Green AI in Fine-tuning Large Language Models with Adaptive Backpropagation

The author presents GreenTrainer, a technique that minimizes FLOPs in LLM fine-tuning via adaptive backpropagation, achieving significant reductions without compromising accuracy.