Installation fails on Jetson AGX Orin 64 GB

[Andymulb]

Hi, thank you for your great work! I already tested SLAM3R on an RTX 4080 SUPER. Installation and running the demos worked flawlessly. However, one of my own demo image sequences exhausted the VRAM, which resulted in a crash of SLAM3R. I'm searching for a SLAM system running on resource constrained hardware. I tried to install SLAM3R on an NVIDIA Jetson AGX Orin with 64 GB of unified memory. Since this machine has much more memory available, SLAM3R shouldn't crash, but unfortunately won't be used xformers. Did you get the chance to test your system on a Jetson and experienced this issue (facebookresearch/xformers#1193) as well? I already opened an issue in xformers repository.

I was able to install triton from source and also installed xformers from source. However, I receive the following error message when running bash python -m xformers.info:

Traceback (most recent call last):
  File "/home/anaconda3/envs/SLAM3R/lib/python3.10/runpy.py", line 196, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "/home/avsjetsonagx1/anaconda3/envs/SLAM3R/lib/python3.10/runpy.py", line 86, in _run_code
    exec(code, run_globals)
  File "/home/diffSLAM/dependencies/SLAM3R/thirdparty/xformers/xformers/info.py", line 11, in <module>
    from . import __version__, _cpp_lib, _is_opensource, _is_triton_available, ops
  File "/home/avsjetsonagx1/diffSLAM/dependencies/SLAM3R/thirdparty/xformers/xformers/ops/__init__.py", line 26, in <module>
    from .modpar_layers import ColumnParallelLinear, RowParallelLinear
  File "/home/diffSLAM/dependencies/SLAM3R/thirdparty/xformers/xformers/ops/modpar_layers.py", line 15, in <module>
    from .seqpar import sequence_parallel_leading_matmul, sequence_parallel_trailing_matmul
  File "/home/diffSLAM/dependencies/SLAM3R/thirdparty/xformers/xformers/ops/seqpar.py", line 10, in <module>
    from torch.distributed.distributed_c10d import _resolve_process_group
  File "/home/anaconda3/envs/SLAM3R/lib/python3.10/site-packages/torch/distributed/distributed_c10d.py", line 22, in <module>
    from torch._C._distributed_c10d import (
ModuleNotFoundError: No module named 'torch._C._distributed_c10d'; 'torch._C' is not a package```