jax-graft: run JAX on your Apple GPU
A custom backend for JAX adding support for Apple GPUs.
Under the hood it treats Metal as an XLA:GPU platform (called "mtl"),
next to CUDA and ROCm. XLA's own GPU compiler fuses your program and
generates the kernels, which the plugin translates to Metal Shading
Language.
This project isn't affiliated with or endorsed by Apple or Google.
Status
jax-graft has wide coverage of the JAX test suite (with the notable
exception of float64 support), but you should expect some sharp edges.
The performance is broadly comparable to MLX and PyTorch's MPS backend
on the workloads we've tried. Please report issues.
Install
pip install jax-graft
Quick start
import jax
import jax.numpy as jnp
x = jnp.arange(4.0)
y = jax.jit(lambda v: v * 2)(x)
print(y, y.devices()) # [0. 2. 4. 6.] {MtlDevice(id=0)}
Just installing jax-graft is enough to set the default backend to
"mtl" on Macs with Apple GPUs. To make CPU the default, set the
JAX_PLATFORMS=cpu,mtl environment variable, and place work on the
GPU explicitly using jax.device_put(x, jax.devices("mtl")[0]).
Some more complete worked examples can be found in the
examples directory, and the docs directory
has a lot more information.
Contributing
Issues and pull requests are welcome. CONTRIBUTING.md
says what makes a useful report, and
docs/development.md covers building and testing.
License
Apache-2.0 (LICENSE). The plugin includes third-party code:
some kernels ported from MLX (MIT),
Apple's metal-cpp headers (Apache-2.0), and a statically linked XLA. Their
notices are in THIRD_PARTY_NOTICES, which ships in
both wheels.
Metadata
Release files for jax-graft 0.0.5
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Built distribution (wheel)
| File | Interpreter | ABI | Platform | Actions Reset |
|---|---|---|---|---|
| jax_graft-0.0.5-py3-none-any.whl | Python 3 | none | any | Details for jax_graft-0.0.5-py3-none-any.whl |
Release files / jax_graft-0.0.5-py3-none-any.whl
| Download URL | jax_graft-0.0.5-py3-none-any.whl |
|---|---|
| Size | 60.2 kB |
| Tags | Python 3 |
|
SHA-256 checksum How to use checksums |
b156b8040a6bd8b208b1c95b20729039e6df712c9efc74d4204e433ed2196742
|
|
BLAKE2b-256 checksum How to use checksums |
f04b3407d33e5eb51b1a8e0eb8203029023ca7908fbceed67d0abef8ece46d5d
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/7.0.0 CPython/3.13.14
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 11, 2026.
Transparency logSecurity for jax-graft 0.0.5
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
Full provenance
1 file in this release has been uploaded with provenance metadata.
Attestation for jax_graft-0.0.5-py3-none-any.whl
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Oct 11, 2026.
Transparency logA note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.