The project attempts to classify a given image as clean or messy. This is an exploratory project wherein we tried to explore various approaches for training and deploying a Machine Learning based binary classification model to be used by an end-user.
We used the following approaches to deploy the project:
- A web app with React frontend and Flask REST API
- A web app with a tensorflowjs model
- A web app with python based framework - Streamlit
React Web app - Click Here
Streamlit Web app - Click Here
- The data for this project was Scrapped from google Images using Selenium and Python. Check out the Scrapper readme for details regarding implementation.
- Also used a publicly available kaggle dataset
We trained a binary classification model in Keras and Fastai and then ported Keras model to tfjs, tflite and coreml. The jupyter notebooks are present here.
The model is deployed using three major approaches :
- Web App - The web-app based on Flask REST API and frontend using React.
- Web App with offline model - The web app based on the tensorflow js for offline classification.
- Web App - The web app based on Streamlit framework.
The apps were deployed on HEROKU. Heroku deployment instructions are available here.