Automatic1111 for Apple metal, 40% speed up sd1.5
dmikey831
65 points
21 comments
August 12, 2026
Related Discussions
Found 5 related stories in 138.0ms across 13,334 title embeddings via pgvector HNSW
- H3-metal – Native MiniMax-H3 inference for Apple Silicon swyx · 172 pts · August 11, 2026 · 50% similar
- Qwen3.8-27B on the iPhone 18 Neural Engine at 15–70 tok/s (1.5 bpw) anemll · 30 pts · October 17, 2026 · 50% similar
- DeepSeek v4.1 flash runs 23 seconds/token on a 2020 16gb M1 Mac Mini ajay-higgs · 14 pts · September 12, 2026 · 47% similar
- LFM2.5-DSpark: Up to 3.2x Faster Inference from H100 to MacB Alephinitesimal · 15 pts · August 21, 2026 · 47% similar
- Show HN: STE80 Humanizer – de-AI text by writing 80% of the way to ASD-STE100 Dodaios · 78 pts · November 09, 2026 · 47% similar
Discussion Highlights (5 comments)
jurgenburgen
Trigger warning: AI content.
vanillax
SD 1.5 is so insanely old at this point im not sure the point of this when you have Mini Max H3 on metal https://news.ycombinator.com/item?id=49252179
smallerize
The speedup includes running LoRAs, right? Which ones are you using?
IronWolve
Automatic1111 wasnt being updated, people moved to Forge Neo, ComfyUI and Maestro.
yogorenapan
A1111 is such a throwback. It's been ages since that & AI image generation was cool. Back before the massive amount of slop and when things were just a fun experiment