Top all-time · #11692
ByteShape
Company blog · byteshape.com/blogsaihardware
ByteShape is the company blog at byteshape.com, writing about ai and hardware. Since 2026, 6 different Hacker News readers have linked it 7 times, most often “Shapelearn Qwen 3.8 27B (13.1 GB VRAM)”. No feed was found, so read it on the site itself.
Too few sampled comments (4) to say how often HN recommends it.?
Why HN links it
HN readers link ByteShape's posts on compressing Qwen models to run on small hardware, such as a Raspberry Pi and a 16GB GPU. Commenters cite its methodology and its quant comparisons.
Summarised by Claude from a sample of 4 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
ByteShape optimizes language, vision-language, and image-generation models for specific hardware and deployment requirements.
Most-linked posts?
- Shapelearn Qwen 3.8 27B (13.1 GB VRAM)2 people · 2 mentions · first linked 2026-09-18
- A 30B Qwen Model Walks into a Raspberry Pi and Runs in Real Time2 people · 2 mentions · first linked 2026-01-06
What Hacker News says?
I run qwen 27b on an old-ass 16gb gpu. It’s very possible using unsloth 2bit and 3bit quants, tho there are a bunch of interesting quants that let you run closer to 4bit on 16gb. This article that’s currently also on the front page mentions a bunch of them while discussing their own quant https://byteshape.com/blogs/Qwen3.8-27B/
Comment linking /blogs/Qwen3.8-27B/
on “Bonsai 2 27B: Near-Lossless Compression in a 9x Smaller Footprint”Submitted Shapelearn Qwen 3.8 27B (13.1 GB VRAM) as a story · 104 points
HN discussion> Would also love to better understand what factors into "accuracy" since there might be some nuance there depending on the measure. It's accuracy across GSM8K, MMLU, IFEVAL and LiveCodeBench. They detail their methodology here: https://byteshape.com/blogs/Qwen3-4B-I-2507/
Comment linking /blogs/Qwen3-4B-I-2507/
on “A 30B Qwen Model Walks into a Raspberry Pi and Runs in Real Time”Submitted A 30B Qwen Model Walks into a Raspberry Pi and Runs in Real Time as a story · 2 points
HN discussionIt's accuracy across GSM8K, MMLU, IFEVAL and LiveCodeBench. They detail their methodology here: https://byteshape.com/blogs/Qwen3-4B-I-2507/
Comment linking /blogs/Qwen3-4B-I-2507/
on “A 30B Qwen Model Walks into a Raspberry Pi and Runs in Real Time”Submitted Every Hardware Deserves a Coder: Devstral Small 2 24B and Qwen3 Coder 30B as a story · 1 points
HN discussionhttps://chatgpt.com/share/695d9ac2-c314-8011-8938-b0d7de7059... You can paste any article and chatgpt (took the most laymen AI thing) and just writing summarize this article https://byteshape.com/blogs/Qwen3-30B-A3B-Instruct-2507/ can give you insights about it. Although I am all for freedom, one forgets that this is one of the few places left on internet where discussions feel meaningful and I am not judging you if…
Comment linking /blogs/Qwen3-30B-A3B-Instruct-2507/
on “A 30B Qwen Model Walks into a Raspberry Pi and Runs in Real Time”