Closed 2226415223 closed 6 months ago
as mentioned above, is it a mistake from other diffrential project so you have to negtive all the x angle before pass to util.rotate_x, the following is the correct matrix
Thanks for highlighting this @2226415223 !
Yes, as you mention, this is an issue we inherited from nvdiffrast https://github.com/NVlabs/nvdiffrast/blob/main/samples/torch/util.py#L28
Fixed in CL a3e7390
as mentioned above, is it a mistake from other diffrential project so you have to negtive all the x angle before pass to util.rotate_x, the following is the correct matrix![image](https://github.com/NVlabs/nvdiffrec/assets/62322001/d21ffd3d-9a3c-45a1-949d-77141254dddd)