https://github.com/shankarj67/Water-classifier-fastai

Deploy your Flask web app classifier on Heroku which is written using fastai library.
https://github.com/shankarj67/Water-classifier-fastai

classifier deep-learning fastai flask flask-api flask-web python3 pytorch resnet

Last synced: 2 months ago
JSON representation

Deploy your Flask web app classifier on Heroku which is written using fastai library.

• Host: GitHub
• URL: https://github.com/shankarj67/Water-classifier-fastai
• Owner: shankarj67
• License: mit
• Created: 2018-12-03T17:51:21.000Z (about 6 years ago)
• Default Branch: master
• Last Pushed: 2019-11-19T02:06:33.000Z (about 5 years ago)
• Last Synced: 2024-08-03T22:14:15.744Z (6 months ago)
• Topics: classifier, deep-learning, fastai, flask, flask-api, flask-web, python3, pytorch, resnet
• Language: HTML
• Homepage: https://water-classifier1.herokuapp.com/
• Size: 75.9 MB
• Stars: 40
• Watchers: 3
• Forks: 50
• Open Issues: 6
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

• awesome-fastai - Water Classifier
• awesome-fastai - Water Classifier

README

        

# Deploy water classifier using flask on heroku

[![](https://img.shields.io/badge/python-3.5%2B-green.svg)]()

> A web app to predict whether water is clean or dirty using fastai library.

------------------

Demo:



  





  



------------------

## Getting started

- Fork and clone this repo on to your system

## Customization

### Put your model inside path/models folder

- Place your trained `.pth` file under path/models/ directory.

### Change the class and path name

- Open "app.py" and search for a variable called classes and change that with your own classes

- If your path name is different that "stage-2.pth", then change the learn.load("with_your_file_name.pth") in "app.py"

### UI Change

Modify files in `templates` and `static` directory.

`index.html` for the UI and `main.js` for all the behaviors

## Heroku Installation

- Make sure that heroku is installed on your system, if you want to install then follow this https://devcenter.heroku.com/articles/heroku-cli#download-and-install

- Go to https://www.heroku.com/ then login or signup into heroku.

- Once you are logged in, you will have to click on new and click on create a new app as given in the screenshot below



  



- Give your app name and click on create app

- Go to the folder where you have cloned my repo and follow the steps given on the page

- Once you are done, you will see the link to check your app on browser like this https://water-classifier1.herokuapp.com/

> Note: If App is showing Application error then refresh the browser,It will work fine.

## Future Roadmap

- Add unit test.

- Add CI/CD integration.

- Create a robust image classifier using the more data or using some advance technique.

- Adding celery and redis as a message brocker to handle the long task.