https://github.com/thomd/spam-detection-with-fastapi-on-azure

Train a spam-detection model with SciKit-Learn and deploy as FastAPI app on Azure Container Instance
https://github.com/thomd/spam-detection-with-fastapi-on-azure

azure docker machine-learning python

Last synced: 26 days ago
JSON representation

Train a spam-detection model with SciKit-Learn and deploy as FastAPI app on Azure Container Instance

• Host: GitHub
• URL: https://github.com/thomd/spam-detection-with-fastapi-on-azure
• Owner: thomd
• Created: 2021-11-21T16:08:34.000Z (over 4 years ago)
• Default Branch: main
• Last Pushed: 2021-11-24T09:09:28.000Z (over 4 years ago)
• Last Synced: 2025-02-05T21:42:34.754Z (over 1 year ago)
• Topics: azure, docker, machine-learning, python
• Language: Python
• Homepage:
• Size: 213 KB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Spam Detection with SciKit-Learn, FastAPI on Azure Container Instance

## Dataset

The UCI [SMS Spam Collection](https://data.world/uci/sms-spam-collection) is a public

set of SMS labeled messages that have been collected for mobile phone spam research.

## Setup

Create environment:

    conda create --name spam

    conda activate spam

    conda install --yes --file requirements.txt

Train spam classifier model:

    python train.py

For local testing, start server and open SwaggerUI

    uvicorn app:app --reload

    open http://localhost:8000/docs

Test local endpoint:

    curl http://localhost:8000/predict?message=I want your Money

## Build Docker Image

Build docker image and run docker container:

    docker build -t spam .

    docker run -d --rm --name spam -p 80:80 spam

## Deploy as Azure Container Instance

Use Azure Container Instances (ACI) to run serverless Docker containers in Azure.

Create Azure Container Registry (ACR):

    az login

    az group create -n  -l 

    az acr create -g  -n  --sku Basic --admin-enabled true

    az acr list -g  -o table

Upload docker image to ACR

    docker tag spam .azurecr.io/spam:v1

    az acr login -n 

    docker push .azurecr.io/spam:v1

Create container (find username and password here: Portal > Container Registry > Access Keys)

    az container create -g  -n  --image .azurecr.io/spam:v1 --dns-name-label  --ports 80

    az container show -g  -n  --query "{FQDN:ipAddress.fqdn,ProvisioningState:provisioningState}"

    az container logs -g  -n 

Open SwaggerUI

    

    open http://..azurecontainer.io

Delete resource group

    

    az group delete -n