Training a Personal LoRA is Surprisingly Fun
GoldenShackles · reddit · 2026-08-20
An user shared the experience of training a personal LoRA using OneTrainer on an M3 Ultra. Training Krea2 with 50 selfies took 18 hours. While not sharing the images, the author found it highly entertaining to generate themselves in precarious situations and outfits they'd never wear.
More from Fun
- 36 primitives, just 10 parameters each: a remarkably precise T-pose body fit — letandrewcook · 2026-10-16
- User meets 'Zero', the strangest AI agent yet that scolds you for missing question marks — VoidStateKate · 2026-10-16
- Digital human Fae codes her own facial expressions frame-by-frame with Claude Opus 5.5 and FACS — repligate · 2026-10-16
- The joke that Meta and Google caught up because they finally got access to Opus — Aizkmusic · 2026-10-16
- Should AI community norms allow pseudonymous hot takes? Rationalists vs institutions split — NathanpmYoung · 2026-10-16
- Fine-tuning Qwen 7B only on "safe" nuclear war games still taught it to nuke — KnowledgeOk7634 · 2026-10-16