NVIDIA-BioNeMo / NVIDIA-BioNeMo/Proteina-Complexa

[Proteina-Complexa] Fresh UV env installs a torchaudio incompatible with torch → import crash

Open Beginner friendly
#42 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
432
Forks
78
Avg merge
10d 46m
Merged PRs (30d)
1

Description

A fresh .venv built from build_uv_env.sh installs torchaudio 2.11.0 against torch 2.7.0. The ABI mismatch throws OSError: undefined symbol: torch_library_impl on first import, which crashes the very first complexa download.

Severity: High · Status: workaround verified; upstream open

Steps to reproduce
  1. Build the UV env from scratch (build_uv_env.sh).
  2. Run complexa download (or anything importing torch/torchaudio).
  3. OSError: undefined symbol: torch_library_impl.
Fix

Pin a compatible torchaudio:

uv pip install --torch-backend=cu126 --reinstall-package torchaudio "torchaudio==2.7.0"

build_uv_env.sh should pin torchaudio==2.7.0.

Environment
  • NVIDIA A100 80GB PCIe · driver 565.57.01 · CUDA 12.7
  • Repo: branch dev @ 916eaae
  • Runtime: UV (.venv + env.sh)

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start in build_uv_env.sh and reproduce the fresh UV environment setup described in the issue. Pin torchaudio to 2.7.0, then verify that complexa download or a torch/torchaudio import no longer raises the undefined-symbol error.

Written by the indexing model from the issue text.

Assessment

Tech stack
python, pytorch
Domain
build-system
Issue type
Bug
Difficulty
1/5
Estimated time
Under an hour
Activity status
Quiet
Clarity
Clearly specified
Newbie friendliness
88/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.