Segmentation fault on Ubuntu 24.04.3

[Azmodeszer]

I'm trying to run exo on a Ubuntu VM, but I don't get very far:

/home/dummy/exo/.venv/lib/python3.12/site-packages/google/protobuf/runtime_version.py:112: UserWarning: Protobuf gencode version 5.27.2 is older than the runtime version 5.28.1 at node_service.proto. Please avoid checked-in Protobuf gencode that can be obsolete.
  warnings.warn(
None of PyTorch, TensorFlow >= 2.0, or Flax have been found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
None of PyTorch, TensorFlow >= 2.0, or Flax have been found. Models won't be available and only tokenizers, configuration and file/data utilities can be used.
Segmentation fault (core dumped)

The VM has four AMD MI50 GPUs passed through from the hypervisor and ROCm is installed; GPUs are recognized fine and compute works like a charm. I realize this is a pretty generic error and not a lot to go on, so please let me know which specs and logs I should share.

Seems to be related to the distro at least tangentially? I whipped up a quick Debian (Trixie) VM on the same host and exo works fine there.

[merll002]

Same here on arch. Tried the latest tag, and fork as well.