Parameter-efficient fine-tuning

Compute interpretation

Adapter and low-rank update methods that reduce memory and optimizer-state costs during adaptation.

Supporting reading cards

• LoRA: Low-Rank Adaptation of Large Language Models (2021, efficient_edge_inference)
• QLoRA: Efficient Finetuning of Quantized LLMs (2023, efficient_edge_inference)

Obsolete or less central under later compute

Track this only through linked reading cards; do not treat this method page as standalone evidence.