Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

interesting approach to tune models for readable text in monospace CLI terminals where basically +/- tune for this does or doesn't meet some fixed width view

could probably do view LoRA at some reasonable level likely also applications for time-series prediction, ARC-like or other puzzle solving, etc. especially if got into fixed height along with fixed width, terminal readable --more-- with good pagination paragraph separation, etc.

QWEN-40-char-terminal DiffusionGemma-80-char etc.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: