2023-11-29 12:54:28 -08:00
2023-11-29 10:52:08 -08:00
2023-11-29 10:52:08 -08:00
2023-11-29 12:45:41 -08:00
2023-11-29 10:52:08 -08:00
2023-11-29 12:38:32 -08:00
2023-11-29 10:52:08 -08:00
2023-11-29 10:52:08 -08:00
2023-11-29 10:30:41 -08:00
2023-11-29 10:42:59 -08:00
2023-11-29 10:42:59 -08:00
2023-11-29 10:42:59 -08:00
2023-11-28 15:33:45 -08:00
2023-11-29 10:30:41 -08:00
2023-11-29 10:30:41 -08:00
2023-11-29 10:52:08 -08:00
2023-11-29 10:52:08 -08:00

MLX

MLX is an array framework for machine learning specifically targeting Apple Silicon. MLX is designed with inspiration from Jax, PyTorch, ArrayFire.

Documentation

Build

mkdir -p build && cd build
cmake .. && make -j

Run the C++ tests with make test (or ./tests/tests for more detailed output).

Python bidings

To install run:

env CMAKE_BUILD_PARALLEL_LEVEL="" pip install .

For developing use an editable install:

env CMAKE_BUILD_PARALLEL_LEVEL="" pip install -e .

To make sure the install is working run the tests with:

python -m unittest discover python/tests

Contributing

Check out the contribution guidelines for more information on contributing to MLX.

Description
MLX: An array framework for Apple silicon
mlx
Readme MIT 466 MiB
Languages
C++ 65.3%
Python 22.5%
Cuda 6.6%
Metal 3.7%
CMake 1%
Other 0.8%