Inspiration
We were inspired to build this project after accumalating trash in our workspace during the first few hours of this hackathon.
What it does
We developed an app which you can upload or take a photo of any disposable item and it will tell you what kind of disposable it is and the action you should take to dispose of it.
How we built it
We built this app using React Native for the frontend, Python Flask for the backend and we made use of a Hugging Face waste image classification model to process our image data.
Model Link: [link]https://huggingface.co/watersplash/waste-classification
Challenges we ran into
We ran into trouble when actually trying to use this Hugging Face image classification model as we were unfamiliar with Hugging Face technologies and the 'Transformers' Python Library it relied on. Our main issue was something that took us a lot of time to realize - the dependencies and other Python libraries needed to import the Transformers module were rather specific. Additionally, we ran into alot of difficulty trying to connect our backend in Python Flask to our front end in React Native - we were unable to transfer our image data correctly between the front end and back end.
Accomplishments that we're proud of
As we had many challenges, we are proud of the ways we overcame these challenges in the given time constraint. We overcame our challenge of being unable to use the Transformers library by using a Python virtual environment for our project which allowed us to install the specific Python version and libraries needed for this technology. Then, we overcame the challenge of moving data between our frontend and backend by learning more about Flask and React Native - technologies which were rather new for us - and we were eventually led to a library called Axios which we used to take care of this issue.
What we learned
We learned about Python Virtual Environments, the how to use AI models from the Hugging Face Transformers Python Library, and how to use React Native/Python Flask and how to connect them.
What's next for Trash Ninja
For further improvement of our app, we could use Google Maps or some other Map API to provide users with the option to view their nearest trash emptying, clothing donation or battery disposal sites.
Built With
- flask
- javascript
- python
- pytorch
- react-native
- tensorflow
- transformers
Log in or sign up for Devpost to join the conversation.