VRAM

LoRA Makes Fine-Tuning LLMs Practical

LoRA cuts LLM fine-tuning to a small adapter layer, reducing VRAM, training time, and cost for teams with modest GPUs.