Add ANN Training Program using PyTorch

[leonking1990]

Description:

This PR introduces the initial implementation of the Artificial Neural Network (ANN) training program for diabetes detection using PyTorch. It includes the following features:

• ✅ AnnModel.py – Defines the ANN model with ReLU activations and a Sigmoid output for binary classification.
• ✅ DataBuilder.py – Handles dataset loading, preprocessing (normalization), and DataLoader setup.
• ✅ Integration with KaggleHub – Downloads the diabetes dataset dynamically, allowing customization of the download path.
• ✅ Training & Evaluation – Implements training with binary cross-entropy loss and Adam optimizer, along with model evaluation.

Changes Made:

1. Implemented DiabetesModel – A fully connected feedforward neural network.
2. Created DataBuilder – Automates data processing and loading.
3. Ensured compatibility with CUDA – Uses GPU if available.
4. Improved training script – Includes proper dataset splitting (80/20) and logging of loss per epoch.

How to Test:

1. Clone the repository and switch to the correct branch:

   git checkout AnnTraining

2. Install dependencies:

   pip install torch pandas scikit-learn kagglehub

3. Run the training script:

   python AnnModel.py

4. Confirm that the model trains successfully and outputs loss per epoch.

To-Do (Future Improvements):

• Add hyperparameter tuning for learning rate and batch size.
• Implement early stopping to prevent overfitting.
• Save and load trained models for inference.

Reviewers:

• Please verify that the dataset downloads and loads correctly.
• Check if the training process runs without errors.
• Suggestions for improving model performance are welcome!