Added preprocessing steps

[aravind-3105]

PR Type

Fix

Short Description

This pull request updates the documentation and configuration for the DVM-CAR dataset preparation and training pipeline in the multimodal representation learning project. The main focus is to streamline and clarify the data preparation process by consolidating previous scripts and notebooks into a single, unified Jupyter notebook (data_preprocessing.ipynb). Additionally, the configuration file now requires users to explicitly set their local data path.

Documentation and Workflow Improvements:

• Added a new notebook, data_preprocessing.ipynb, to guide users through downloading, cleaning, and preprocessing the DVM-CAR dataset, including tabular data processing, image conversion, and missing mask creation. This replaces references to multiple external scripts and notebooks, consolidating the workflow.
• Updated the README to reflect the new data preparation workflow, including step-by-step instructions for using data_preprocessing.ipynb for all preprocessing tasks.
• Clarified the order and description of the primary Jupyter notebooks in the project, now listing three: data preparation, training/fine-tuning, and evaluation/comparison.

Configuration Changes:

• Modified the data_base field in configs/config_dvm_TIP.yaml to require manual user input for the local data path, removing the hardcoded default.

Tests Added

...