1. Understanding Gradients in PyTorch Neural Networks
   In neural networks, we train the network by adjusting its internal parameters (weights and biases) to minimize a loss function