Inference & Serving · 2025

LoRA Without Regret

John Schulman

Shows that LoRA applied to all layers with the right hyperparameters (notably a learning rate ~10x higher than full fine-tuning) matches full fine-tuning's learning efficiency and final quality across typical post-training regimes.

Editorial record

Plain-language summary

Shows that LoRA applied to all layers with the right hyperparameters (notably a learning rate ~10x higher than full fine-tuning) matches full fine-tuning's learning efficiency and final quality across typical post-training regimes.

Knowledge graph

Relationships

Descendants

Source record

Provenance

Record ID
P-657
Record created
2026-07-13
Last reviewed
2026-07-14
Record version
2

Citation caveat: Citation metadata is approximate and marked unverified in the source dataset.