https://arxiv.org/abs/2402.12354 LoRA+: Efficient Low Rank Adaptation of Large ModelsIn this paper, we show that Low Rank Adaptation (LoRA) as originally introduced in Hu et al. (2021) leads to suboptimal finetuning of models with large width (embedding dimension). This is due to the fact that adapter matrices A and B in LoRA are updated warxiv.org기존 LoRA가 A,B 모두 같은 학습률을 가졌다면 여기서 A,B는 다른 학습률을 가져..