Add demo UI, token attention rollout & top5 table; clean ignores

fix model paths, retrain, perform inference

Auto-format code with Black

made formatting changes for CI pipeline

changes to incorporate text only, image only model modes

hyperparameter tuning with optuna, wandb

added init.py in src to make it a pkg, made changes in multimodal_model.py

Created a multimodal model to combine text and images