Caenorst
Caenorst
Hi @ERICXUCHI , there is no `TriangleMesh` in Kaolin 0.13, I assume you want to use 0.1 which is a very old version of Kaolin?
Actually from [this issue](https://github.com/pytorch/pytorch/issues/74734#issuecomment-1086418310), it looks like the issue is from the int16. I'll look into it.
Hi @jloveric ! Planning that for next release
Hi @ankarako , if I understand well, you could use https://kaolin.readthedocs.io/en/latest/modules/kaolin.ops.spc.html#kaolin.ops.spc.points_to_corners on the truncated point_hierarchy (use pyramids to get the bounds)
@XiongFenghhh , If you are using nvdiffrast, you could use `nvdiffrast.torch.DepthPeeler`
Hi @prokolyvakis ! Thank you for raising this issue. I will create wheels for CPU, but It may take a little bit of time to set that up since I...
Hi @odusseys , We could add nvdiffrast as dependency, however I believe it will build it on the fly since nvdiffrast has no wheels
Hi @aradhyamathur , you can actually compute sdf by combining [check_sign](https://kaolin.readthedocs.io/en/latest/modules/kaolin.ops.mesh.html#kaolin.ops.mesh.check_sign) and [point_to_mesh_distance](https://kaolin.readthedocs.io/en/latest/modules/kaolin.metrics.trianglemesh.html#kaolin.metrics.trianglemesh.point_to_mesh_distance). Please be careful to apply `torch.sqrt` to `point_to_mesh_distance` as the output is squared
Hi @2601705041, how did you install Kaolin?
@2601705041 , please show command-line the installation instructions are showing multiple ways to install