Articles by tag: apple-silicon

Local LLMs for Code: Privacy Without Speed Loss

How to Choose and Set Up Local LLMs for Code Generation. Comparison of Formats, Optimization for Apple Silicon, Integration with Agents. Keep Data Privacy Without Slowing Down Workflow.

FLUX.2-dev GGUF Q4_K_M: 29 GB VRAM on Apple Silicon

Breakdown of 29 GB VRAM consumption by FLUX.2-dev GGUF model on M3 Pro. MPS-overhead, dequantization, optimizations reduce peak by 1.5 GB. Instructions for developers.

SD 3.5 Medium on M1: optimization to 11.6 GB

Analysis of memory issues in Stable Diffusion 3.5 Medium on Apple Silicon. Sequential loading of T5 and DiT reduces peak from 21 to 11.6 GB. Code and benchmarks for M1 16 GB.