Summary of On the Compressibility Of Quantized Large Language Models, by Yu Mao et al.
On the Compressibility of Quantized Large Language Models
by Yu Mao, Weilan Wang, Hongchao Du, Nan Guan, Chun Jason Xue
First submitted to arxiv on: 3 Mar 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 As machine learning educators, we’ll summarize a research paper on deploying Large Language Models (LLMs) on edge or mobile devices. The challenge lies in reducing the model size while maintaining good performance. Quantization is an effective method, but even after quantization, LLMs may still be too big to fit into limited memory. In this case, I/O latency becomes a bottleneck. This paper takes a preliminary step in studying data compression techniques to speed up inference on memory-constrained devices. We’ll explore the compressibility of quantized LLLMs and the trade-off between compressibility and performance. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This research paper is about making Large Language Models work on smaller devices like smartphones or computers. Right now, these models take up too much space and are slow because they need to be loaded from storage. The goal is to make them faster by using special techniques that shrink the model size while keeping its performance good. This is a first step in figuring out how to do this. |
Keywords
* Artificial intelligence * Inference * Machine learning * Quantization
