Hacker News new | ask | show | jobs
LISA: Layerwise Importance Sampling for Memory-Efficient LLM Fine-Tuning (arxiv.org)
3 points by convexstrictly 810 days ago
1 comments
convexstrictly 810 days ago
An author claims better performance than LoRA in 50% of the time.

https://twitter.com/Rui45898440/status/1772996453557997924

link