/deploy-python-ml

Primary LanguagePython

DEPLOY-PYTHON-ML

Basic machine learning (ML) project to illustrate steps to deploying to AWS Lambda using Docker.

This repo is part of the YouTube video How to Deploy a Python Machine Learning App using Docker + AWS Lambda

Youtube logo

Data Science Process

Some super basic data science/machine learning project. I basically went on Kaggle, looked up "nlp dataset" and downloaded the first one with sufficient data. Then I copied and pasted old code on here.

How do I build the model?

Make sure you have conda installed. Then, install the requirements with bash install_requirements.sh.

Finally, you can build the model by running python build_model.py on the terminal.

How do I test the app?

RUNNING_LOCAL=True python -m app.main or RUNNING_LOCAL=True FILENAME=model-dev/data/emotion-labels-test.csv python -m app.main where FILENAME is any file with text in the header.

Directory structure

deploy-python-ml/
├── app
│   ├── model
│   ├── preprocessing
└── model-dev
    └── data

Deployment

General steps to deploy your code include

  1. Build your machine learning model and pipeline
  2. Create/setup a AWS account
  3. Package your code in a Docker container
  4. Upload your Docker image to AWS Elastic Container Registry (ECR)
  5. Create your AWS Lambda to run the ECR image
  6. Run/test/configure your AWS Lambda
  7. Deliver your results to others who may need the results

Deployment Process