Jeremy Reizenstein

This is much too little information for us to be able to help you. You need to make pip output the actual error being experienced, and -v should help. There...

Looks like a compiler wasn't found. Do you have dev tools installed? Is nvcc working? (try `nvcc --version` and `g++ --version`).

This isn't a pytorch3d-specific problem. It looks like your pytorch version is too new for your version of the nvidia driver, so PyTorch isn't able to use cuda at all....

@Neiizo please open a new issue and report everything the template encourages you to report. Windows is really different. g++ isn't used for this on windows. python extensions are built...

This looks like the version of pytorch used at pytorch3d build time is different from the version used at runtime. Could environments have got confused somehow? Or could you have...

Can you try replacing the line ``` faces = faces.type(torch.int32) ``` with ``` faces = faces.type(torch.int64) ``` ?

Can you paste your current code? The verts need to be float32, the faces int64

pytorch3d is not installed at all. It's not a version problem. Note that installing pytorch3d on windows is a bit of a pain. What is the output of `conda list`?

If you want to use `normals_packed()` to return the value calculated by a previous `estimate_nomals()`, you just need to set `assign_to_self=True` in that call to `estimate_nomals()`. If you don't set...

Please paste the output of `conda list`.