I’m trying to infer a sparse model on TX2 https://github.com/NVIDIA/MinkowskiEngine , but it fails and causes the error discussed on https://forums.developer.nvidia.com/t/tx2-gpu-obsolete-for-pytorch/158330?u=sathya . Any help or guidance?

Train/Infer Sparse Models on TX2?

AastaLLL November 2, 2020, 3:06am 3

Hi,

Please check this comment to add the TX2 GPU architecture and recompile the library.

Thanks.