Bedtime Story Completion Model
Laura Madrid, Mateus Moreira, Michael Sheinman, and Katia Semyanovskaya
2022
My team and I created two models to assist writers in crafting short bedtime stories, for our Neural Networks & Deep Learning (CSC413) Final Project: a short-story generator that continues story prompts and a title generator using a sequence-to-sequence model. These models, fine-tuned from a pre-trained GPT-2 model and trained from scratch on an RNN for title generation, aim to help writers overcome writer’s block and enhance their speed in creating bedtime stories, particularly within the context of children’s books.
