Summary of Bisup: Bidirectional Quantization Error Suppression For Large Language Models, by Minghui Zou et al.
BiSup: Bidirectional Quantization Error Suppression for Large Language Models
by Minghui Zou, Ronghui Guo, Sai Zhang, Xiaowang Zhang, Zhiyong Feng
First submitted to arxiv on: 24 May 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
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 BiSup method addresses the challenges of efficient deployment of Large Language Models (LLMs) by introducing a bidirectional quantization error suppression technique. This approach optimizes the results of single matrix multiplication, accounting for vertical and horizontal error accumulation in LLMs. The model utilizes prompt mixed-precision quantization strategy to preserve high precision for key-value cache and suppresses error propagation through fine-tuning with a small amount of data. Experimental results on Llama and Qwen families demonstrate improved performance over state-of-the-art methods, enabling practical applications of low-bit weight-activation quantization. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Large Language Models (LLMs) are getting bigger and more important for everyday life. To make them work faster on devices like phones and computers, scientists have been working on a way to shrink the size of these models without losing their power. This is called “quantization.” A new method called BiSup helps solve some big problems with quantization by making sure errors don’t build up in the model. It does this by using a special kind of training that makes the model better at handling errors. The scientists tested this new method on different types of LLMs and found it worked really well, which is important for making these models useful for things like language translation and text summarization. |
Keywords
» Artificial intelligence » Fine tuning » Llama » Precision » Prompt » Quantization » Summarization » Translation
