=True)

Its an external library so you have to install it: pip install easy-dwpose

[Honey-666]

We don't actively maintain this repo that much (only patches to fix things), but more importantly, the dwpose code is the same as the original dwpose and comes from OpenMMLab. If any other repository has an update to this I can look into it or even better, you can open a PR yourself if you want (or someone from the community).我们并没有积极维护这个仓库（仅进行补丁修复），但更重要的是，dwpose 代码与原始 dwpose 相同，源自 OpenMMLab。如果其他仓库对此有更新，我可以调查，或者更好的是，如果你愿意（或者社区中的其他人），你可以自己提交一个 PR：

Also recently a community member did a PR with the onnx version of it which doesn't have the commercial restriction the original one has:此外，最近有社区成员提交了一个包含 onnx 版本的 PR，这个版本没有原始版本的商业限制：

from easy_dwpose import DWposeDetector

device = "cuda:0" if torch.cuda.is_available() else "cpu"
dwpose = DWposeDetector(device=device)

skeleton = dwpose(input_image, output_type="pil", include_hands=True, include_face=True)
Its an external library so you have to install it: pip install easy-dwpose这是一个外部库，因此你需要安装它： pip install easy-dwpose

ok! thanks.