How to see the 3D reconstruction results (e.g. multi-view images)

[JuewenPeng]

No description provided.

[liphao99]

Currently supports exporting mesh , see https://github.com/aigc3d/LHM#-inference-pipeline.

[lingtengqiu]

As we do not add any hard supervision in the canonical space, the 3D avatar model in the canonical space may have a small probability of having slight artifacts on T-posed space.

[JuewenPeng]

I found this code can only export the T-pose mesh. Is it possible to obtain the result under the pose of the reference image?

[lingtengqiu]

I will update the interface to export gaussian model under target-view pose.