loom-py-rt-cuda

The CUDA backend for loom-py-rt. One shared library in one small wheel; there is nothing here to import.

pip install "loom-py-rt[cuda]"

import loom

loom.devices()                                   # a CUDA device now appears
model = loom.Model.from_file(path, device="gpu")

The base package is unchanged by installing this — it discovers the library on sys.path at import and device="auto" starts using it. That is what GGML_BACKEND_DL buys: no second copy of the runtime per accelerator.

If it appears not to have worked

Check loom.devices(). A backend whose driver is too old, or which finds no supported device, loads without error and registers nothing — the only other symptom is a model running at CPU speed.

device="gpu" asks for an offload device with its own memory, preferring one the kernel confirms is a GPU. It is not a promise that CUDA specifically was chosen. Pass device="CUDA0" to require this backend.

Versioning

This package pins the base with ==, not ~=. It carries a libggml-cuda.so that is dlopened beside the base wheel's libggml-base.so, and ggml makes no ABI promise across revisions — so a base release that moves its ggml pin invalidates every backend wheel published before it. The exact pin is what stops pip from pairing two libraries that do not agree.

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

loom_py_rt_cuda-1.0.0rc4-py3-none-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (75.3 MB view details)

Uploaded Aug 16, 2026 Python 3manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

loom_py_rt_cuda-1.0.0rc4-py3-none-manylinux_2_27_aarch64.manylinux_2_28_aarch64.whl (51.0 MB view details)

Uploaded Aug 16, 2026 Python 3manylinux: glibc 2.27+ ARM64manylinux: glibc 2.28+ ARM64

File details

Details for the file loom_py_rt_cuda-1.0.0rc4-py3-none-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.

File metadata

Download URL: loom_py_rt_cuda-1.0.0rc4-py3-none-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl
Upload date: Aug 16, 2026
Size: 75.3 MB
Tags: Python 3, manylinux: glibc 2.27+ x86-64, manylinux: glibc 2.28+ x86-64
Uploaded using Trusted Publishing? Yes
Uploaded via: twine/7.0.0 CPython/3.13.14

File hashes

Hashes for loom_py_rt_cuda-1.0.0rc4-py3-none-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl
Algorithm	Hash digest
SHA256	`7d268fcb22670bdc21ac6598c6697dde0cb52de3c496e8e16106f4ce2dc4d816`
MD5	`d1583f07ea0ebda7366b491c201c66b2`
BLAKE2b-256	`c8e5d578adf2c370535a28f7a00ff3faaf9f56e13fe333497aa07974f1af9152`

See more details on using hashes here.

Provenance

The following attestation bundles were made for loom_py_rt_cuda-1.0.0rc4-py3-none-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl:

Publisher: wheels.yml on loom-ai-org/loom-py

Attestations: Values shown here reflect the state when the release was signed and may no longer be current.

Statement:
- Statement type: https://in-toto.io/Statement/v1
- Predicate type: https://docs.pypi.org/attestations/publish/v1
- Subject name: loom_py_rt_cuda-1.0.0rc4-py3-none-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl
- Subject digest: 7d268fcb22670bdc21ac6598c6697dde0cb52de3c496e8e16106f4ce2dc4d816
- Sigstore transparency entry: 2481964542
- Sigstore integration time: Aug 16, 2026
Source repository:
- Permalink: loom-ai-org/loom-py@92d106bfcddb322dd03af22de6ed2474c53b71a5
- Branch / Tag: refs/tags/1.0.0-rc4
- Owner: https://github.com/loom-ai-org
- Access: public
Publication detail:
- Token Issuer: https://token.actions.githubusercontent.com
- Runner Environment: github-hosted
- Publication workflow: wheels.yml@92d106bfcddb322dd03af22de6ed2474c53b71a5
- Trigger Event: release