I trained a small transformer in 1.5hrs and it beats many LLMs

·Hacker News··

You are using an outdated browser. Please upgrade your browser to improve your experience. Mithil Vakde's HomepageAbout MeBlogMisc Mithil VakdeIIT Bombay '23Engineering Physics Follow Bangalore, India Twitter LinkedIn44% on ARC-AGI-1 in 67 centsI trained a small transformer from scratch in 1.5hrs on a 5090 Beats many LLMs, and scores the same as TRM/HRMThis is an upgrade to my previous model Faster, better, cheaper and still open source.Also gets 7% on ARC-2Discussion on Twitter, Code on github

Read full article →

Related Articles

The ChatGPT/Codex app bundles a full copy of LibreOffice
timpera · Hacker News · 13h ago
Show HN: Running 104GB Qwen3.8-Flash-Next on 48GB Mac with at ~12 tok/s
carloslfu · Hacker News · 16h ago
The Emergent Symbolic Structure of Artificial Neural Networks
schmuhblaster · Hacker News · 5h ago
Refurbishing a Tektronix TDS7104 Oscilloscope
jwise0 · Hacker News · 13h ago
Omarchy: Any User Process Can Escalate to Root
trap0xcc · Hacker News · 2d ago