Introduction
Tensors are like containers for numbers. We create them to hold data for machine learning models.
When you want to store fixed data that does not change during training.
When you need a tensor that can change values during training, like model weights.
When you want to create a tensor filled with zeros to initialize data.
When you want to create a tensor filled with ones for bias or masks.
When you want to quickly create tensors of specific shapes for experiments.