HYPPO: A Surrogate-Based Multi-Level Parallelism Tool for Hyperparameter Optimization
We present a new software, HYPPO, that enables the automatic tuning of hyperparameters of various deep learning (DL) models. Unlike other hyperparameter optimization (HPO) methods, HYPPO uses adaptive surrogate models and directly accounts for uncertainty in model predictions to find accurate and reliable models that make robust predictions. Using asynchronous nested parallelism, we are able to significantly alleviate the computational burden of training complex architectures and quantifying the uncertainty. HYPPO is implemented in Python and can be used with both TensorFlow and PyTorch libraries. We demonstrate various software features on time-series prediction and image classification problems as well as a scientific application in computed tomography image reconstruction. Finally, we show that (1) we can reduce by an order of magnitude the number of evaluations necessary to find the most optimal region in the hyperparameter space and (2) we can reduce by two orders of magnitude the throughput for such HPO process to complete.
- Research Organization:
- National Renewable Energy Lab. (NREL), Golden, CO (United States)
- Sponsoring Organization:
- USDOE Office of Science (SC), Advanced Scientific Computing Research (ASCR)
- DOE Contract Number:
- AC36-08GO28308
- OSTI ID:
- 1847268
- Report Number(s):
- NREL/CP-2C00-82274; MainId:83047; UUID:36199796-1b14-4a6b-a253-0090fd01b2be; MainAdminID:63962
- Resource Relation:
- Conference: Presented at the 2021 IEEE/ACM Workshop on Machine Learning in High Performance Computing Environments (MLHPC), 15 November 2021, St. Louis, Missouri
- Country of Publication:
- United States
- Language:
- English
Similar Records
DeepHyper: Asynchronous Hyperparameter Search for Deep Neural Networks
DeepHyper: Asynchronous Hyperparameter Search for Deep Neural Networks