Skip to content

pegasystems/nlp-model-containers

Repository files navigation

Table of contents

General info

This is an example dockerized container which comes with the dependencies required to run your machine learning and deep learning models

This uses flask to serve the models using a rest end point secured with Authlib for OAuth2 (using https://docs.authlib.org/en/latest/flask/2/) as well as an end point without auth.

Notes: Please note that it is a sample Docker Container. scripts can be modified and deployed with/without authentication.

The Repo contains 2 containers.

  • Machine Learning Container
  • Deep Learning Container

Please find the steps to be followed:

  • Step1: Train the model
  • Step2: Save the model
  • Step3: Dump the saved model into specific location
  • Step4: Initialize the Flask API via Docker container
  • Step5: Predict and access the endpoints

Note: Detailed steps are available in the Machine Learning and Deep Learning respective README.md files

Machine Learning: https://github.com/pegasystems/nlp-model-containers/tree/master/machine-learning-nlp-container

Deep Learning: https://github.com/pegasystems/nlp-model-containers/tree/master/deep-learning-nlp-container

Machine Learning NLP container

The project contains:

Algorithm Supported(Scikit/SKlearn):

  • Boosting Algos( XGBoost/LightBoost/AdaBoost/CatBoost)
  • Decision Tree
  • Random Forest
  • Naive Bayes(Multionomial and Gausian)
  • SVM
  • KNN

Deep Learning NLP container

The project contains:

Algorithm Supported(Keras Framework):

  • Tensorflow
  • Microsoft Cognitive Toolkit (CNTK)
  • Theano