Description:
Array framework for Apple silicon
|
|
Type: Formula
|
Latest Version: 0.30.0@0
|
Tracked Since: Dec 18, 2025
|
|
Links:
Homepage |
formulae.brew.sh
|
|
Category: Ai ml
|
|
Tags:
machine-learning
apple-silicon
array-framework
ml
python
|
|
Install:
brew install mlx
|
About:
MLX is an Apple silicon-optimized array framework designed for efficient machine learning model development. It provides a NumPy-like interface that leverages the Unified Memory architecture of Apple chips for zero-copy operations between CPU and GPU. The framework supports automatic differentiation and dynamic graph construction, enabling high-performance training and inference without data transfer overhead.
|
Key Features:
- Unified Memory Architecture: Zero-copy operations between CPU and GPU
- Familiar API: NumPy-like syntax for easy adoption
- Automatic Differentiation: Built-in gradient computation for ML models
- Apple Silicon Optimization: Native performance on M-series chips
- Dynamic Graphs: Flexible computation graph construction
|
Use Cases:
- Developing and training machine learning models on Apple Silicon
- High-performance numerical computing without GPU memory transfer overhead
- Research prototyping with Python-native ML frameworks
|
Alternatives:
-
PyTorch
– Cross-platform but lacks unified memory optimization for Apple Silicon
-
TensorFlow
– Broader ecosystem but higher overhead on Apple hardware
-
NumPy
– General-purpose array computing without ML-specific features
|
| Detected |
Version |
Rev |
Change |
Commit |
| Dec 18, 2025 2:42am |
0.30.0 |
0 |
VERSION_BUMP |
dc3b0187 |
| Sep 26, 2025 11:42pm |
|
0 |
VERSION_BUMP |
00610835 |
| Sep 16, 2025 6:00am |
|
0 |
VERSION_BUMP |
bc833b69 |
| Sep 12, 2025 4:25am |
|
0 |
VERSION_BUMP |
8028fc5d |
| Jan 10, 2025 6:09am |
|
0 |
VERSION_BUMP |
2683a9d3 |
| Nov 3, 2024 2:07pm |
|
0 |
VERSION_BUMP |
3bd76b18 |
| Oct 31, 2024 6:38am |
|
0 |
VERSION_BUMP |
a0233ffe |
| Oct 31, 2024 1:15am |
|
0 |
VERSION_BUMP |
c0d64e03 |
| Oct 30, 2024 10:03am |
|
0 |
VERSION_BUMP |
c40a022a |
| Oct 28, 2024 10:15pm |
|
0 |
VERSION_BUMP |
f8c33faf |
| Oct 11, 2024 5:53pm |
|
0 |
VERSION_BUMP |
27fffe3d |
| Sep 20, 2024 4:22pm |
|
0 |
VERSION_BUMP |
0dcf9d3a |
| Sep 14, 2024 4:32am |
|
0 |
VERSION_BUMP |
a5221c31 |
| Sep 13, 2024 3:11am |
|
0 |
VERSION_BUMP |
11190c79 |
|