ra2157218-boop/GEKO

Intelligent training framework that automatically skips mastered samples and gives 5× more compute to hard ones. Up to 80% compute savings on LLM fine-tuning.

1 best points · 1 comments · 1 Hacker News post · 1 GitHub stars · View on GitHub

Hacker News discussions

• Show HN: GEKO (up to 80% compute savings on LLM fine-tuning) (1 points, 1 comments, by SyedAbdurR2hman, 2026-02-28)