Fast CNN Tuning with AWS GPU Instances and SigOpt | AWS AI Blog

Date: May 22, 2017 Published by

  • SigOpt can significantly speed up and reduce the cost of this tuning step compared to standard hyperparameter tuning approaches like random search and grid search.
  • As for cost, model training on the NVIDIA-K80-enabled p2.xlarge AWS instances (90¢/hour was $0.05 per iteration, and a bit over $2.50 on the m4.4xlarge CPU instance (86.2¢/hour).
  • Although you need domain expertise to prepare data, generate features, and select metrics, you don’t need special knowledge of the problem domain for hyperparameter tuning.
  • Choosing these parameters, fitting the model, and determining how well the model performs is a time-consuming, trial-and-error process called hyperparameter optimization or, more generally, model tuning.
  • Black-box optimization tools like SigOpt increase the efficiency of hyperparameter optimization without introspecting the underlying model or data.