Kaggle Competition Starter Notebook

Welcome to our 2025 Kaggle Competition of AI Applied to Medicine at UC3M repository!

This GitHub repository contains the starter Jupyter Notebook designed to help you get up and running with the following:

1. Data Loading: Shows how to load images from HDF5 files (train-image.hdf5 and test-image.hdf5) and read the accompanying metadata.
2. EDA & Preprocessing: Performs a minimal exploratory data analysis (EDA) and basic preprocessing (resizing, normalization, etc.).
3. Modeling: Demonstrates how to load a pretrained ResNet50, then optionally fine-tune it (placeholder code for illustration).
4. Submission Generation: Creates a submission.csv file with the required format (isic_id,target) for the competition on Kaggle.

Quick Start

1. Clone this repo or download the Jupyter Notebook (.ipynb file).
2. Open the notebook in Jupyter (e.g., JupyterLab or Jupyter Notebook) or directly in Kaggle’s environment.
3. Follow the cells step by step—each section guides you through:
   • Loading data from the .csv files and HDF5 images.
   • Setting up PyTorch datasets, transforms, and data loaders.
   • (Optional) Training or fine-tuning a pretrained model.
   • Generating predictions for the test set and creating a valid submission.csv.

How to Use

• Install Dependencies: Make sure you have Python 3.x, pandas, numpy, torch, torchvision, opencv-python, matplotlib, and h5py installed.
• Data Access: Place the HDF5 files and metadata CSVs in the appropriate directories as indicated by the notebook.
• Customize: Adjust hyperparameters, training loops, and transforms to improve your model’s performance.
• Submit: Upload your submission.csv to the internal Kaggle competition page to see how your solution ranks!

Contributing

Feel free to fork this repository and tailor the notebook to your own approach. If you have improvements or useful scripts, consider sharing them via pull requests so everyone can learn.

Good luck with the competition!