Package Details: python-bitsandbytes 0.49.2-1

Git Clone URL: https://aur.archlinux.org/python-bitsandbytes.git (read-only, click to copy)
Package Base: python-bitsandbytes
Description: k-bit optimizers and matrix multiplication routines.
Upstream URL: https://github.com/bitsandbytes-foundation/bitsandbytes
Licenses: MIT
Conflicts: python-bitsandbytes
Provides: python-bitsandbytes
Submitter: taotieren
Maintainer: taotieren
Last Packager: taotieren
Votes: 1
Popularity: 0.060113
First Submitted: 2025-08-30 13:15 (UTC)
Last Updated: 2026-02-16 22:06 (UTC)

Latest Comments

lubosz commented on 2026-02-11 19:54 (UTC)

Only 3 months ago you have added python-pynvml as a new dependency.

I now get the following when importing torch (python -c "import torch"):

/usr/lib/python3.14/site-packages/torch/cuda/__init__.py:65: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
  import pynvml  # type: ignore[import]

Will bitsandbytes also work with python-nvidia-ml-py or does it need changes upstream?

taotieren commented on 2025-10-21 05:45 (UTC)

@lubosz Thanks. I'll update later.

lubosz commented on 2025-10-20 20:09 (UTC)

The native library needs to be rebuilt since CUDA 13. Please take my patch: https://github.com/lubosz/python-bitsandbytes/commit/2a4c38872a49d224a28a2e842d4132c3879d4c6d

Before my patch:

$ python -c "import bitsandbytes"
bitsandbytes library load error: Configured CUDA binary not found at /usr/lib/python3.13/site-packages/bitsandbytes/libbitsandbytes_cuda130.so
Traceback (most recent call last):
  File "/usr/lib/python3.13/site-packages/bitsandbytes/cextension.py", line 313, in <module>
    lib = get_native_library()
  File "/usr/lib/python3.13/site-packages/bitsandbytes/cextension.py", line 282, in get_native_library
    raise RuntimeError(f"Configured {BNB_BACKEND} binary not found at {cuda_binary_path}")
RuntimeError: Configured CUDA binary not found at /usr/lib/python3.13/site-packages/bitsandbytes/libbitsandbytes_cuda130.so
WARNING: Logging before InitGoogleLogging() is written to STDERR
I20251020 21:53:57.723321 140326377933760 SymmetricMemory.cpp:41] Destroying Symmetric Memory Allocators

After:

$ python -c "import bitsandbytes"
WARNING: Logging before InitGoogleLogging() is written to STDERR
I20251020 22:05:43.728649 139869615320000 SymmetricMemory.cpp:41] Destroying Symmetric Memory Allocators