MLX
Apple’s framework for running and training models on Mac unified memory.
Fit.
Catalog snapshot
- Platforms
- Macos
- Operating systems
- macOS
- Ram floor
- 8 GB
- Gpu
- Optional · 8 GB VRAM helps
- Docker
- No
- Setup
- Intermediate
- Deploy as
- Local
- Commercial use
- Allowed
Unified memory counts RAM and VRAM together; 8 GB fits small models only.
Project.
Fetched 1
About.
MLX is Apple’s array framework for machine learning on Apple Silicon, and mlx-lm wraps it into local LLM chat, serving, and LoRA fine-tuning. It uses unified memory the way CUDA stacks cannot. Framework, not a finished product.
Metal kernels · unified memory · mlx-lm chat and serve · fine-tuning recipes
