MLX Apple Silicon Skill

"Unified memory means no GPU↔CPU transfers - arrays live in shared memory."

Trit: +1 (PLUS - generative) Color: Warm (optimistic/fast)

Overview

MLX is Apple's ML framework for Apple Silicon:

• Unified Memory: No GPU↔CPU data transfers
• Lazy Evaluation: Compute only what's needed
• Metal Backend: Native GPU acceleration
• 4-bit Quantization: 75% smaller models

MLX-LM provides high-level LLM APIs.

Quick Start