How to optimize a Machine Learning Model?

Optimize a machine learning model with SageMaker Neo

Documentation

SageMaker Neo enables to tune ML model once and uses the model to run anywhere in the cloud and edge. It optimizes the model for increased speed and less memory consumption. Neo provides the compilation of trained models from different frameworks. The following figure shows a summary of Neo compilation of trained models.

sagemaker-diagram

The pre-trained model formats to be uploaded for compilation should be

enter image description here

Neo compilation from SageMaker notebook instance

  1. Create or Open a Notebook instance. Please follow steps 1-4 in Amazon SageMaker to create a new notebook instance.

  2. Model optimization for Person detection model can be downloaded from SageMaker Neo compilation example. Please refer to How to work with notebooks in Jupyterlab to get an overview of Jupyterlab.

  3. Run the notebook cells for Neo compilation. Once the notebook cells have been run, the optimized model can be downloaded from the S3 bucket specified in the notebook.

In the SageMaker → Compilation Jobs page, we can see the current status of our job.

neo_menu

Select the desired Compilation job from the list by clicking on the name. Here we have the job details.

details

MX-ssd

Info

Project State

Public Project

Licences

Software Licence: Project has no software
Hardware Licence: Project has no hardware

Project Tags

Admins

SelenaS
emeusel
Nina_Boehm
MeyerMel
nwilson
jstoltz
kilian-hohm
TomE
abd

Members

Does this project pique your interest?

Login or register to join or follow this project.

Comments
Back to top

Ready to join the project?

You'd like to participate ... Show more