feat: Modularize Quantum Acoustic QSVM & 4-Class QCNN Pipeline

[MayuriYatagiri]

Summary of Changes

• Code Modularization: Converted quantum acoustic classification workflows from Jupyter notebooks into structured Python scripts (main_pipeline.py and train_qcnn.py).
• 4-Class QCNN Architecture: Implemented an 8-qubit hierarchical Quantum Convolutional Neural Network with parameterized unshared convolution layers and quantum pooling.
• Optimization Dynamics: Added Adam optimization with a step-decay learning rate scheduler, early stopping, model checkpointing (best_weights), and Euclidean distance decoding in Pauli-Z expectation space.
• Diagnostics & Benchmarks: Achieved 100% cross-validation accuracy on QSVM and a stable ~75% multi-class accuracy on QCNN across 4 audio categories (Nutcracker, Choice, Trumpet, Vibe Ace).
• Project Structure: Added requirements.txt and README.md with usage instructions.

Addresses maintainer feedback from @Ria-K912 regarding script modularity and reproducibility.