Building Well-Architected Machine Learning Solutions on AWS with Phil Basford
ML Platform Podcast
00:00
The Cost of SageMaker
SageMaker, yeah, there's definitely a SageMaker tax involved. It's even on the instance size. We would typically run those kind of models in Lambda for a better different architectures or low volume architecture. The main weakness being lack of GPU and that's expensive in its own rights. But for most of our use cases we'd have gone that down that route.
Play episode from 27:58
Transcript
