NVIDIA Developer Forums

Using Genetic Algorithms for Hyperparameter Search is a real thing #DidYouKnow

AI & Data Science Data Science of the Day

btegegn January 28, 2021, 2:00pm 1

RAPIDS cuML trains models so fast that we can use genetic algorithms to optimize hyperparameters. Make a population of models and select the fittest model. Then combine them with mutations to make another generation of models.
Click the image to read the article
Find more #DSotD posts

Topic		Replies	Views	Activity
Hyperparameter tuning with RAPIDS on GPUs yields higher accuracy models #HPO Data Science of the Day python , fun-facts , rapids , data-science , machine-learning	0	1147	January 19, 2021
Deep Learning Hyperparameter Optimization with Competing Objectives Technical Blog	1	364	June 18, 2019
Hyperparameter Optimization TAO Toolkit	5	1428	November 30, 2021
PyTorch, MLflow & Optuna: Experiment Tracking and Hyperparameter Optimization Data Science of the Day pytorch , fun-facts , machine-learning	0	1323	November 19, 2021
Deep Learning Hyperparameter Optimization with Competing Objectives Technical Blog	0	234	August 21, 2022
Hyper-parameter auto-tuning~ TAO Toolkit	2	478	May 30, 2022
Learn about Combinatorial Optimization with Google OR-Tools Data Science of the Day python , fun-facts	0	1284	March 9, 2022
RAPIDS Is in a Constant State of Improvement. Benchmarking Provides the Metrics to Measure This Data Science of the Day ai , fun-facts	0	762	June 10, 2021
Genetic Algorithm CUDA Programming and Performance	28	32703	September 16, 2021
Stacking Generalization with HPO: Maximize Accuracy in 15 Minutes with NVIDIA cuML Technical Blog	1	33	May 1, 2025