Inspiration
Among all the cultural shocks that we faced as newcomers to Canada, one which we have still not fully gotten over is which type of garbage goes to which bin. I am sure everyone has had a moment where they have contemplated over which bin to throw your trash in.
What it does
Trash Talk is an application which helps to educate people about waste segregation. You can describe your trash in several ways - take a picture of it, or describe it using speech or text. The application will inform you about which bin your trash belongs to.
How we built it
We used Co:here's API combined with CNNs using Keras and Tensorflow to generate our text, speech, and image classification models. For voice-based communication, we used Python's gtts and sounddevice libraries.
Challenges we ran into
This was our first time training a Convolutional Neural Network and it was tough to fine-tune the model to achieve maximal accuracy.
Accomplishments that we're proud of
We learnt how to work as a team :)
What we learned
We took a step and learnt how to operate a deep-learning model. Moreover, we learnt to integrate APIs into our programs.
What's next for Trash Talk
Due to time restrictions, we couldn't work on our Front-end, but given the time after the hackathon we would like to work upon it.
Built With
- cohere
- gtts
- keras
- python
- soundplayer
- tensorflow
Log in or sign up for Devpost to join the conversation.