Radio: Rate–Distortion Optimization for Large Language Model Compression

Sean I. Young

TL;DR: A rate distortion-theoretic framework leads to improved LLM compression outcomes.

Abstract: In recent years, the compression of large language models (LLMs) has emerged as a key problem in facilitating LLM deployment on resource-limited devices, reducing compute costs, and mitigating the environmental footprint due to large-scale AI infrastructure. Here, we establish the foundations of LLM quantization from a rate–distortion theory perspective and propose a quantization technique based on simple rate–distortion optimization. Our technique scales to models containing hundreds of billions of weight parameters and offers users the flexibility to compress models, post-training, to a model size or accuracy specified by the user.

Lay Summary: In recent years, the compression of large language models (LLMs) has emerged as a key problem in facilitating LLM deployment on resource-limited devices, reducing compute costs, and mitigating the environmental footprint due to large-scale AI infrastructure. Here, we establish the foundations of LLM quantization from a theory perspective and propose a quantization technique based on simple optimization. Our technique scales to models containing hundreds of billions of weight parameters and offers users the flexibility to compress models, post-training, to a model size or accuracy specified by the user.

Link To Code: https://github.com/seannz/radio

Primary Area: Theory->Optimization

Keywords: rate–distortion theory, optimization, compression, quantization

Submission Number: 508