Final Project for LT2326 This repository contains the code, report, and presentation for the final project of LT2326.

Notebooks The project is divided into two main notebooks:

Notebook 1: image captions.ipynb

This notebook focuses on generating image captions using a custom model. It includes detailed documentation for most functions to aid in understanding the code and workflow. Notebook 2: VQA VizWiz.ipynb

This notebook explores Visual Question Answering (VQA) using the VizWiz dataset. Similar to the first notebook, it is well-documented to help follow the experiments. Usage Instructions The notebooks are intended to be run sequentially, as some dependencies and functions may carry over from one to another. In cases where two functions share the same name (e.g., preprocess_img), they represent different experimental setups. Be sure to execute only one version to avoid overwriting. Models The trained models could not be saved in github. If needed, please reach out to me directly to obtain the saved models.