Cost-Effective Hyperparameter Optimization for Large Language Model Generation Inference

Chi Wang, Xueqing Liu, Ahmed Hassan Awadallah

Keywords: Large Language Model, Text Generation, Hyperparameter Optimization

TL;DR: The first study of holistically optimizing the hyperparameters for text generation inference using large language models under budget constraints

Abstract: Large Language Models (LLMs) have sparked significant interest in their generative capabilities, leading to the development of various commercial applications. The high cost of using the models drives application builders to maximize the value of generation under a limited inference budget. This paper presents a study of optimizing inference hyperparameters such as the number of responses, temperature and max tokens, which significantly affects the utility/cost of text generation. We design a framework named EcoOptiGen which leverages economical hyperparameter optimization and cost-based pruning. Experiments with the GPT-3.5/GPT-4 models on a variety of tasks verify its effectiveness. EcoOptiGen is implemented in the "autogen" package of the FLAML library: https://aka.ms/autogen.

Submission Checklist: Yes

Broader Impact Statement: Yes

Paper Availability And License: Yes

Code Of Conduct: Yes

Reviewers: Yes

CPU Hours: 0

GPU Hours: 0

TPU Hours: 0