This project implements the deployment of a functional Machine Learning.
Need: a Machine Learning template to better organize and deploy a Machine Learning model.
Solution: My software is a suggestion on how to organize a Machine Learning experiment, with the capabilities of train
a Machine Learning Model. It also includes CI/CD on how to deploy training to AWS SageMaker.
This software does not requires specific hardwares. This project is based on Python 3.8, pipenv, Makefile, and Docker.
I provide the followin Makefile
target commands:
make data
- download the training data,make train
- train the machine learning model with the data and sabe the Machine Learning artifact,
This is a service software, the idea is to use its final result an API. If you intend to develop I provide detailed package requirements in the requirements.txt
file or Pipfile
and Pipfile.lock
for pipenv. Also Makefile
target make install
will install requirements using pipenv and docker.
% code samples
% config tips
% FAQs
% Bugs
% Community
% Explain How
The repo provides the capability of training and serving a sklearn
model capable of predicting divorce according to several features available for the "Divorce Predictors data set Data Set" dataset publicly available here.
After succesfully applying the deployment at EKS I obtained the following terminal screen: