Dataset from OpenBook QA

Contributors:

Haiyun Lyu has been responsible for the model training and connection for the frontend and backend. She used PyTorch with Transformer and Lora to fine-tune Llama-3 and Flask for the server. She outputs results from testing the chatbot. She wrote the conclusion, experiments and results, server and inference, and model training of method. Her model training version is used as this project’s final version.

Freya Liu has been responsible for the de- velopment of the front-end. She enabled user interaction through implementation of JavaScript functionalities and advocated, utilized FastAPI in this project for agile development. She wrote Introduction, Problem Definition, Related Work, API Encapsulation and Javascript Deliverables in Method, and part of User Feedback and Efficiency.

Ziyin Zheng has been responsible for filtering and cleaning the dataset, as well as transformations. She utilized pandas to clean and transform the dataset by simplifying questions and accurately extracting answers. She outputs for datasets that can be used for training. She wrote the data preparation and cleansing of method, and data section.