LoRA (Low-Rank Adaptation) fine-tunes large models by updating small rank decomposition matrices.
- True
- False
Answer: True
LoRA adds low-rank matrices to weight updates, reducing trainable parameters by 100-1000x vs full fine-tuning. Enables efficient adaptation of LLMs on consumer hardware. Critical for resource-constrained AI.