Summary of Cora: Optimizing Low-rank Adaptation with Common Subspace Of Large Language Models, by Xiaojun Xiao et al.
CoRA: Optimizing Low-Rank Adaptation with Common Subspace of Large Language Models
by Xiaojun Xiao, Sen Shen, Qiming Bao, Hongfei Rong, Kairui Liu, Zhongsheng Wang, Jiamou Liu
First submitted to arxiv on: 31 Aug 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed CoRA method optimizes Low-Rank Adaptation (LoRA) training by leveraging shared knowledge from large language models (LLMs). By freezing the substitute matrix B and halving parameters, CoRA achieves the same efficacy as original LoRA fine-tuning while being more efficient. Additionally, using the substitute matrix B as an enhanced initial state for the original matrix B leads to improved results with the same parameters. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary CoRA is a new way to make large language models work better without needing as much computer power. It does this by sharing knowledge between different parts of the model. The CoRA method has two steps: first, it freezes some parts of the model so that they don’t change during training, which saves time and energy. Then, it uses the frozen parts to help train the rest of the model, making it work even better. |
Keywords
» Artificial intelligence » Fine tuning » Lora » Low rank adaptation
