deployment

PEFT

Parameter-Efficient Fine-Tuning, a family of techniques that adapt large models by training only a small subset of parameters. PEFT methods include LoRA, adapters, and prefix tuning, making fine-tuning accessible with limited compute resources.