Cost-Effective Hyperparameter Optimization for Large Language Model Generation Inference
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: \url{https://aka.ms/autogen}.
Cite
Text
Wang et al. "Cost-Effective Hyperparameter Optimization for Large Language Model Generation Inference." Proceedings of the Second International Conference on Automated Machine Learning, 2023. doi:10.48550/arXiv.2303.04673Markdown
[Wang et al. "Cost-Effective Hyperparameter Optimization for Large Language Model Generation Inference." Proceedings of the Second International Conference on Automated Machine Learning, 2023.](https://mlanthology.org/automl/2023/wang2023automl-costeffective/) doi:10.48550/arXiv.2303.04673BibTeX
@inproceedings{wang2023automl-costeffective,
title = {{Cost-Effective Hyperparameter Optimization for Large Language Model Generation Inference}},
author = {Wang, Chi and Liu, Xueqing and Awadallah, Ahmed Hassan},
booktitle = {Proceedings of the Second International Conference on Automated Machine Learning},
year = {2023},
pages = {21/1-17},
doi = {10.48550/arXiv.2303.04673},
volume = {224},
url = {https://mlanthology.org/automl/2023/wang2023automl-costeffective/}
}