Hi @m.tharani.193, Jetson is single-GPU architecture and doesn’t support NCCL:
You can recompile PyTorch with USE_DISTRIBUTED enabled (try mounting additional swap if it hangs), but it will use MPI instead of NCCL.
Hi @m.tharani.193, Jetson is single-GPU architecture and doesn’t support NCCL:
You can recompile PyTorch with USE_DISTRIBUTED enabled (try mounting additional swap if it hangs), but it will use MPI instead of NCCL.