Inspiration
Empowering children's creativity using language models.
What it does
Generates and tells stories to children based on their interests.
How we built it
We used an Ionic-React Frontend paired with a Flask Server for the application. We used the Open-AI API to query and fine-tune the model, and the Google Cloud Speech Synthesizer API to tell the stories.
Challenges we ran into
It was challenging to combine the speech synthesizer and the Open-AI API, whilst providing an intuitive user interface for children.
Accomplishments that we're proud of
Producing novel engaging stories of up to 2-3 minutes in length.
What we learned
We learnt a great deal about how language models work.
What's next for Build-A-Story
Our next steps will include using pictures or animations generated by models such as DALL-E, to make the stories more engaging, and adding more options to allow our users to be more creative with their stories.
