MLCOST-04: Tradeoff analysis on custom versus pre-trained models - Machine Learning Lens
Implementation planDocumentsBlogs

MLCOST-04: Tradeoff analysis on custom versus pre-trained models

Optimize the cost through tradeoff analysis based on custom versus pre-trained models. This tradeoff analysis should keep the security and performance efficiency in perspective and within the acceptable thresholds.

Implementation plan

  • Use Amazon SageMaker built-in algorithms and AWS Marketplace - Amazon SageMaker provides a suite of built-in algorithms to help data scientists and machine learning practitioners get started on training and deploying machine learning models. Pre-trained ML models are ready-to-use models that can be quickly deployed on Amazon SageMaker. By pre-training the ML models for you, solutions in the AWS Marketplace take care of the heavy lifting, helping you deliver AI- and ML-powered features faster and at a lower cost. Evaluate the cost of your data scientists’ time and other resource requirements to develop your own custom model vs. bringing a pre-trained model and deploying it on SageMaker for inferencing. The advantage of a custom model is the flexibility to fine-tune it to match the needs of your business use case. A pre-trained model can be difficult to modify and you might have to use it as is.

  • Use Amazon SageMaker Jumpstart to access pre-trained models and accelerate the ML development process. SageMaker JumpStart provides a set of solutions for the most common use cases that can be deployed readily with just a few clicks. The solutions are fully customizable and showcase the use of AWS CloudFormation templates and reference architectures so you can accelerate your ML journey. Amazon SageMaker JumpStart also supports one-click deployment and fine-tuning of more than 150 popular open-source models such as natural language processing, object detection, and image classification models.

Documents

Blogs